Whatnot's CPO on Why the Company "Regrets That Product Management Exists"
Lenny's Podcast
Tom Verrilli is chief product officer at Whatnot, the live-stream shopping platform. They were previously chief product officer at Twitch for a long stretch and director of product growth at Twitter. In a conversation with Lenny Rachitsky, Verrilli explains a line they wrote that surprised many readers: Whatnot's product team was built on the premise that "we regret that product management exists."
The phrase is not a call to eliminate PMs. Verrilli argues that product management should be used deliberately for specific high-leverage needs rather than assigned to every team by default. Senior people, including managers and executives, should stay close to the actual work. AI makes that more feasible than before. Verrilli also stresses that this model fits Whatnot's product and culture, and that it may not fit everyone.
Where product management came from, and why Whatnot "regrets" it
Verrilli starts with history. Product management originally didn't exist. The business side, often a founder-CEO, talked directly with engineering and design about what to build, and they executed together. Internet businesses scaled faster than any businesses before them, so at some point someone had to take over the specifics of execution because there was too much for one person to hold. In that early form, Verrilli says, the role was a working-it-out function. You had to explain less to engineers and designers because they understood the context or had been involved in forming it.
In Verrilli's view, the idea that tech needs a specialist "decision-making class of humans" is a more modern development than a pure necessity. Whatnot's position is that it would be better if engineering and design had enough context, and were close enough to users, to make the same decisions a PM would make.
The only argument Verrilli knows of for making product management a specialist function is that it's "a trade, not a qualification." It is a skill built through repetition, like a muscle. The flip side is that when engineers and designers are kept away from that work, their own muscles stay underdeveloped. Verrilli says product management can be deployed as important leverage on things that must go very well, and that doing so lets design and engineering focus on their crafts in those situations. But where possible, they consider it optimal not to have a PM, and to have design and engineering go through those steps themselves.
Lenny asks whether PM was valuable early and then got diluted by too many mediocre PMs. Verrilli observes that most engineers and designers remember the great PMs they worked with mostly because they worked with so many bad ones. They frame this as honest self-criticism of the function, not an insult. They blame much of the dilution on a scaling pattern: at some point, an "HR ratio" of a pod appeared. For every six engineers, you add a designer, a PM, and an EM. Once a company serves a billion daily users, it has many engineers and therefore many PMs, including in places that don't need one, such as notifications infrastructure. Verrilli says that hiring so many PMs "infantilizes the engineers and the designers who are perfectly capable of making good decisions, but just never had to because there was always a PM to babysit them."
Specialization cuts both ways
Lenny raises the obvious objection. Much of PM work is tedious: alignment meetings, docs, note-taking. Engineers want to code and designers want to design. Verrilli agrees that specialization works both ways. An infrastructure lead can reasonably say they want to think about scale, not alignment debates. Skills also don't always transfer. Someone good at building mental models of how infrastructure scales may not be good at listening to a customer and finding the core problem behind what they said.
So "we regret" is phrased that way on purpose, Verrilli explains, to remind the team that you "don't hire a PM just for the sake of hiring one, you hire one with a really specific need." The culture has to match that. Whatnot's internal documents on how it ships say anyone can bring change forward, and a designer or engineer can be DRI for new product development. But everyone goes through the same process, including product review, because "it is real work." If someone doesn't want to do that work, PMs can be moved around. Instead of mapping PMs to teams on the assumption that a PM will always handle that work, Whatnot maps them to problems or core projects. As a result, some engineering teams have gone a year or more without a dedicated PM despite ongoing product work.
Lenny notes that this kind of talk usually comes from developer-tool companies like Linear, where it's easy to see why engineers might be enough. Coming from someone who has built very consumer-oriented products, it carries more weight. Verrilli says they learned this over time. Twitch was "very Amazon two-pizza team, ratio-driven." Twitter had tight alignment and "a lot of alignment meetings." In their view, much of that structure exists because management can't see what's going on, so it hires more people and adds layers, and "systems beget systems." What is changing, they say, is the realization that leverage no longer requires stacking more PMs under a senior person. It is now easier to understand what's happening in a business and to talk directly with the codebase and with engineers. They no longer think consumer versus enterprise is the dividing line. The culture of the organization is what matters.
How much of this is AI?
Verrilli says the shift isn't explicitly caused by AI, but AI makes it much easier. They describe two factors.
The first is the huge leverage AI gives individual contributors. Verrilli recounts talking with senior leads who feel so capable with AI that they feel pressure about the work they aren't doing, since a couple more hours could move a lot. With well-honed judgment, AI tooling makes it easy to move fast. As an example, they say you can now pull data through a Hex thread that would have taken an Amazon L7 data scientist a week or two in 2017.
The second is cultural. The ratio-driven model came with slogans like "hire great people and get out of their way" and bottoms-up roadmaps. Verrilli thinks tech has recently shifted toward recognizing that top-down can be quite good. People at the top can decide quickly, cut out alignment debate, and usually have more macro context. If they are "genuinely in touch with ground truth" and good enough, having senior leaders involved in many decisions is efficient. One very senior PM covering more areas can be more effective than three entry-level PMs. AI makes this easier again.
Verrilli adds that PMs themselves benefit from not being tied to a single team. Moving across different problems builds more "muscle groups" than staying attached to whatever one EM owns.
How Whatnot's product org is structured
Whatnot has just passed 20 PMs, with 21 or 22 in the building. Verrilli calls this small given the GMV its sellers move. The PMs are loosely organized into three groups: buyer, seller, and trust and risk (standards, payments, safety, and similar areas). Within those groups, PMs have loose alignments, such as broadly growth or broadly discovery, but they are reassigned often.
That comes from how Whatnot plans. Every six months, the CEO, Verrilli, and other senior leads define what needs to be true over the next six months, both outcomes and critical projects. They then go through the list and ask who the DRI is. That is mostly how PM work gets allocated. Verrilli notes that a planning cycle had finished that very morning. Often, at the end of such a review, they find something that is the second or third priority on several teams' roadmaps, clearly important, with no owner. A PM gets pulled in and told, in effect, "congratulations, this is the thing we need you to deliver over the next 6 months." The assignment is rarely a specific feature. It's more like working out what needs to be true. Not every owner is a PM, but most are, chosen for the fit of their skills: financial, algorithmic and recommendations-oriented, or core user features.
31,832 applicants, one hire: what Verrilli looks for
Lenny quotes another line Verrilli wrote: "In the last 2 years, 31,832 people applied to be a product manager at Whatnot. We hired one."
Verrilli is clearest about what is trending down. That's candidates who spend much of their interview talking about driving alignment and stakeholder management. They say there is "a group of PMs," which they were part of earlier in their career, whose specialty "wasn't technical or customer-oriented, it was politics." An answer to "tell me about a time you failed" like "I didn't keep the CEO up to date and that led to a pivot" is an anti-pattern to them.
What stands out positively is seeing both macro and micro thinking over the interviews and case study. The candidate can describe how a system works and an end state, then show near-impatience about how they'd validate it quickly and where they'd push to get it done. Verrilli also looks for specificity about what the candidate actually built. Many people from large companies (they mention FAANG and Uber) "babysat things that existed" and polished the edges. Verrilli wants people who were given a problem and came up with something novel, or iterated through a complicated change and made decisions along the way. In a large organization with inertia, they note, it's easy to go along with what's happening instead of being an agent of change or a decision-maker, "which is ultimately what you need PM to do."
Later, Verrilli describes a pattern they see often. Everyone hired at Whatnot, in any role, does a hands-on case study. Candidates who present very well get a prompt and some data, are asked to return with a PRD, and then have to defend it verbally. Verrilli finds it telling how quickly the thinking falls apart for people who are good at the theater but not the specifics.
Developing systems thinking: "know then go"
Lenny mentions that Netflix's Elizabeth Stone also named systems thinking as the most important rising trait. Her advice was to step back one level and consider how your manager sees the problem. Verrilli treats systems thinking as a skill built through mental exercise.
One small version happens in product review. When someone proposes an experiment, Verrilli asks what the team will do if it's green and what it will do if it's red. If the person doesn't know how their strategy would change, they haven't thought it through yet. The same applies to building local solutions: ask what happens with a thousand times the expected usage, or what knock-on effects might follow, before putting pen to paper or code into the system.
Verrilli argues that this speeds people up. PMs often get slowed by worries that legal, finance, or another team might object. Whatnot's internal motto is "know then go": think through everything that could go wrong, understand where scale will break, and then move anyway. Having thought it through, you'll likely preempt many of the problems. "You don't have to solve all of them, you've just got to think through all of them and then you end up solving more than you think." In a high-growth environment, Verrilli adds, you're not looking for a 5% statistical win but for something that moves the whole business and compounds over time.
Senior people back to IC work
The change Verrilli seems most excited about is PMs, including senior ones, returning to hands-on work. In "ratio land," a successful PM got promoted to director and told to stop being hands-on and just coach. "We took all of our A players and then promoted them out of doing things," Verrilli says. Those people then spent their time in alignment and in reviews, producing what Verrilli calls a "yo-yo" process: someone does the work, it goes to review, gets rejected, and bounces back and forth. They acknowledge this comes from their own "scar tissue."
At Whatnot, four or five people manage other PMs, and Verrilli says all of them spend over 90% of their time on IC work. Verrilli estimates they personally still spend about 50% of their time on IC work. They give several reasons.
First, a VP-level product person with 10 to 15 years of experience should have well-honed instincts and can decide faster than a structure where a problem is split among several PMs who must align across engineering teams that debate each other.
Second, a senior person can cover the work of several junior PMs. Seeing more of the board makes the intuitively correct trade-off more likely. Verrilli's example is an evergreen e-commerce question: how should the discovery algorithm treat sellers who ship slowly? Someone who understands both seller management and discovery can make the right call from either side. They recall that at Twitch, the discovery team and the ads team were "always at war for impressions" over where ads go in the feed and how they affect discovery metrics and ad revenue. Verrilli admits being among the people most at fault. One of their first moves was to put ads into discovery with the same PM accountable for both. With a shared goal, such as value generated from the feed through either organic or paid placement, that person makes the trade-off naturally. When one person spans multiple areas, Verrilli says, they align those areas organically and eliminate months of back-and-forth and politics that turn things "from being company first to career first."
Third, keeping VPs, directors, and the CPO in IC work keeps everyone connected to "what's actually true as opposed to what seems true in a review." Verrilli sums it up with a football analogy: why wouldn't you want Messi playing for your team rather than always bringing the academy along?
Asked what IC work means for a PM, Verrilli answers: whatever is required to ship most effectively. They have shipped some production code at Whatnot but don't think it's the best use of their time. They're sure someone quietly reworked their code to fix linting, localization, and the other things "me and Claude Code did not get right." For them, IC work mainly means being in the support tickets, knowing current customer problems, pulling the data yourself, sitting with engineering and design, querying the codebase directly, writing the spec, and running the stand-up.
Who wants to go back, and how pay works
Lenny points out that many people who climbed to VP don't want to become ICs again. Verrilli expects a split across the industry. Some large organizations may still work best without everyone hands-on. The culture has to match, and every startup reflects its founders. Verrilli's recruiting targets people who used to love product management and are sick of alignment meetings. They pitch CPOs and VPs of product with, "Don't you miss actually doing things?" Verrilli says they've found that many people who spent five or six years as senior directors at places like Meta, stuck in alignment meetings, miss talking to customers and engineers and shipping.
Lenny mentions reports of former CTOs, including from Workday, Instagram, and Box, joining Anthropic as members of technical staff. Verrilli says their greatest wish is for Whatnot's product bench to look like that.
On pay, Verrilli offers what they call a "spicy take": being paid well as an IC is easier in this model. Add up the compensation for five L5s under an L7, and four L7s under a VP, then ask why three people couldn't each be paid VP-level money if they're having that much impact.
"How do you know that was fraud?": staying close to ground truth
Asked why this model is better, Verrilli says the product improves. People have long told Whatnot that leadership staying hands-on won't scale and that it will need many more layers. Verrilli says they found that isn't true, though it takes a different muscle and a deliberate effort to understand ground truth.
Their recurring example comes from growth meetings. Someone says a pattern was fraud. How do you know? It's labeled fraud in the dataset. How does it get labeled? Probably someone in ops. Do you know the SOP they follow? No. Then you don't know it's fraud. An experienced product director pushed this way will go find out, and the system agents use for data labeling gets stronger because a smart, invested person is now helping guide it. Verrilli's broader belief is that "growth can hide all sins." Companies grow, everyone works from averages, and problems go unexamined. The cultural commitment is to "whole ass few things" (crediting Ron Swanson), put the best people deep in the details, and fix things along the way. Over time, Verrilli says, this is more efficient because you understand how things work the first time.
AI adds leverage here too. As a junior PM, Verrilli spent a lot of time asking engineers how hard things would be, which took time away from actual work. Now they can ask Claude for rough levels of effort. If they suspect some "car crash of models" is hitting new users when they open the app, they can load the feed logic and ask who sees what, in what order, and when each thing fires. A tenured person who sees that the result is bad can just make the call. Verrilli says that saves a kickoff meeting, an alignment meeting, a week of writing PRDs, and experiment time.
Advice for PMs in a tough market
For experienced PMs struggling to find roles, Verrilli's main advice is to start doing IC work in their current job. That keeps the skills sharp and, they bet, will help where you are as well as where you go next. They acknowledge the online talk about "PMs are engineers now" and call coding a good muscle to build. They pushed production code to understand the process themselves. But before building, they say, the priority is getting back to scoping the right thing, understanding the problem, and defining what good looks like. Senior people should use their wider view to contribute more. Verrilli also criticizes a coaching style where directors watch junior PMs cycle through reviews on a PRD while already knowing the answer, the "lead a horse to water" approach. They think leaders should help people see what good looks like quickly instead of running an "endless review yo-yo."
Lenny says their own view has shifted: PMs have more leverage improving what the team ships than shipping code themselves. Verrilli agrees. The mechanics of git commits and similar tasks take them much longer than working out and describing the problem. They still try things themselves, for example to judge whether dev tools have really gotten easier instead of taking someone's word for it. But understanding the codebase and working with someone who can execute avoids "a thousand classic traps" engineers learned to avoid early in their careers.
Beyond prototyping: data science, the codebase, and live feedback loops
Asked for the biggest AI productivity gains other than prototyping, Verrilli names data science first "by a country mile." Whatnot uses Hex Threads. They say it's hard to remember working as a PM without tools that let you pull nuanced cohorts, look up an individual user after a report, examine logs to see exactly what that user did and saw, count how many similar users exist, estimate impact, and quickly build sensitivity models, forecasts, and regressions. The same tools help Whatnot ship faster because regressions and odd interactions between products show up sooner. In large, complex systems, those problems usually cause slowdowns such as release trains. Verrilli says that over the past year they spent less time talking to data scientists than ever in their career, while spending perhaps ten times more time in the data than ever before. They say they don't know what the future of data science looks like.
The second gain is asking Claude how the codebase works instead of interrupting engineers. Early in their career, Verrilli aimed to understand systems at the "boxes and lines" level. Now, they say, there's no excuse not to understand them at a more detailed level.
The third may be specific to Whatnot. Verrilli has worked on live products for about a decade and has long been able to watch customers use a feature for the first time. With current AI tools, when a user says they're having a problem, you can examine the codebase live and tell whether it's a real bug or a comprehension gap. That combines a video of the user, real-time code analysis, and observability into "a feedback loop on steroids."
They tell a story from their second week at Whatnot. Sellers can pick auction formats. In "sudden death," the auction ends when the timer runs out. In the classic format, a bid in the last five seconds adds ten seconds. A seller on stream complained that a 7-second timer was too slow and wished it were 3 seconds so they could move more product. Two engineers looked at each other, said it was a config change, updated it in real time, and posted in the show's chat telling the seller to refresh the app. Verrilli says they thought, "I'm with my people." They add that AI makes this kind of responsiveness accessible to many people, but they themselves wouldn't touch production code that way. It's great for "qualified humans," not for them.
The data scientist problem, and which roles are growing
Lenny relays a data scientist friend's complaint: their job has turned into checking flawed analysis from non-data scientists, and half of it is wrong. Verrilli says they're sympathetic and have seen this happen. They see it as another argument for fewer, more senior PMs who have more experience. They also think much of the confusion comes from organizations underinvesting in data engineering, data structures, and labeling. Many of Whatnot's best data scientists now focus on making tracking and attribution correct so analyses are harder to misread. Using AI to find data, like using AI to write code, doesn't remove responsibility for the result. It gives leverage to people already inclined toward that work.
On which roles are rising and falling, Verrilli notes that Whatnot has still hired plenty of data scientists. "Fewer" means fewer for the same output, not necessarily fewer overall, because companies using these tools well can grow faster and take on more. Verrilli would be surprised if the net number shrank. The trend is more like fewer per unit of customer impact. The rising role, in their view, is something like a tech lead running a very small team, an "engineering manager light." As trying things gets cheaper, Whatnot has been incubating small teams to "sit in the corner" and attempt things once considered too hard. Verrilli says this isn't quite prototyping, has increasingly high leverage, and sits between a member of technical staff and a full EM.
The product team of the future
Verrilli says they don't know what product teams will look like everywhere. At Whatnot, they expect the core to look much as before: specialist designers, engineers, and PMs working together on specific high-priority projects, or where the company has high conviction it has a path forward. Around the edges, there will be more free space for anyone, whether designer, engineer, PM, or data scientist, who is confident they can meaningfully improve something. "If you're sitting there on a Friday afternoon and you can't focus on the PRD you're writing, but you're pretty sure you can go and fix something, go for it."
Lenny brings up their own essay from about two years ago arguing that PMs are best positioned to thrive with AI because the key skills are PM skills: deciding what to build, communicating requirements, prioritizing, giving design feedback, and understanding go-to-market and business strategy. Verrilli agrees that product skills are probably the most durable. Their addition is that the industry hasn't rewarded those core skills much over the past five years. It rewarded storytelling, alignment, and strategy frameworks for leadership. Many people with the PM title haven't been building the core skills: understanding the customer, the business, and the technology, and connecting them. Lenny mentions Marty Cagan's term "product theater." Verrilli admits they're guilty of it too and says it's unsurprising after years of being rewarded, but "there's not a lot of place to hide in that anymore."
Why "hire great people and get out of their way" fails
Verrilli calls the maxim wrong, "nuance required." They think it became cover for letting teams set the roadmap and define the problems with near-total delegation. Better hires do deserve more trust, but Whatnot operates on "verify then trust," not full trust or "trust but verify." Verrilli is probably better placed than their direct reports to see how buyer, seller, and trust pieces fit together. The reports are almost certainly better placed to understand how individual features work. Verrilli says they'd get the discovery model's weightings wrong compared with anyone on that team. But they consider it their responsibility to learn this over time, since they approve the team's work and ask the team to make decisions.
The example comes from founder-CEO Grant LaFontaine. In a review, he'll say he doesn't think something is right, pause, and say he's clearing the rest of his day so they can figure it out together. Then he sits with the team and goes through tickets, code, and data line by line, which AI data tools make easier. Verrilli says this keeps him current and sets the tone that the goal is seeking truth. They contrast it with review cultures where PMs just try to get a green light for credibility with their engineers. Planning aligns the company on priorities and is mostly about resourcing, which Verrilli is accountable for. But they can't know resources are in the right places without periodically going deep, down to details like what logic fraud detection uses to invalidate a referral, and whether address signals come from normalized addresses or free text. "You can't make good macro decisions without the micro." Verrilli describes trying to be T-shaped: broad in general, very deep when needed.
Lenny summarizes this as top-down, founder-led involvement producing better results despite micromanagement's bad reputation. Verrilli adds a condition: top-down works when leadership is in the details and "specifically correct." It fails, and becomes what people call micromanagement, when leaders manage from above without knowing ground truth. They don't know any junior engineer or entry-level designer who isn't excited to work directly with the CPO or CEO, since there are no alignment meetings and the feedback is better. The open question is how to do this in enough places. Grant can go deep because work is divided up. He owns three or four top priorities at a time, and anything besides hiring can be cleared so he can nail those.
Working as CPO under product-minded founders
Verrilli has worked for three product-minded founders in a row, and Whatnot has two, Grant and Logan Head, which they call a blessing. Their first rule is not to double up. If Grant or Logan is on something, Verrilli confirms they're accountable and steps back. A PM once jokingly called the alternative the "two dads problem," with conflicting instructions and repeated reviews. So there are long stretches where half of Verrilli's team is on work whose daily status Verrilli couldn't report, and they're fine with that. The second rule: in a founder-led company, "it's not your company, it's theirs." You find a rhythm by asking whether the founder is open to feedback or has decided. Verrilli credits founders' insight and customer intuition as a big reason founder-led companies do well, and sees their job as covering the areas the founders aren't in.
For someone considering a CPO or first-PM role under a founder, Verrilli's advice starts with asking why you want it. The idea that a CPO decides all the roadmaps is "not strictly true." Spend time with the founder to learn how they discuss ideas and how they like to be pushed. Before joining Whatnot, Verrilli had five or six coffees with Grant, went through interviews, and then spent a full day in LA with Grant and Logan working through problems and the roadmap. They tried to be their "most ordinary self" rather than their interview self. Verrilli says they enjoy product leadership because it translates a founder's vision and intuition into reality. The skill is pushing without competing. They've seen CPO-CEO relationships fail when the CPO competed for the vision.
On pushing back effectively, Verrilli says not to treat it as a trick. The aim is to seek truth, not to get a yes. Start with curiosity: does the other person have context you lack? Verrilli asks whether they're understanding the other person's prior correctly and whether some missing context informs it, and they coach their PMs to do the same with them. Then ask whether the person's mind is made up or they're open to input. If it's made up, don't "pick a fight for the purposes of picking a fight." If they're open but want data, bring data or go get it. If there's no data and it's two opinions, "the CEO's opinion is going to win," and that's fine: "check your ego at the door." Finally, make sure you're debating for the right reasons. Wording sometimes matters, but often it doesn't, and fighting over framing can be part of product theater.
"Play the accordion"
Colleagues of Verrilli's mentioned the phrase "play the accordion." Verrilli says they aren't big on frameworks but find this mental model useful. It targets two failure modes. One is iterating forward and "throwing spaghetti at the wall" without a sense of direction. The other is writing long multi-year roadmaps and vision documents, which throws away what Verrilli sees as software's advantage over other industries: learning, such as updating your understanding through A/B tests. With an accordion, you pull it open to take in air before playing, which here means deciding what you're trying to do. But you only make music when you press the keys and push it closed, meaning shipping V1. For the next phrase, you pull it open again (what did we learn, what do we do now?) and then build again. Pulling it open doesn't create value by itself, but without doing it repeatedly you're probably playing the wrong thing. Verrilli's summary: "you've got to do both. You can't over-index on either."
Verrilli's example concerns listings. Most marketplaces depend on listings. Amazon without them would be nearly empty. In live commerce, a seller can hold up a pair of AirPods, describe them, and start bidding at a dollar, and the buyer has what they need. A listing takes about three and a half minutes to create. Showing an item takes zero. So a first conclusion might be that listings aren't needed. Zooming out, though, new buyers will expect search to work, and if Whatnot doesn't know what a seller is offering until it's sold, it can't send someone looking for AirPods to the right stream. The next idea is to require listings for everything. Zooming out again, three minutes per listing drastically reduces how much a seller can sell per hour, which hurts sellers. The point is to keep stretching toward longer-term implications and knock-on effects.
Live commerce versus agentic commerce
Lenny asks how agentic commerce, where AI agents buy on your behalf, relates to Whatnot's human, live model. Verrilli welcomes agents for routine purchases like light bulbs and air filters, and for high-intent searches, such as a specific computer cable, outdoor lights, or black shoes that must arrive by Thursday for a wedding. But they argue most commerce isn't high intent. By their account, about 30 years into e-commerce, it has never exceeded 20% of US retail spending. They put the UK at roughly 75/25. Much shopping is low intent: going to the mall to see what's available. Verrilli describes the value of a shoe store owner's taste, curation, service, and window display, and says shopping is pleasant and social.
With US retail at what Verrilli says is $7.5 trillion, they don't see it as winner-take-all. They describe live commerce as the first combination of the internet's scale and convenience with the social and cultural experience of physical shopping. At Twitch, they recall, a stream under about a thousand viewers was assumed to be uneconomic because of CPM-based economics. On Whatnot, a stream might have 30 to 50 people, and if a mall shoe store had 50 people inside, "you'd never close," because commerce economics differ from entertainment economics. Verrilli sees live commerce as meeting a different customer need rather than competing with agentic commerce.
Lessons from Twitter
Verrilli compares asking someone about Twitter in 2015–16 to a therapist asking about childhood. They say Twitter had nine heads of product in their two years there. A colleague on the partnerships team described it as "chaos which had lived in the events." Besides lasting friendships and what they call a remarkable product diaspora, they took away two lessons.
The first: when you truly achieve product-market fit, it doesn't matter how badly you mess up the organization. Twitter had PMF strong enough that "you could emotionally feel how much people loved your product," and Verrilli found it valuable to learn early that this is what PMF means, not graphs that look okay. Lenny adds that Twitter's network effects survived Elon Musk changing almost everything, and Verrilli agrees.
The second: when people say something is really complex, it often isn't. "Leadership's just weak." During their two years, everyone knew the 140-character limit had to go. There were repeated working groups under the "beyond 140" project, and the team knew, for example, that Japanese users tweeted six times as often as Western users, which conversations with those users attributed to kanji fitting more meaning into the same character count. But no one would make the call, so there was sprint after sprint. Verrilli says it happened roughly a year and a half to almost two years after they left, "and it turns out, you know, nobody died." They imagine editing tweets went the same way.
Failure corner: averages mean nothing to the individual
Asked about a failure, Verrilli says they've probably failed more than they've succeeded. The internal "how we build" document attached to their blog post sets batting .500 as the goal. The common thread in their worst mistakes is: "averages mean nothing to the individual." It's tempting to see that only 3% of users use a feature and remove it. But for some group, that feature may be 100% of what they do, and removing it for convenience can wreck their use case, with network effects amplifying the damage. In e-commerce, where the platform is someone's business, Verrilli compares careless deprecation to a mall "turning off the power in the lead-up to Christmas." Their decisions they're most disappointed in usually came from trusting averages without looking at the individual use cases hidden beneath them. Lenny cites Jeff Bezos: when data and an anecdote disagree, trust the anecdote. Verrilli agrees.
Closing caveats and lightning round
Verrilli closes with a caveat. There isn't one way to do product management, and they don't think AI will shape the industry in one way. They're confident that fewer, more senior, more autonomous PMs is right for Whatnot's product and culture, but they don't assume it holds everywhere, and they acknowledge people who love management and places where that remains valuable. What they believe is broadly true is that this is the best time to return to the core of product work and leave the theater behind. They add that listeners should "assume that at least half of what I've said is wrong," just as, in their estimate, half of what they've shipped wasn't right.
In the lightning round, Verrilli recommends The Hard Thing About Hard Things, which they call the best book about product management. They also recommend Rick Warren's The Purpose Driven Church, which Twitch CEO Emmett Shear used to make sure people read, as a study of how groups come to invest emotionally in something and useful for anyone building a community product. For fiction, they pick R.F. Kuang's Babel. Their TV pick is Apple TV+'s Star City, which they describe as the Soviet counterpart to For All Mankind. Their product pick is the Service NSW app from Australia, which handles tasks like license renewals and title transfers in one place. They're surprised to be praising a government app. The conversation briefly turns to Australia's solar surplus, and Verrilli credits a regulation from about 20 years ago requiring solar panels on new buildings. Their motto is "I'll figure it out": most things are easier than they seem, and if you won't commit the time and effort to solve something, it probably isn't a big problem.
Their favorite recent Whatnot purchase was a live California spiny lobster. A seller with a seafood store on the San Diego docks live-streams crates coming off the boats each morning, auctions them on Whatnot, and ships overnight on ice by UPS. Verrilli uses it to make a point: they had no intention of buying a lobster that morning, which is why they think e-commerce will be large but not all-encompassing.
Verrilli says the 31,832-to-one statistic wasn't meant to discourage applicants. It was meant to show that the growth of the PM title hasn't produced a matching growth in the skills discussed. Whatnot had hired two people the day before and is recruiting PMs, with payments and logistics as current priorities. That includes, as Verrilli puts it, the surprisingly nuanced product problem of shipping lobsters overnight.
As tech companies scaled, somewhere along the line, this HR ratio of a pod popped into being. Every time you hire six engineers, you add a designer, you add a PM. Hiring so many PMs infantilizes the engineers and the designers who are perfectly capable of making good decisions, but just never had to because there was always a PM to babysit them.
Something you wrote online that surprised a lot of people: at Whatnot, product team was built on the somewhat simple premise, we regret that product management exists.
Not a thing you probably hear from a lot of CPOs. We articulated that way to force ourselves to remember that you don't hire a PM just for the sake of hiring one, you hire one with this really specific need. It is better to not assume we need a PM in every place. The only argument for why you would want product management to be a specialist function is really it's a trade, not a qualification. It's something you get good at by doing. It's a muscle, but the flip side of that is the more you abstract your engineers and your designers from doing the same thing, their muscle gets underdeveloped.
They also wrote this brutal quote, "In the last 2 years, 31,832 people applied to be a product manager at Whatnot. We hired one." What do you look for in the folks that you hire?
I can tell you what's definitely trending down. Folks who spend a lot of the time in their interviews talking about driving alignment and stakeholder management because there's definitely a group of PMs whose specialty wasn't technical, it was politics.
You're very excited about this move to IC work. PMs moving away from this big org management world. If you were really successful as a PM, you got promoted into being a director. We took all of our A players and then promoted them out of doing things. Why wouldn't you want Messi playing for your team rather than trying to have the academy coming along all the time?
Today, my guest is Tom Verrilli. Tom is chief product officer at Whatnot, former long-time chief product officer at Twitch, and former director of product growth at Twitter. He's also someone I've been wanting to get on this podcast for so long. Tom has a lot of hot takes and really unique and important insights on the future of the product role, what he's seeing the best PMs doing differently these days and what he's looking for when he's hiring product people for his team now. If you're not familiar with Whatnot, it's a live stream shopping platform. It's the fastest growing US marketplace business of all time. At one point Tom shares how he bought a fresh lobster from a fisherman on the platform. Before we get into it, don't forget to check out lennysproductpass.com for a free year of the hottest and most beautifully crafted AI product in the world available exclusively to Lenny's newsletter subscribers. With that, I bring you Tom Verrilli.
Tom, thank you so much for being here and welcome to the podcast.
Thank you so much, dude. It feels kind of surreal after years of watching.
I hear that. I hear that people come on the podcast. Here you are.
To be first time long time.
That's right. I want to start with something that you wrote online that surprised a lot of people and I think will surprise a lot of people coming from a long time chief product officer, long time product builder, someone that's built a lot of very successful products and teams. What you wrote is since its earliest inception, the Whatnot product team was built on the somewhat simple premise, we regret that product management exists.
Yes, sir. Not a thing you probably hear from a lot of CPOs.
No. Talk about why you feel this way. Talk about how you got to this place. Talk about what this means.
Sure. I mean, I always try and start with like history in order to understand kind of things. And one of the things when you come to product management is if you go all the way back, like product management didn't exist, right? It was like the business and, very often, founder CEO types talking directly to engineering and design about what we needed to build and then executing it together.
And internet businesses, it turns out, scaled a lot faster than any other businesses in history. And so at some point scale meant delegating the specifics of execution to somebody or trusting somebody else to work out what comes next because there's just too many things on for somebody to sit with. But, if you think about most startups or you think about how that evolution worked, that was really a specialist role where somebody was working it out. It was usually you said less things to the engineer where you had to describe in less detail what you were trying to get done to a designer because they understood or they'd been involved. And this idea that we need this kind of specialist decision-making class of humans in tech is really a more modern function than it is a like pure necessity.
You know, the way I would think about it and the way that Whatnot has always treated it is like it would actually be way better if engineering and design had the context that they needed to just make great decisions there, if they were so in touch with users and what they needed that they could kind of make the same decisions that a product manager is. The only argument as far as I'm aware for why you would want product management to be a specialist function is really it's a trade, not a qualification. And what I mean by that is it's something you get good at by doing. It's a muscle, for want of a better term, and as every private trainer has ever told me, muscles are built by reps. And so, the more you do it, the better you get at it. But, the flip side of that is the more you abstract your engineers and your designers from doing the same thing, their muscle gets underdeveloped.
And so, I think product management plays a really important role, and I think you can deploy product management to have a really important leverage on particular things that need to go really well. And in a lot of ways doing that lets design and engineering be the best they can be at their craft in those situations. But, wherever possible, it's kind of optimal to not have a product manager and instead have design and engineering going through those kind of steps and making sure that their muscles are well ripped.
Do you feel like product management was very helpful early on and was important, and then it kind of went through a period of, well, there's way too many PMs that are not that amazing. And now there's kind of this coming back to okay, what is actually an amazing PM and maybe we need fewer of them.
Yeah. The funny thing is I think if you talk to most engineers or even designers, they will remember the great PMs that they've worked with mostly because they've worked with so many bad ones. And I don't mean that as a disheartening pejorative for folks, but I think it's more, you know, you're being kind of really, really self-critical as a function. Does the average PM add a ton of value to folks around them in the way that having really high-quality product management can disseminate clarity and absorb ambiguity in the ways that they can, or like help people make really timely decisions.
And I think a lot of it is this function of like as tech companies scaled and wound up hiring so many engineers, somewhere along the line, and I don't want to pin it on like HR, but somewhere along the line this kind of HR ratio of like a pod popped into being. You know, every time you hire six engineers, you add a designer or you add a PM, you add an EM, and the nucleus exists. And then when you're now serving billion DAU products, you got a lot of engineers. And so now all of a sudden you got a lot of product managers. And in most cases, you probably don't need a PM for notifications infrastructure, right? Engineers are perfectly capable of understanding how that works. And in a lot of ways, as we just described, hiring so many PMs infantilizes the engineers and the designers who are perfectly capable of making good decisions, but just never had to because there was always a PM to babysit them.
This episode is brought to you by our season's presenting sponsor WorkOS. What do OpenAI, Anthropic, Cursor, Vercel, Replit, Sierra, Clay, and hundreds of other winning companies all have in common? They are all powered by WorkOS. If you're building a product for the enterprise, you've felt the pain of integrating single sign-on, SCIM, RBAC, audit logs, and other features required by large companies. WorkOS turns those deal blockers into drop-in APIs with a modern developer platform built specifically for B2B SaaS. Literally every startup that I'm an investor in that starts to expand upmarket ends up working with WorkOS. And that's because they are the best. Whether you are a seed stage startup trying to land your first enterprise customer or a unicorn expanding globally, WorkOS is the fastest path to becoming enterprise ready and unblocking growth. It's essentially Stripe for enterprise features. Visit workos.com to get started or just hit up their Slack where they have actual engineers waiting to answer your questions. WorkOS allows you to build faster with delightful APIs, comprehensive docs, and a smooth developer experience. Go to workos.com to make your app enterprise ready today.
Something that I find people run into eventually when they think this way that I want to get your take on is engineers and designers don't necessarily want to be doing the work of a PM because a lot of the work of a PM is kind of annoying and not fun. And you know, there's a glamour part like making decisions.
Less glamorous than people think, yeah.
Yes, exactly. How do you think about that, especially as the company scales, engineers having to be in alignment meetings, having to write docs, aligning everyone, taking notes, all these like kind of minutiae parts of PM. And also just the big, like, engineers want to build, engineers want to code, designers want to design. How do you think about that element of they may not actually want to be doing that work?
I think this is where like specialization works both ways, right? It is useful to have folks who are well honed in making decisions. It's totally reasonable as well for someone to say, "Listen, I'm an infrastructure lead and I want to think a lot about scale and I do not want to have to spend my time debating alignment or getting those minutiae right." And certain skill sets don't necessarily translate super well, right? If you're really good at building big mental models in your head of how infrastructure should scale, you may not have the skill set of listening to a customer and actually understanding the core problem as opposed to the thing that they said, which again, is just a thing that we've built over time.
So, my supposition of "we regret" isn't to say that we don't want product managers, just we recognize in those situations that you do need to kind of help other functions specialize, but we articulate it that way to kind of force ourselves to remember that you don't hire a PM just for the sake of hiring one, you hire one with a really specific need.
And I think what goes with it, Lenny, is like you have to build the culture of the organization around that. So, for example, when we say we regret product management exists, every time we write documents about how we ship or what we're doing, we're very kind of clear that anyone can bring change forward, that we can have DRIs of new product development that isn't an engineer or is a designer, but that everybody goes through that same level of function of like, you've got to go through product review. You've got to actually do the work because it is real work. And if people don't want to do that, if they kind of want to do something else, we can always move product management around. And so, rather than mapping PMs to teams where you kind of assume that the PM will always do that, we tend to map them to kind of problems or kind of like core projects. And that means that there will be a year or more where there isn't a PM attached to a particular engineering team even though there's lots of ongoing product work to do.
What I love about this is we often hear people at like eng-oriented products talk like this. Like developer tools where we don't need PMs. Why do we need PMs? And it often comes from companies like Linear and, I don't know, like companies that are building developer tools where you could see why engineers are enough. And so, it's really interesting to hear from your perspective because you built very consumery products, very, very consumery products that require what you think are very strong PM skills. So, it means a lot even more so coming from you that you find that PMs aren't as necessary as people may think in building something great.
Yeah, I think there were like two parts that drove it. And certainly a lot of what I'm kind of talking about here are things that I definitely learned over time. I spent seven years at Twitch prior to my time at Whatnot, and that was very Amazon two-pizza team, ratio-driven. I started in the valley at Twitter, and that was similarly very like you always had this kind of very tight alignment. And God, there was a lot of alignment meetings, so you can imagine why engineers didn't want to be in them. And it's kind of evolved over time to understand that actually a lot of that is just a function of management not being able to see what's going on. And so you hire more folks, and you build more layers, and your systems beget systems.
And I think what's really changing is people are realizing that it's not true anymore that the only way to get leverage is just to hire more PMs underneath you and be more senior. We've got a better understanding of what's happening in businesses now. It's easier to kind of like converse directly with the code base and talk to your engineers than it ever has been. And so you don't actually have to get into this like all-scale thing. And I don't really think consumer or enterprise is the cutting point anymore. I think it's more kind of the culture of the organization itself.
How much of this shift in your mind is AI-driven because AI now enables non-PMs to do PM-y work? And how much of it was like pre-AI? And then I want to talk about just what this actually looks like at your team, but let me ask that question first.
I don't think it's explicitly because of AI, but I do think AI makes it a lot easier. I think it's two things. I think AI certainly means that there's an enormous amount of leverage for an IC now. I don't know how much other people feel this, but I was talking to a couple of our senior leads the other day, and we feel that we're so capable of doing things now with AI that you almost feel a bunch of pressure of the stuff you're not doing because you know that if you carved out a couple more hours, you could move a lot of stuff.
It's certainly easier to move fast with AI tooling if you've got well-honed judgment. You can call data now on a Hex Thread that is basically what it would take a week, two weeks with an Amazon L7 data scientist in 2017. And so, if you're empowered and have good judgment, you can basically make really quick decisions and just kind of keep people moving, which is extraordinary.
But more than just AI, I think it's also a lot of those kind of like ratio pushes also came with a bunch of other cultural pushes like, you know, hire great people and get out of their way and bottoms-up roadmaps and all of those pieces, where I think there's been a cultural shift in tech over the last little while, which is like, actually top-down's pretty good because folks at the top, generally speaking, can make quick decisions and remove all of this alignment debate. They have probably more macro context than most folks. And assuming that they are genuinely in
touch with ground truth and are good enough, it's actually a really efficient model to be able to kind of have senior leadership involved in a lot of those decisions, which means you it's not just the tooling that makes them efficient, but you can have one very senior PM across more things and they can be more efficient than having, you know, three relatively entry-level PMs. But certainly AI makes all of that a lot easier again.
So, just again coming back to the broad premise, which I think is very important to clarify, this point about regretting product management exists isn't saying we don't want or think PMs are useful. It's that it is better to not assume we need a PM in every place and it is great to enable other functions to do the PM work. And also just PMs kind of take away the reps from people being able to do the things that PMs do. And if they can do that work, they can actually execute better, they'll build better products.
I think it's also better for PMs to not be mapped specifically to a team than it is to say, we go with the workers. You will build more and better reps, you're going to work out more muscle groups by like moving around on the things you work on as opposed to saying, I'm attached to whatever this EM owns.
So, let's follow up on that. What does the team look like? What does the PM/design team look like at Whatnot? What does this org look like in this world view?
Yeah, we've just passed 20 PMs. We have kind of I think 21, 22 PMs in the building today, which for those following Whatnot's trajectory is pretty small considering the volume of GMV that our sellers move.
We're very loosely organized into three groups, kind of buyer, seller, and what we would call kind of trust and risk. So, the folks looking after our standards, you know, payments, kind of safety, etc. And then within those kind of like broad groups, we basically reassign the PMs pretty regularly. So, people have like loose, you know, alignment. You might be like broadly a growth PM or broadly kind of like work on discovery, but even amongst those groups, it kind of gets allocated and moved around pretty quickly. And that's because the way that we do planning is like every 6 months, you know, the CEO, myself, some other folks and senior leads will sit down and just define what needs to be true over the next 6 months. What do we have to get done as a company? Both in terms of like outcomes and critical projects. And then we sit down and we go through that list and say, who's the DRI? Who's accountable for that? And that's mostly how we end up allocating PM work. And so, quite regularly you'll have something, you know, we've just finished a planning cycle literally this morning, and you'll quite regularly go through those and you'll get to the end and say, cool, there's this thing that's like second or third priority in a bunch of different teams' road maps. That feels really important. Who owns that? We actually don't have an owner for that. And we'll go and grab a PM and be like, congratulations, this is the thing we need you to deliver over the next 6 months.
It's rarely you need to build a feature that works exactly this way, that does this thing. It's a little more high-level than that. But like, how do you go and work out what needs to be true? And then you map the human beings that you think can kind of go get that. Not every name is a PM name, but overwhelmingly when you go through that planning process, it tends to be PMs and it tends to be kind of PMs who have the right skillset vis-a-vis what it is, whether it's more financial, whether it's more kind of like, you know, algorithmic and recommendations-based, whether or not it's like core user feature.
Let me follow that actual specific thread at the end there around what you look for in product managers that you hire. So, you also wrote this brutal quote. In the last 2 years, 31,832 people applied to be a product manager at Whatnot. We hired one.
Yes, sir.
Pretty great. Pretty great. So, there's a few things here I want to talk about. One is just what is it you look for in the folks that you hire, especially these days? What do you find is kind of like trending up in what you find you need in really successful PMs at Whatnot and what's maybe trending down.
I can tell you what's definitely trending down. It's folks who spend a lot of the time in their interviews talking about those like alignment meetings and driving alignment and stakeholder management and those pieces because there's definitely a group of PMs, and I certainly used to be one of them earlier in my career, whose specialty wasn't technical or customer-oriented, it was politics. And so, folks who tend to kind of like naturally lean towards like driving alignment, building relationships, you know, tell me about a time when you failed and like, "Oh, I didn't keep the CEO up to date with something and that led to a pivot," is definitely kind of a thing that is a bit of an anti-pattern that trends down.
What we tend to find kind of really jumps in a PM kind of interview is over the course of your interviews or your case study, can we see both the macro thinking and the micro thinking? I think the system works something like this and I can describe an end state, but can I, you know, almost exude impatience on like, "And here's how I would validate that very quickly. Here's where I would push to get that done."
And are you specific about the things that you've built? All right, there's an awful lot of folks who've worked at, you know, FAANG, Uber, pick any scale company where they babysat things that existed and they maybe, you know, polished the edges of it as opposed to the idea of saying, we were given this problem and I had to go and come up with something unique and novel or I had to kind of really iterate our way through a complicated change. But I made decisions along the way and we moved, because I think it's easy in a very large organization with inertia to kind of go along with what's happening and not necessarily be an agent of change or be a decision maker, which is ultimately what you need a PM to do.
There's something you said there. I just had Elizabeth Stone on the podcast. She's CPTO at Netflix and I asked her what's the trait that is most trending up and she said exactly what you said initially, which was the systems thinking, thinking big picture, thinking about the bigger business. And her advice there to work on this, and I want to ask you if you have any other advice here. Say someone hears this, they're like, oh wow, I got to work on my systems thinking skills. Her advice is take one click back from your problem and think about, okay, for my manager, what do they think about this problem and how does that impact the rest of the business? Thoughts on just how somebody might develop the skill and get better at systems thinking?
I mean, I think the skill is exactly the right way to say it. I think you can get good at it from just a mental exercise. So, like, you can do it in small ways. One of the things I ask a lot in product review when someone says, hey, we want to run an experiment, is, okay, what do we do if it's green? What do we do if it's red? And if folks are like, actually, I don't know how my strategy would change, like, cool, we haven't really thought about that. So, like, stop and go and do the mental exercise of how it would work.
And then I think it's the same thing if you can build quick local solutions if you have done the mental exercise in your head of saying, well, what would happen if we had a thousand times more usage of this than we expected or what are the unexpected knock-on effects that this could have? And how do you just start doing all of that in your head before you get pen on paper, before you get code on the system? And I think it helps actually unblock people to move faster, too, because the other thing that PMs are often slowed down by is like oh, risk, oh, legal might, finance might, another team might.
And I think one of the monikers we use internally is know then go. As in just like think through all the things that could go wrong, understand where scale will break, understand the things that might happen and then make the like move on anyway because if you've thought through all the things that could happen at scale, you're probably going to preempt a bunch of them. So, I really like Elizabeth's quote there, but mine is like just do the mental exercise of just like playing out if it gets really widely adopted, if it happens, if there is something that goes wrong, what will it be? You don't have to solve all of them, you've just got to think through all of them and then you end up solving more than you think.
I love that. So, it's essentially don't just focus on will this be an impactful experiment and think about what comes next and what comes next if this is true.
Like, okay, I moved it. Particularly in a higher growth environment, you're not really looking for a 5% stat-sig win. You're looking for something that kind of like totally moves the business and that has kind of compounding vector over time. And so, you just got to think through what that is.
Something else you said that is changing in how PMs operate that you're very excited about is this move to IC work, PMs moving away from this kind of big org management world to actually doing the work. Talk about that and what that means for the role of product management.
I mentioned a little bit earlier, but there was this thing in the, you know, in the ratio land where like what you did if you were really successful as a PM is you got promoted into being a director and then all of a sudden it was like don't be hands-on anymore. Your goal is just to coach and guide. And so, we took all of our A players and then promoted them out of doing things. And they spent all of their time in alignment and they spent all of their time kind of like coaching and tweaking what their team was doing and you get this really yo-yo development process where somebody does all this work, it goes through a review, it gets told no and you're just kind of going back and forward in reviews. You can see my scar tissue coming through.
On our team, there are managers. There's like I think four or five people across the team who manage other PMs. All of them would spend 90 plus percent of their time doing IC work. I'm still probably 50% of my time doing IC work personally.
And I think there's kind of a couple real advantages of it. The first is if you are a kind of VP of product where you've got a decade plus, maybe 15 years of experience building things, hopefully your instincts as to what's going to work or not are fairly well honed at this point, you can just make decisions more quickly than people. You can have real impact very, very quickly. And it's really great for the organization to have somebody who can do that as opposed to the idea of like working through three layers of, you know, we divide the problem up against a couple PMs. Those folks need to get into alignment. There's different engineering teams debating stuff. You just tend to go.
So, that's wonderful. I think the second thing is there's two ways that that ends up driving leverage. One is a VP in theory can, you know, handle the workload of multiple, kind of like more junior PMs just because, as I said, they're more efficient. And that means that you see more of the board at any given point in time, and so you're far more likely to make the intuitively correct decision for how should we tune the discovery algorithm vis-a-vis people who ship slowly, which is a, you know, an evergreen thing in e-com. Well, if you're thinking about how we, you know, manage sellers who ship slowly, and you know about the power of kind of like discovery, you can in either case make the kind of correct decision.
You know? One of the things that I remember plaguing Twitch for a long time, and I was probably one of the people more at fault of it than ever, was the discovery team and the ads team were always at war for impressions. Right? Like, where do ads go in the feed? What's the impact to discovery metrics? What's the impact to kind of ad dollars? One of the first things I did when I got to Twitch was just put ads in discovery, and make sure that there's the same PM who's accountable for both. Cuz they're going to make the natural trade-off that say the goal is GMV generated from the feed. One of them is through organic, one of them is through kind of like a paid substitution. Solved. And when you put the same person across multiple things, they tend to organically align those things and you just cut out months and months and months of back and forth and the politics that tends to kind of take it from being company first to career first. And so, having VPs mostly in IC land, having directors mostly doing IC work, even having me having to grapple with IC work keeps everybody connected to the ground floor of what's actually true as opposed to what seems true in a review. But also means that you're more likely to make the kind of intuitively correct decision early. Just because why wouldn't you want, you know, Messi playing for your team rather than trying to kind of have the academy coming along all the time.
I love this. So, when you talk about IC work for a PM, what does IC work for a PM in this context mean? Does it mean shipping code, building, or is it like running a team, owning a road map, writing the strategy doc? Imagine it's the second bucket.
Whatever is required to kind of like most effectively ship is the short answer. Have I personally shipped some production code at Whatnot? Yes. Do I think that's really the best use of my time? Not really. You know, I'm certain that quietly somebody is reworking most of my code in order to ensure that the linting was correct and the localization worked and all of the nuances that decades of software engineering has taught you that, you know, me and Claude Code did not get right. But I do think it starts with like are you literally in the support tickets? Do you know what customer problems we're having? Have you pulled all of the data yourself so that you actually understand it? Have you sat with engineering and design, you know, have you queried the code base directly in order to understand how things work and then have you written the spec? Are you then running a stand-up in a week? I think all that is just like core individual IC work.
There's a lot of people that have worked their way up the ladder of product, become a VP, and it doesn't feel exciting to go back to being an IC. Some people, like clearly you, love it, you enjoy it. A lot of people are like, "Ah, I thought I was done with this. I could just work through people. I could think big picture." How do you feel about that? And what would you say to folks in that bucket?
I think there are probably still a lot of organizations where that is really valuable and that they will go. My recruiting tends to be the folks who are, "Oh my god, I used to love product management and I'm so sick of sitting in alignment meetings," and I'm out there pitching CPOs and VPs of product to be like, "Don't you miss actually doing things? Do you want to come back?" And I think it's okay for us to acknowledge that there'll be a bifurcation across the industry. I do think that there are organizations that are sufficiently large that maybe everyone being hands-on isn't right. I also think, you know, I mentioned earlier that like you have to have a matching culture to kind of go with this kind of environment where that's the expectation, you know, where people are like, "I don't want to talk about it. I want to just, you know, let's go and do that thing." And in a lot of ways, you know, every startup is a reflection of their founders. And so that naturally tends to be, you know, how do they think and how do they want to run the organization? But I've actually found that for a lot of the cases you go and talk to somebody who's spent the last five, six years as a senior director
at, you know, Meta who spends their entire time in alignment meetings and they miss actually talking to customers and talking to engineers and shipping things.
What makes me think about it, I imagine you've seen this list of all of these chief technology officers that have gone to become just engineers at Anthropic. I pulled up this list as you were talking. The CTO of Workday is just a member of technical staff at Anthropic. CTO of Instagram, Box's CTO, super.com's CTO, they're just like engineers now at Anthropic.
Yep. This is my greatest desire, that the Whatnot product bench basically looks like that.
Mm.
All of these people with like great skills and great understanding and actually end up coming and building as ICs.
What about the comp of this path, you know? That's people's dream, move up to VP, make millions of dollars. Is there a world where you can still do that and be an IC?
I actually think it's easier. Spicy take, but like go and take the comp required to have five L5s reporting to one L7 and then four L7s reporting to one VP and now total the comp of that product org and turn around and say, what if I had three people? Why can't I pay them all, you know, D2 VP money, particularly if they're having the level of impact that those folks are having there? Like why not?
And just to fully understand why this is happening, why this should happen, what I'm hearing there's kind of many combinations. One is AI is enabling this, which is great, perfect timing.
Yes.
What are the other motivations to do this? Is it just the product ends up being better? Is it fewer people?
I certainly think the product ends up being better. Like one of the things that folks have been telling us for a long time is yeah, that won't scale. Like oh, leadership isn't going to be able to stay hands-on with what's going on. You're going to need to go and hire tons more layers and what we found is that's actually not true.
Right.
It requires a different muscle. You have to make an effort to make sure you genuinely understand, you know, like ground truth. An example we use all the time is we'll be talking in a growth meeting and someone will say, oh yeah, but that was fraud. You know, turn around and say, how do you know that was fraud? Oh, it's labeled in the data set as fraud. Okay, do you know how it gets labeled? I assume someone in ops does it. Okay, do you know the SOP or how they label that? No. Okay, so you don't know it's fraud.
And if you push, you know, a really experienced product director, they're going to go, good point, I don't. I'm going to go find out. And then you invariably end up strengthening the system that agents are using to kind of, you know, data label because suddenly there's a very smart person who's very invested in like understanding how we do that and helping guide it.
And so, just that attitude that says like, we're going to do fewer things, we're going to make sure we execute the hell out of them, and we're going to kind of push our best people to be in the weeds everywhere means that you fix lots of things as you go, and you don't end up kind of just making loads of trade-offs. And I think there's a general belief that like it's too easy otherwise for growth to hide all sins. Like you get bigger and you just end up scaling and everybody's kind of like working off of averages. So, culturally I think you got to be really committed to like let's, you know, whole ass few things, as I tend to say sometimes. Shout out Ron Swanson. And then push your best people to be really in the weeds of stuff. And actually what you find over time is you end up being more efficient by doing that because you actually understand how things work the first time and you make the best decisions.
And then listen, AI, huge leverage for all of this. I can't think of how much time I spent as a junior PM asking my engineers how hard something would be and, you know, distracting actual velocity in order to help scope future stuff. And now I can sit and, you know, talk to Claude and understand roughly LOEs. I can sit there and go through and be like, it feels like there's some car crash of models that must be hitting new users as they open up the app, and you can literally just go through and be like, let's load up the feed and talk to me about the logic of who sees what in what order and when does this thing fire, and you can get answers really quickly.
And having, as I said before, folks with enough tenure and enough reps that they can see that and say, you know what, that's bad, let's just make a good decision and change the ordering of those. I've saved now a kickoff meeting, an alignment meeting, a week writing PRDs, experiment time, all of that by just having somebody who's kind of empowered to go and make a decision.
So, coming back to people that are trying to get a job as a PM, whether you're new, let's actually hold off on new people, but people that are, say, managers, senior folks that are just like, wow, the market has really shifted. A big thing we're hearing right now is you need to be comfortable with moving back into IC, giving up your fancy title. Is there anything more along those lines of just people looking for a job, struggling to find a job? Any other advice for them?
Start doing IC work in the role you're in would be my push, like get back to the basics of like make sure that you're taking on kind of practical work. Cuz I just think it's like good to make sure that you're keeping those muscles, you know, well honed. I think it also starts with like pushing internally for those things. Like I would bet that if you started bringing that level of productivity back into the role that you're in, it probably helps you where you are in addition to kind of help you where you might move to.
But I think there's a lot of chatter online obviously about like PMs are engineers now. And I think that's all well and good. Like it's a great muscle to go and hone. As I said, I've pushed some production code because I wanted to go through the exercise of understanding it. But I think there's also, before you get to like building things, it's like how quickly can you get back into the muscle of scoping the correct thing, understanding the problem, being able to define what good looks like. And like take advantage of the scale and the scope that you've got, that you can see more and bring more to things.
I think most folks who have sat in, you know, a director plus role will know the pain of sitting there and watching a junior PM yo-yo back and forth on the same PRD back to review where you kind of know what the answer is. Somewhere along the way we decided that, you know, lead a horse to water as opposed to kind of help them understand the answer and then keep moving. And I think there's like something in our coaching styles that we can get back to of like help somebody understand what good looks like relatively quickly as opposed to just endless review yo-yo.
This point you make about PMs not shipping to production, I so agree with. My mind changed on this recently with a previous podcast guest with this point that PMs are really like the leverage. PMs have so much more leverage if they're not sitting there trying to ship to production. Enabling the team to ship better and faster and making sure the things that are shipping are better is a much better use of PM's time than sitting there shipping stuff.
I mean this sounds really silly, mate, but like it takes me substantially longer to go through the minutia of like git commit and all of the kind of like pieces that are second nature to an engineer than it does to actually work out what the problem is and be able to describe it. So yes, like good practice to try and, you know, make sure you're doing something. For example, I don't want to judge if our dev tools have gotten easier or not based on what someone tells me. I'm going to go and try it and be like, yep, that was easier than last time I did it. But I do think you can get an awful long way understanding the code base and then talking to somebody who can actually execute well, you know, otherwise you're going to fall afoul of like a thousand classic traps that every other engineer learned how not to do when they were in L4.
So following that thread a little bit, what are some of the ways that AI has enabled you and/or your team to move faster and be more productive, other than prototyping, which is the very clear benefit of AI for PMs and product teams? What else would have been the top three of like, wow, this has really unlocked our productivity and the quality of what we do?
I mean the first one I think by a country mile is data science. You know, we use Hex Threads internally at Whatnot. I'm sure there are other kind of like comparable products, but I think it's almost hard to remember time as a PM before you had tooling like that where you could genuinely start pulling very nuanced cohorts of data, where you could kind of grab an individual user where you've heard a report and actually understand, you know, like let's pull logs. Help me understand exactly what this user did and saw. How many other users look like this? You know, what would impact be? And then all of a sudden you can build pretty meaningful sensitivity models or like forecasts of what might happen, regression models, etc. Really, really, really quickly, which is like incredibly powerful.
The other thing that we found is it helps us move much more quickly with shipping things because you can spot regressions and weird knock-on effects of two products intermixing more quickly than you used to be able to. And I think in, you know, really large complicated systems, that's always one of the things that ends up slowing you down, of like release trains and all of that, versus like if you build the right AI tool, you can spot regressions really quickly, which basically lets people just kind of like go. So, I don't know what the future of data science looks like, but I think as a product manager, I've spent less time in the last year talking to a data scientist than I ever have in my career, even though I've probably spent 10 times more time in data and understanding actually how the product's working than I ever have in my career. So, that one I think is really powerful.
Second one I've already mentioned, which is like stop bothering engineers with how does the code base work and actually just go and talk to Claude and understand it, which is really helpful. You know, I used to say earlier on in my career that the goal was always to understand your systems at the boxes and lines level of, you know, which system drives which thing, and now there's no excuse not to understand that or a nuance layer.
But the other one, and this might be very specific to Whatnot, so I don't know that this will help everyone, but one of the things that I've been lucky to do in my career is basically worked on live products for a decade now. And so, it's always been really cool to be able to ship a product and then watch a customer use it and watch them kind of figure it out. So, I think people have just gotten this experience with like Listen Labs and, you know, others in that cohort of watching people use your product, but I've always been able to sit and watch people use a thing for the first time and go through that new user comprehension gap.
What's really cool with a bunch of the AI tooling right now is as they're describing, oh, I'm having a problem, you can literally be watching the code base live and work out is that actually a bug that's happening right now, or is that a comprehension gap where it doesn't work as expected. And suddenly you've got this video artifact of someone using your product. You can be analyzing the code base in real time. And you can be just talking kind of through AI to the code base to understand what's actually happening, and it's like, you know, a feedback loop on steroids because all of a sudden you know exactly what's going on customer side, code side, and observe side as a viewer in real time, which is really cool.
That sounds both awesome and very stressful, to be building products that live in real time. I think about Netflix where they invest in live now, and that's just like all you've done for 10 years, and how big of a deal that was for them. I know the scale is different, but
Honestly, my second week at Whatnot, I remember sitting in a room watching... so, for those not familiar, kind of commerce auction, live auction platform, but live commerce platform. And there's varieties of different ways to run an auction, but one of them is what's called sudden death, which is like when the timer ends, it ends. Right? Otherwise, the classic auction environment, somebody bids in the last 5 seconds, it adds 10 seconds back on the clock. And a seller can decide what auction model they want.
And I was sitting there watching a seller who was like, "These are taking too long. I wish this 7-second timer was actually 3 seconds cuz I want to move more product." And I watched two engineers in the office look at each other and be like, "That's a config. We could totally do that." And so they went and updated it in real time and then jumped in the chat of a show and just said, "Refresh your app." And then all of a sudden, bang, it was operating that way and I was like, "Cool. I'm with my people. I'm in like the right place." Because that's the level of like responsiveness that you can get in a live environment, and obviously AI means that that's really easy for lots of people to take on. Not that I would touch production code in that kind of way cuz that's a disastrous idea. But for qualified humans, it's a wonderful one.
That is very cool.
This episode is brought to you by Mercury, radically different banking loved by over 300,000 entrepreneurs. And now with Command. I've been a customer of Mercury's for over 6 years. I have never once thought about leaving. Mercury is basically what happens when banking is built by product people, not by bankers. They make it so easy, dare I say fun, to send invoices, move money around, set up virtual cards for folks on my team. Does your bank have an API, a terminal native CLI, or an AI-ready MCP server? I don't think so.
And just recently they launched Command, a conversational interface built directly into Mercury, which acts as your financial operator. I've been using Command to transfer money around, to figure out what categories I've been spending the most money in, analyze my cash flows, and just today I used it to find out how much I've made from a specific sponsor over the past year. I just asked, "How much have I made from X over the past year?" 10 seconds later I have an answer. It is so freaking cool. Visit mercury.com to learn more and apply online in minutes. Mercury is a fintech company, not an FDIC-insured bank. Banking services provided through Choice Financial Group and Column NA, members FDIC.
Coming back to this data science point, I have a friend who's a data scientist, and he said it's a rough time for data scientists because of this exact thing. And to add a little more color, basically their time used to be: asked to do some data analysis, do some work with the data, come back, here's the results, I'm confident in this conclusion. Now their time is basically seeing like half-assed data science work from non-data scientists, and just like, show me, is this right? And half the time it's wrong. And they're like, what the hell is my job now? It sucks.
Yeah, so I have a lot of empathy for that because I've certainly seen it. Another plug for why having fewer, more senior PMs is helpful, because there are folks who've seen more of those reps and understand. But I also think a lot of the time that lack of clarity comes from the other thing, which is organizations have historically underinvested in data engineering and data structures and good data labeling, and whether or not you've got your kind of not just your tax on a
maze, right? But whether or not you really understand whether or not your data systems and structures are set up well. And so what we found is a lot of our best data scientists are pushing in that direction of, like, are we actually correctly updating all the ways in which our tracking and attribution works so that it's less easy for people to kind of misunderstand those? And then, yeah, using an AI tool to find a piece of data, much like using an AI tool to write code, doesn't absolve you of responsibility to make sure that that was good analysis, good code. It just tends to be leverage for those people who are naturally inclined that way.
So if you think about these different functions, data science, user research, design, engineering, PM, what I'm hearing so far is we'll need probably fewer PMs, we'll need fewer data scientists. Are there any other roles that you think are kind of trending down in terms of what we'll need, and are there any roles trending up, like, wow, we're going to need a lot more of this kind of person?
Well, funnily enough, like I say, I haven't spoken to a data scientist in a while. We've certainly still hired plenty, because I do think all that tracking and attribution and measurement is really, really powerful. I also think one of the flip sides of when we say we need fewer is, like, fewer of for the same output doesn't necessarily mean fewer of in macro, because if you are using these systems right, you can just grow more quickly, you can build more things, you can take on more stuff. So, you know, I would be surprised actually if we ended up with net fewer. I think it's more like net fewer vis-a-vis customer impact for both.
Certainly, I think trending up over time within these roles are these kind of, and I don't remember the exact label that people used to use, but this idea of tech leads that are kind of a hybrid EM, where you're running a very small team, kind of running at things. Because the same way that we say, you know, the cost of trying something has come down, the idea that says you can have core focus, which is a thing we're very big on, we've also found incubating a lot of smaller teams to kind of go over and sit in the corner and just try and build this thing, going, you know, it's not quite prototyping, but it's like go and take a swing at a thing that we've historically thought was too hard, is getting cheaper and increasingly has very high leverage. So the idea of engineering manager light, for want of a better term, is a thing that I definitely think we'll see more and more of. It's not quite the technical member of staff, although that must be lovely, but certainly not the full-on EM role. I think it's one that definitely will come about more often.
So maybe just to close the loop on this part of the conversation, there's this, you know, trend towards everyone's kind of a builder, PMs are shipping a little bit, being a little more engineer, engineers are taking on more of the PM work. How do you think this just kind of maybe plays out over the next couple years in terms of what product teams look like broadly? Is it still PM, engineers, a designer, data scientist somewhere? How do you envision the canonical product team over the coming years?
Great question, and I don't know that I know what it looks like everywhere. I think what it'll look like at Whatnot is it probably still looks mostly like it has historically, which is, you know, there are reasons that you'd have a specialist designer, specialist engineers, specialist product management. I think those very concrete teams will largely be reserved for very specific projects, or things that we have high confidence or conviction that we need to solve, or we're pretty high confidence and conviction that we have a path forward on a thing and we want to make good progress.
And I think at the edges around that is going to be a lot more free space for people to play in, like, "Hey, I'm reasonably sure I can go and make a meaningful improvement to this thing," and it doesn't matter if you are a designer, an engineer, a product manager, a data scientist, you can and you should. Right? If you're sitting there on a Friday afternoon and you can't focus on the PRD you're writing, but you're pretty sure you can go and fix something, go for it. And so I think it probably doesn't morph in the more formal sense, but I do think there's just a lot more free space, and people who are well-versed in the customer problems, well-versed in the code base, and understand some of that macro context will just be empowered to do more and more things.
Two or two and a half years ago I had this post I put out where I said, "Why PMs are the best positioned role in tech to thrive in an AI world." And I feel like, even though the beginning of our conversation was like, we should live in a world where we regret PM exists, I feel like we agree on this idea that the skills that seem to matter most and are going to be most valuable, whether it's a PM doing them or an engineer or designer, are very PM-y skills. I'll share a few examples from this post. Like, who is really good at these things? Identifying what to build, distilling and communicating requirements, prioritizing everyone's ideas for the highest ROI opportunities, giving feedback on designs to improve impact, developing go-to-market strategy, understanding business strategy. To me this is what PMs do, and it feels like that's becoming more and more important. So I guess the question to you is if you agree with this idea that the PM-y skills seem to be the most valuable now as AI takes on the building.
Definitely agree. Also, props for bringing in the receipts even with a timestamp, my friend. Well done. I think my only build on that would be to say I think those core PM skills aren't necessarily the things that we have rewarded PMs for over the last 5 years, versus storytelling, alignment, strategy. And so I do think you're 100% correct that, like, can I genuinely understand the customer, can I genuinely understand the business, can I genuinely understand the tech, and can I translate the three together for, you know, optimal efficiency? That is the point of leverage when doing things gets cheaper, trying things is cheaper, etc. So absolutely agree that product skills are probably the most durable. My build would just be there's a lot of people who have the title PM who haven't spent a lot of time building those skills in the last 5 years, but have gotten really good at communicating frameworks to leadership. And so I'd just kind of push us back as a function into that core work.
Yeah, Marty Cagan calls this product theater. A lot of people just do the things that PMs should be doing.
Or, and listen, I'm guilty of this. We have rewarded it for so long that it doesn't surprise me that product theater is a core skill set for a lot of folks. I just think that there's not a lot of places to hide in that anymore.
And this comes back to those 32,000 people who applied for jobs. I imagine a lot of that is just people who think they're PMs or have the title PM, but are exactly what you described. They just focus a lot on alignment, writing docs, meetings, things like that, and not actual building, understanding what it takes to build a successful product.
Yeah, the number of people who do really well in a product interview, Lenny, and then you give them a case study. So everyone who gets hired at Whatnot in any role has to do an actual hands-on case study, but the number of people who present incredibly well, and then you give them a prompt and some data and ask them to come back with a PRD on something, and then we make them verbally defend it, how quickly the thinking decays from folks who are good at the theater but not the specifics is really telling, I think.
Okay. So going beyond the hiring step, say you hire somebody, what are some things you've learned about how to get the most out of the people you hire? You mentioned this kind of contrarian take that you don't agree with this "hire great people, get out of their way." So I want to hear more about that. And just, is there anything else you've learned about elements to building a very successful, world-class product team?
I mean, nuance required in my "hire people and get out of the way is wrong," I will say, but a couple of pieces to build off of it. I think in general "hire great people and get out of their way" became this kind of macro saying for: let them work out what the road map is, let them work out what the problems are, just completely devolve, you know, what's going on. And I think the real answer is obviously the better people you hire, the more you can totally trust that they know what they're doing. But we tend to live in a verify-then-trust land, as opposed to a totally trust or even trust-but-verify, which is like, I'm probably in a better position than any of my directs to understand how all of the different pieces of our system, buyer, seller, trust, fit together. They're almost certainly in a better position than me to understand the nuance of how any of those individual features work. You know, if somebody quizzed me today on exactly all the weightings in the Whatnot discovery model, I would definitely be wrong vis-a-vis any of the engineers on that team, vis-a-vis any of the PMs on that team. Good. But it's actually kind of incumbent on me to learn that and understand that over time, because I'm asking them to make decisions and I am approving things that they're doing. And so time spent actually working alongside those teams, in the trenches, trying to solve something, is really, really powerful.
One of the things that I saw earliest when I joined Whatnot that I've seen Grant, who's our founder CEO, do is he'll sit in a review and be like, "I don't think this is right." And then he'll pause and say, "I'm going to clear the rest of my day. Let's sit and figure it out." And he'll actually end up sitting with the team and going through, you know, again, easier with AI data tools: let's literally pull up the tickets, let's literally pull up the code, let's go through the data line by line and understand what's actually happening so that we can make a decision there. And it means he's very up to date with what's going on. It very culturally sets the tone for the team that we're just seeking truth. And it makes it very much a "us" versus "you." Reviews got very into "listen for yes" for a little while there, where all you're trying to do as a PM is just get a green light so that you can go back to your engineers and say, "I have some credibility. I can get, you know, the CPO to approve what we're building," as opposed to this idea that says we're just trying to find out the right answer, and in theory everyone wants us to come up with the right answer.
So, you know, we use planning to align the company on what are the really important things we have to solve. That's mostly a resourcing discussion, right? Like, if we pick the right things and stack rank them all, ultimately I'm most accountable for making sure that we have the right resources in the right places to hit things. But I don't know if those are the right places if all I do is delegate to the team to go figure it out and I'm not actually periodically very deep with them on exactly how that works. You know, how do our referrals work? Literally, what is the logic that fraud might use in order to invalidate one? Oh, based on address signals. How do we calculate address signals? Is that a Google-normalized thing, or is that free text that's put in? And if you don't actually push yourself down to sit alongside your IC engineers and your IC designers and your IC PMs, you don't actually know that stuff. So you can't make good macro decisions without the micro.
So increasingly, I think, you know, I push myself to try and be T-shaped. I can go very, very deep when required, but I'm mostly broad across pieces. And I think the idea of just hiring people and then not asking any more questions and delegating all of the detail just isn't really the most successful model for getting the most out of an organization.
With a very product-minded founder, CPO classically is a very challenging role for people, because you're basically the person between a very opinionated founder and the team building it. What have you found works in creating a, you know, environment where you are happy in that role?
Yeah, and funnily, that's been most of my career. Actually, I worked for three founders in a row who are all very product-minded. And at Whatnot I've got two, which is a blessing, actually. In general, I try not to double up. If, you know, Grant or Logan, our founders, are on a thing, it probably doesn't need me. Like, what's the advantage of an extra layer? You know, I think jokingly one of the PMs on the team has referred to it as the two dads problem, where you've just got two people dishing out conflicting instructions, or somebody wants to review, and then you do all this work to present it to me, and then you go back and it gets a different thing. So my first thing is, if Grant or Logan are on it, I check that they're watching it, they're accountable for it, and I step out. So there'll be long periods of time where fully half my team could be working on something and I couldn't tell you day-to-day where it is, because, you know, it's with Grant, it's with Logan, and that's totally fine. I don't have to be across all of the things they're doing. We just need to make sure that there is somebody doing that bar raising.
So we spend a lot of time doing that alignment. I think the second one is, ultimately, if you are a product leader in a founder-led company, you have to understand it's not your company, it's theirs, and you just find the right balance of, you know, "Hey, are you open to feedback on this? Have you made up your mind? Are you open to a push on this?" And you just find that rhythm working with folks. Generally speaking, you know, there's a reason that founder-led companies do so well in our industry. The insight required and the customer intuition to make the thing in the first place and make it work tends to be really important. And then I just view my job as, you know, making sure that we've got coverage on the places where our founders aren't.
Awesome. So a couple things you've learned here for building, and I think in consumer this is especially important, is, counterintuitively to how maybe people think things should work, you're finding that the best teams, companies, products end up coming from top-down, founder-led, almost, you know, micromanagement is a dirty word to a lot of people, but it's basically being in the weeds, in spite of how people may feel, actually ends up leading to better stuff.
Top-down works well if leadership is good enough to be in the weeds and be specifically correct. I think where it falls apart is where you don't actually know ground truth and then you attempt to manage people from above, and that's where I think the term micromanagement comes from. Otherwise, if you're working from the same data and you have it, I don't know a junior engineer or an entry-level designer who isn't stoked to sit there and work alongside the CPO or the CEO and ship something, because you're just unblocked. There's no alignment meetings, there's nothing to do. They also find that you can give way better feedback if you're literally in the detail. I think the nuance is just how do you make sure you can do that in enough places? There's never been a better time to attempt to be in the detail on things, because you can literally query it in real time.
I think that's such an important nuance here. The story told is so powerful, this idea of Grant just, okay, I'm going
to clear my day, I'm going to spend time going deep on this stuff. That feels like a very necessary ingredient for someone at the top to be making right or wrong decisions because, to your point, if they don't have all the details, they're not making decisions out of real data.
And we can do that because, as I said, we plan what we're doing and then we allocate and then we're dividing and conquering. So he's probably trying to nail three or four most important things at a time and, you know, he's the CEO, he's going to call the ball and say these are the four things I own right now, and I'm going to be like, "Great. I'll be over here then." And then within those, what am I doing with the rest of my day that is more important than nailing the five things I've said we'll get done this half? If it's anything other than maybe hiring, standing meetings, any of that stuff, you can just clear it and set the tone for the team that until we understand it, we can't do anything else.
Say somebody is looking for a CPO role or a first PM role, kind of similar, basically working for a founder. What would be your advice for them to land in a place where they're happy and not just super frustrated by this kind of middle layer where they just don't actually have any agency?
Yeah, I think the first thing you've got to work out is why do you want it? Right? There's this idea that in the CPO's job you get to decide all of the road maps and all these things, and I've got bad news for you: that's not strictly true. But I think it also comes down to spending some time with that person to work out how do we jam on a topic? How do they like to get pushed? How do they not? And then you spend a lot of your time calibrating.
So before I joined Whatnot, I think Grant and I had five or six different coffees where we talked about how do you get, you know, this type of team to move? How do you work on those things? And then I obviously went through a series of interviews and actually came down to LA, where Grant and Logan were based at the time, and spent a whole day with him. Just in a room going through a couple different problems, talking through different things in the road map, just really getting into it. And I tried through that to be my most ordinary self as opposed to my interview self, if that makes sense. Because you've got to ask yourself, do I really want to spend my whole time having this discussion and this debate?
But I like being a CPO or any kind of product lead because in a lot of ways what I'm doing is helping translate that vision and that intuition into reality. And then, you know, there is an art to learning how to push somebody without competing with them. And I think a lot of the times I've seen the CPO-CEO relationship go badly, it ends up with the CPO competing with the CEO for vision, and they end up at loggerheads. And I think, you know, that's not your job.
I want to ask more about that, this art of pushing back and, you know, nudging things in a direction. Is there one trick or one tip you might share with folks to get good at this? Because a lot of people deal with execs and they're always trying to get them to agree to what they want.
I think the first thing is don't treat it as a trick. You're not trying to get an answer and a yes. You're trying to seek truth, is my first statement. And so I wouldn't say it's especially common that we start out of alignment, but I do think, you know, part of your role, if you're one of the more senior product folks in the room and the CEO, or if you're a director and it's the CTO in the room, is pushing the team in a way that you don't really expect or you don't really understand, is to start from a place of curiosity. Does that person have more context than you, or is there a thing that you're not aware of that is guiding it? So I often try and start with, am I hearing you right that this is your prior? Is there a piece of context or something that I don't have that helps inform that prior? Cool. Make sure everybody's on that same baseline. And, you know, I coach my PMs to do this with me all the time too. If I'm coming at you from an angle you don't expect, pause and make sure that you understand why.
And then after that, I try and do a couple different things. People are people, right? They have good days. They have bad days. Sometimes it can be as simple as, is your mind made up on this or are you open to input? If the answer is, no, I'm pretty set that this is the answer, shut up. Don't pick a fight for the purposes of picking a fight. If the answer is, actually yes, you're welcome to push me, but I'd need to see data. If I don't have any data in the room, if it's just my opinion, then the terms of the debate are pretty clear. If I have fresh data, bring it. If I really believe a thing and I don't have data, question why, or go get it and then go back to them with the data. But if nobody's got data, it's just two opinions, the CEO's opinion is going to win. And that's okay. Check your ego at the door, get to the answer, but if you've got data, bring it.
And then sometimes it's just make sure that you're debating for the right reasons. I think it's, you know, easy in the PM theater that you're describing to want to make sure that you set the framing and the tone. And there are genuinely times in our industry where nomenclature matters and exactly what word is used can be really important. But a lot of the time it doesn't.
There's a concept that I heard come up a bunch when I talked to people who work with you. The phrase was "play the accordion." Explain.
So, I mentioned earlier I'm not huge on frameworks, but I do think there's a mental model which is quite useful that goes to the thing that you and I discussed earlier on: think through a problem, do the mental work, even if you're not shipping at all. Obviously the magic of code is that you can ship things and iterate really quickly and get data as you go. And I think one of the trappings that comes with that is you're just iterating forward and throwing spaghetti at the wall and you don't really know what direction you're going in. I think there's a second failure mode, which is people sit down and write out these long road maps and strategic vision docs of what we'll do over the next two, three years, which loses the comparative advantage we have over every other industry, which is learning. A/B tests mean that you can update your understanding of a problem and therefore change your direction as you go.
So the point of the analogy of play the accordion is, if you think about a piano accordion, before you can play a note you've got to stretch it all the way out, bring the air in. So, what is it we're trying to get done here? But you don't make music until you press the key and push it all the way back into V1. And then when you go to play the next progression, you pull it all the way out again. Okay, given what we just learned, what do we do? And then you go and build the next thing again. And you've got to get used to this motion that says this isn't creating value. Pulling it out doesn't actually do anything. All the value is created here, but unless you are going through this motion constantly of saying we'll re-evaluate what we understand, how does this change what we're doing, you're probably not playing the right thing.
And I know I'm probably bastardizing it; somewhere in the comments there's going to be a musician who points out to me that this is actually also one of the ways you make music. But I've found it just a generally very valuable memory device for PMs: you've got to constantly be zooming out and pushing back in, zooming out and pushing back in.
Yeah, that is such a fun analogy because, you know, even though you make music expanding it, it's like inward, or it's like learning from the
Yeah, the compression. What are we learning? How do you communicate differently? Back to shipping.
A different kind of music. Internal music, external experiment.
Yeah, because genuinely there are people who are really big on road maps and there are people who are really big on iteration. I think the honest answer is you've got to do both. You can't over-index on either.
So the lesson here is run experiments, but then make sure you think about the implication of the result at the bigger picture, and then
What are we learning? Have a plan. Like, the system works this way. I have this belief that if we change the way that discovery algorithms work, for XYZ reasons it will have this impact. What's the smallest thing I can do to prove that? Okay, it worked. Great, no change to plan. V2. Oh, that didn't work. V3 is going to have to be different. And you just want to always be going through that motion.
For people not watching on YouTube, Tom is moving his hands.
Just gesticulating wildly.
What's a context where you had said that to someone, "Hey, go play the accordion," or "We're playing the accordion"? What are they usually doing wrong?
It tends to be that you're shipping something in a very local sense without necessarily understanding impact. So an example right now is, you know, the core of most marketplaces is listings, right? If you try and think about amazon.com without listings, there's basically nothing there. It's a left rail or right rail and some videos.
In video commerce and live commerce, in Whatnot, you don't really historically need listings. If I want to sell you a pair of AirPods, I could literally hold them up to the screen and show you them and describe them and say they're AirPods, I'm going to start them at a dollar. And as a buyer, you now have all the information you need in order to make a purchase decision, which is great.
And so you may not have to invest in making listings the way that someone else does, and that's probably net good for a seller, because it takes three and a half minutes to make a listing and it takes zero minutes to describe a thing and hold it up.
But you then zoom out and say, "Oh, new buyers are going to come to live commerce and expect search to function." If I don't know what you're selling until after you've sold it, there's no possible way that I can put somebody in the right stream for AirPods, because we didn't know you had them, and I therefore can't direct people to it. And so you go through this exercise of, we don't need to fix it, it works. And then you're like, okay, great, what implications does that have for other things? Okay, then what would we do? Oh, we're going to make everyone make listings. And you're like, zoom back out again. If every seller is now spending three minutes for every listing that they're making, the number of things they can sell per hour is now dramatically, dramatically lower, which means it's bad for sellers. Okay, we can't do that either. And so you just have to keep stretching through: what are the longer-term implications, or what are the knock-on effects of things we're building?
That is extremely helpful. What's interesting about Whatnot is there's this spectrum: there's Whatnot, and then there's this whole trend of agentic commerce, where agents are going to be buying things for you and working with each other, creating a whole new agentic economy. And this is the opposite: humans live, talking to each other, buying from each other. Thoughts on that trend and how it might impact you guys?
Listen, I for one welcome our agentic overlords for things like, I don't want to have to think about light bulbs, air filters, any of the programmatic stuff that I need to run my life, right? I'm also pretty great with it for really high-intent things. I need a particular cable for my computer. I'm looking for, you know, outdoor lights for my house. Things where there are really taxing searches. I'm thinking about, you know, I've got to go to a wedding, which means I need black shoes, but delivered by Thursday, because if they don't get here before Thursday, they're no good for me. Those kinds of things, absolutely.
But most of commerce in America isn't actually high intent. How many years are we into e-commerce now? Thirty years into e-commerce, and e-commerce has never exceeded 20% of retail spend in America. Right? The vast, vast, vast majority of retail shopping in the United States is still people going in person and buying things. In the UK, it's somewhere like 75/25.
And it turns out actually that a lot of shopping is low intent. I'm going to the mall. It might be because I've got a wedding coming up and I don't have anything to wear, and I'm going to wander around and see what people have available because I don't actually know specifically what I want.
And actually the value of stores is that the person who runs that shoe store has agency and taste, and she curates a selection of shoes. She has, you know, a level of customer service. She has a window display which can show me the types of things she has, and I can wander around the mall and actually work out what it is I want to buy, or be educated by it. And as it turns out, it's quite pleasant. There's a reason that the mall is a social thing.
And so I think, you know, agentic is going to be huge. Retail is a seven-and-a-half-trillion-dollar industry in the US, though, so I don't think it's a winner-takes-all thing. What I think live commerce does is it's the first time that we've ever managed to bring together the scale and convenience of the internet and also that same social, cultural experience of physical shopping.
Because, you know, at Twitch we used to basically assume that if it's less than a thousand people on the stream, it's non-economic, because you're operating on CPMs. But you can go to a stream on Whatnot and see 30, 50 people in that stream, and imagine for a moment that you're running a shoe store at a mall and you had 50 people in your store. You'd never close. You would literally never shut down the store, because that's so much more foot traffic than you can ever imagine being in a store, because commerce has totally different economics to entertainment and CPMs aren't the thing that you have to worry about. So I think it's not really in competition with agentic. I think it's a completely different kind of customer need.
Amazing. There's room for everyone. I want to end maybe with a question about Twitter. So you were a PM at Twitter working on growth. I feel like everybody that worked at Twitter as a PM was just scarred from the experience of Twitter. Everyone's like, do not do things this way. What's something that stuck with you from that experience? What did you learn? What did you unlearn?
It is funny. I have said to a friend before that asking someone who was at Twitter in the 2015–16 era is a little bit like a therapist asking somebody to tell them about their childhood. You know the trauma that you're bringing up. Yeah. For those not familiar with that lore, I think we had nine heads of product in the two years that I was there. It was that level of chaos. I think my favorite quote was from someone on the partnership team who said it felt like chaos which had lived in the events. That was how often drama was going on about the place.
I think I probably took two overwhelming things away from my time at Twitter, other than some incredible friends, and I will say that the product diaspora from that era of Twitter is pretty incredible and everywhere.
The first one is, if you truly find product-market fit, if you manage to bottle lightning, it doesn't matter how badly you screw up the organization of it, it's huge. And like
Twitter genuinely had that level of product market fit that you could emotionally feel how much people loved your product. And it's a really powerful litmus test to learn relatively early in your career that that's what PMF is, not like "Hey, the graphs look okay." Right? There's like a level of fervor that goes in.
I think maybe on the flip side of like less positive is like what I definitely learned is like most of the time you hear it's really complex. It isn't. Leadership's just weak.
So, you know, the two years that I was there, everyone knew we were going to have to lift the 140 character limit, right? There was working group after working group, like the project beyond 140 was like everywhere, because we knew for example that like people in Japan tweeted six times more often than people in Western markets, and it was largely, when we spoke to them, because kanji let you say heaps more in the same number of characters than, you know, a Romance language did. We knew that was the inevitable end state, but there's obviously a bunch of like work that was needed and trade-offs, and just no one wanted to make the call, and so there was just like yet another design sprint, yet another cycle.
And it was another year and a half after I left. I think it was maybe even almost two years after I left before someone actually did it. And it turns out, you know, nobody died. The soul of the place didn't fall apart. I imagine the same discussion went on for editing tweets, which took another two and a half years. Like, sometimes things aren't actually that complicated. It's just weak leadership.
Yeah, it's funny how far it's come. Now you can write an entire blog post on Twitter, like articles. It's a big bet. On the product market fit piece, I think even more important there is the network effects of a Twitter, which you spent a lot of time thinking about building marketplaces. Like to me, watching Elon basically change everything. I forget who tweeted this, but just like everything changed. The brand, the name, the website,
Number of people working there, everything.
People, the team, like what was the thing that stayed the same? And it was basically the network effects of Twitter.
Yep.
It's the place to be so everyone's there and that's hard to break.
And even in spite of how much you tried to mess it all up, it's still kicking. That's it. If you've bottled lightning, genuinely you can tell.
Yeah. Okay, I'm going to take us to a recurring corner of the podcast, failure corner. Why I like doing this is because people see people like you coming on the podcast, have this illustrious career, everything's going great constantly, you're just killing it, and they don't see the things that don't work out and the times that you failed, and in their life things often go wrong. So the question for you is just, what's an example of a time in your career where things failed, something you built, some career move you made that didn't work out, and then what did you learn from that experience?
Honestly mate, and it's very nice of you to say nice things, but I feel like I fail more often than I succeed across the course of my career. Genuinely, to the blog you referenced right at the start, I published in the back of that the actual document we use internally to talk about how we build, and it starts with like batting 500 is like the goal. So like you're hoping to be right as often as you're wrong.
So there's probably just too many specific examples of times I've screwed up in my career, but there probably is a really common thread to it, and I think it's probably an easy trap for any PM to fall into, which is like averages mean nothing to the individual is probably the thing that I've like really scarred by.
In any sizable population it's really attractive to go and look at like average utility or average adoption of something, and then you find that like, you know, only 3% of people use something. And you're like, cool, we can probably get rid of that feature, it's not used widely. But if you don't go a layer deeper and be like, actually, you know, it's only 3% of something, but there's a group of people for whom it's 100% of what they do. This is their core use case, and for expediency's sake, because somebody doesn't want to maintain a feature anymore, you're just going to deprecate it.
And then it turns out you like blow up the use case of that group of humans, and then to your last point about network effects, the ongoing spiral effect of that can be enormous. You know, I think about it a lot in e-commerce of like, this is somebody's business. Right? If we're just not reliable or like deprecating a feature, it's kind of like a Westfield Mall just turning off the power in the lead-up to Christmas without thinking about it. And so like there are real downstream impacts to people's businesses that often come from just like a lack of nuance in understanding metrics, particularly averages. Like they just lie to you all the time, and I think, you know, I've probably screwed up in all of the ways in my career, but most of the time I've made genuinely like I'm-disappointed-in-myself levels of decisions, it's typically that I've relied on averages without thinking about the individual use cases that are hidden underneath.
Makes me think about, Jeff Bezos has a quote, "When you have data and an anecdote, trust the anecdote."
Yep. That's exactly right.
Well, Tom, we've gone through everything I wanted to talk about. Is there anything else that you wanted to share, anything you want to leave listeners with before we get to our very exciting lightning round?
I think all I'd add, mate, I've really enjoyed the discussion, so thank you, is like I don't think there is one way to do product management and I don't think there is one way that AI will shape the industry. So like we're pretty confident that for Whatnot, the product we're building and the culture of the company that we have, that, you know, this model of like fewer PMs who are more senior, a lot more autonomy, is right for us. I don't pretend to, you know, presume that that will be true for the entire industry, but I do think there has never been a better time to go back to the roots of like the actual product work and getting out of that theater, and I think that probably is true everywhere.
Even if you are still, you know, there are people who are wonderful people managers and really derive their satisfaction from doing that and growing and coaching, and I'm sure there'll be loads of places where that's still valuable. So assume that at least half of what I've said is wrong, on the same basis that half of the things I've probably ever shipped are not correct.
That's why I love these conversations and why I think this work that we do together is important, is we are living through the wildest time in our careers. So much is changing. So much is being rethought, as you've described, and it feels like the only way we can make our way through this successfully is to learn from how other people are approaching it, see what they've learned, see what has not worked.
Take bits that resonate, ignore the bits that seem bombastic or not applicable.
Exactly, cuz like no one knows exactly where... Elizabeth Stone had this great way of putting it. There's a storming phase and a norming phase, and we're in the storming phase of holy... Like I remember in my PM career, as I started writing and stuff, everyone was always asking me, "How has product management changed in the last decade?" I'm like, "It hasn't changed. It's the same basically." But it feels like now it actually has significantly changed.
Although to your point earlier, actually the core thing will be the same.
Yeah.
Yeah.
Yeah. So that's why these are so useful, just for people to see. Here's how a team is operating and what they've learned, and here's things to try, and it may not work for you. But this is how we learn from each other.
With that, we have reached a very exciting lightning round. I've got five questions for you. All right, here we go. What are two or three books that you find yourself recommending most to other people?
Okay, not to be clichéd, but The Hard Thing About Hard Things I still think is the best book written about product management. Just from a like breadth of things that you have to go through and try and do and screw up a lot.
Slightly left field, but the best book anyone's ever recommended to me to read, which I now recommend to others, is The Purpose Driven Church by a pastor called Rick Warren. Emmett Shear, the CEO at Twitch, used to basically make sure that people read it. It's this wild examination of why people emotionally invest in something and how to engineer emotional investment into it from a group of people. It's literally a like how-to-go-and-build-a-church guidebook written in the '90s. Definitely worth a read if you're in a community product of any form.
And the last one is, if you're looking for kind of like fiction or fantasy, which is where I tend to go in the evening, Babel by R.F. Kuang.
11 books have never been mentioned before get added to the canon of recommended books.
It's very fun.
Favorite recent movie or TV show that you've really enjoyed?
Star City on Apple TV. If you liked For All Mankind, it's kind of like the flip of that, but it's the Soviet side. It's like watching the Americans and For All Mankind mixed together.
Favorite product that you have recently discovered that you really like?
This is a very deep cut, so I will apologize to most listeners. Hopefully you've detected the accent. I'm told constantly that my Australian accent is going, but one of the things that I love is every time I go home, I realize actually a bunch of the government services apps in Australia have become phenomenal. You ever had that kind of concept where you're like, I wish the government has all my data, but I wish there was just like one place where I could with one click get my driver's license renewed, I could like transfer titles, and do all of the admin that slows you down in life? Service New South Wales actually nailed it. Bizarre to me that I would ever come on a podcast and say, actually a government-run app in Australia, of all places, is it. But I was home recently and had to do all of my life admin, and it's incredible.
Wow. Something I heard recently about Australia, while we're on that topic, real quick tangent from lightning round, is with a solar panel buildout that has happened there, there's more electricity available in Australia than they can use.
Yep. Solar is real huge in Australia.
So it's like they're giving people free electricity in the middle of the day cuz there's so much available, and they're like, use all your stuff in the middle of the day cuz otherwise it's going to go to waste.
There is a pitch somewhere that says if AI actually needs loads of electricity in order to feed data centers, Australia's entire 21st century economy should be powered.
Incredible. That's just like such good news that we are finding ways to generate so much energy from solar panels.
A small piece of regulation, what, 20 years ago that said if you're building a new property you got to put solar panels on the roof, and it turns out it works great.
Oh my god. I love this. I love this optimism of the future because, you know, climate energy...
Technology is the way through. Not the problem.
Mhm. Hear, hear. Okay, two more questions. Do you have a favorite life motto that you often come back to in work or in life?
In my uni days, or back in college for American translation, I used to have a party trick where I'd memorized "If—" by Rudyard Kipling, cuz I thought it was like deep and really meaningful. But I think if I'm being really honest, "I'll figure it out" is probably the closest. You know, like it turns out most things are not as hard as people think, like we'll work it out. I'll figure it out. If you're willing to devote the required time, money, effort, energy, you can solve almost anything. And if you're not, then it's probably not that big of a problem.
Final question. I imagine people ask you this a lot, but I'm also just curious. What's something you bought on Whatnot in the past month or so that was just awesome, delightful, surprising?
The most fun thing I bought recently, I'm not joking, is a live lobster. So...
Was that a hook?
We've recently... one of the things that's really taken off on Whatnot is our like fresh and specialty foods kind of category, and so there's this wonderful seller who goes by e-fish-co who has a seafood store down on the dock in San Diego, and every morning as the boats come in he literally goes out and live streams all of the crates of seafood coming in off the boats, and then he auctions them off live on Whatnot and ships next day to your door. So I got a California spiny tail lobster shipped direct to my door courtesy of a live stream.
And how's this work? They put it in ice, ship it next day, same day?
An ice kind of like container, and it's UPS overnight on my door the next day. And that is why I think e-commerce is going to be great but not all-encompassing, because I had no intention of buying a spiny California lobster that morning.
And my free agent decides he needs a lobster today and we...
I got to be honest with you, it was delicious.
Amazing. I didn't know that you could buy stuff like that. Tom, where can folks find you online if they want to follow your writing / hiring? Talk about what you're hiring for, and finally, how can listeners be useful to you?
If you're looking for me online, I'm tdrobbo, T-D-R-O-B-B-O, on Twitter, or X, I guess, which is where I do most of my musing and some mix of product management and yelling about Warriors games, so apologies in advance. And then LinkedIn is the other place I've got most of my kind of like work writing. I will say I try for quality over volume, so don't expect daily drops from me on either.
Just bang those. Just bang those once a year.
That was the nicest thing anyone said about me in ages. Just dropping casual bangers.
And how can folks be helpful? Honestly, would always welcome feedback around Whatnot and how people are finding it and what more we can do better. So like hit me up on either of those platforms with, you know, hot takes, feedback or thoughts.
And then you're hiring PMs, even in spite of the many strong opinions about product management. Maybe just talk about that and where folks can apply.
Sure. Listen, we are constantly looking for PMs. The intent of kind of like explaining how many people applied versus hired was not to discourage people. It was more along the lines of like the proliferation of product management doesn't itself naturally lend to the people with the skill set that you and I spent the better part of 90 minutes talking about right now. But, you know, we hired two people yesterday. I'm very excited for them to come start, and we've always got a variety of PM roles open.
Whatnot, if you just kind of Google Whatnot jobs, the right kind of application process will pop up there. Or I've found that, you know, the valley is small enough and product management is not that kind of obscure that you can probably find one of the 20 or so humans who work at Whatnot and reach out to them. But we're looking at a variety of things right now. Payments very high on our list of what we're doing, as well as logistics. So if anyone is really excited to come work on the future of shipping lobsters overnight, it turns out there's quite a lot of nuanced product that has to happen there.
Amazing. Tom, thank you so much for being here.
It was a pleasure, mate. Thanks for having me.
Bye, everyone.
Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review, as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com. See you in the next episode.
Article published · Updated