Dan Hockenmaier
Head of Strategy and Analytics at Faire, former Director of Growth at Thumbtack and Reforge partner
Where to hear and read about Dan Hockenmaier1 episode or test
Dan Hockenmaier is Head of Strategy and Analytics at Faire, a former Thumbtack growth leader who has built 20 to 30 growth models and worked on more marketplaces than almost anyone. The thing most worth learning from him is to build your own growth model, because it shows that retention and early customer experience drive growth, and then to treat a marketplace like a garden with a light touch.
If you think about running a marketplace, you're basically like a gardener. You have to have a very light touch.
Developing a growth model + marketplace growth strategy | Dan Hockenmaier
Dan Hockenmaier: the story and the lessons
1Dan Hockenmaier's path through marketplaces
Hockenmaier joined Thumbtack at about 30 people and helped scale it. He stayed a little over three years, and the business grew roughly tenfold in that time. He now leads strategy and analytics at Faire.
He learned marketplace growth by running it, first at Thumbtack and then through consulting and Faire.
He built growth models for dozens of startups through his consulting firm. His firm, Basis One, helped dozens of startups build growth models and growth strategies. He has built 20 to 30 growth models himself.
He understood Faire's problem because he had run an e-commerce business. He knew the pain of line sheets, PDFs and wrong pricing from dealing with a hundred suppliers.
He says he would not have seen why Convoy would work, because he knows nothing about trucking. He thinks this is why B2B marketplaces are rarer: fewer founders live the problem.
2How Hockenmaier builds and uses a growth model
He builds the model himself from first principles, never from a template. A growth model is a spreadsheet linking acquisition, retention and monetization to show where growth is sensitive. Transactional businesses add transactions per month, AOV and COGS, and marketplaces add supply acquisition and retention.
The learning is the point, so a template undercuts it. He says: I think 50 percent of the value you get from it is simply building the model. He suggests starting with a smart analyst or finance partner and iterating for quarters, so the gap between what teams ship and what the model predicts builds intuition.
He uses the model to compare opportunities, not to forecast. In annual planning an analyst runs each team's projected metric gains through the model, so gains in different metrics become comparable in one currency.
With 10 to 20 pods, planning can be zero-based. Many assumptions are in play, so the output varies a lot, which is why it suits opportunity assessment better than prediction.
He keeps one simple company model and gives each team a mini model. Each product pod or go-to-market team models the inputs to its own North Star metric. A pod's strategy doc should include that mini model.
He advises against stitching a master model for a marketplace, since stacked assumptions make it junk in, junk out. Teams managing a tension, like marketplace quality or credit underwriting, model the trade-off rather than a funnel.
Linking viral and paid loops shows why payback period beats LTV to CAC. With virality, the coefficient drives growth speed. With paid marketing, contribution margin is reinvested to buy the next customer.
The faster money comes back, the faster the next customer can be bought, so payback period predicts paid growth better.
Thumbtack's first model moved resources from top of funnel to retention. Built with the finance team, it showed extreme sensitivity to the repeat rate of new customers, which set LTV and what Thumbtack could pay to acquire them.
Almost all traffic came from targeted SEM and SEO for one specific job, so customers were hard to upsell. One team had shipped hundreds of conversion experiments, but little had gone into cross-selling customers into other categories.
3What Hockenmaier knows about retention
He says retention is won in the core product, not in emails and pushes. Retention is the sum of the whole product experience, and the biggest wins come from inflecting the early user experience. In a marketplace the key lever is often depth of supply.
At Airbnb, analysis showed that moving retention one percent would hit all the goals, but it proved very hard to do.
He looks for first-week customers having a bad experience they shouldn't. A new driver may earn little through bad luck, then conclude the platform pays three dollars an hour and never return. Uber and Lyft compete to guarantee first-week or first-month earnings.
By month 3, 6 or 12 a customer has formed a firm opinion, while early on there is still room to teach why the product is valuable.
He distrusts making other users copy what best users do. Our best users do X, so make others do X almost never works, because something unique about those customers drives both the behavior and the retention.
He has rarely managed to move users from the worse bucket into the better one.
He waits on resurrection campaigns until earlier funnel work is exhausted. A big pool of churned users makes even 1% back look huge, but they already tried the product and decided against it.
4How Hockenmaier measures a marketplace
He treats liquidity as the top metric, measured in what customers care about. Liquidity is how reliably people can do what they came to do: a four to five minute wait for a ride, or search-to-fill conversion in commerce. Until a marketplace is liquid, nothing else matters.
That is why he advises narrowing to one geography or category first. He also tracks GMV, unit economics and share of wallet. Instacart and Uber lost money on every order or ride early on, so understanding the components matters more than the starting level.
He takes depth over breadth and tracks share of wallet. Share of wallet is the share of a buyer's spend, or a seller's business, that goes through the marketplace. At Faire it is the share of a retailer's shelf that came from Faire.
A high share makes multi-homing less likely. Both sides rarely reach a high share at once, so pick one leverage point. Given 10% more customers or 10% more of current customers' wallets, he takes the wallets.
He prices supply into customer acquisition with a dual-sided ROI equation. True CAC is the rider's CAC plus the driver's CAC times the driver to rider ratio. With one driver per ten riders, that means the rider's CAC plus ten times a driver's CAC. Compare it to LTV to get payback, then push acquisition to your payback threshold.
He has grown less focused on balance metrics such as buyer to seller ratio. He still watches externalities the equation can't capture, such as disillusioned drivers who switch to Lyft and complain on social media.
5How Hockenmaier grows and expands a marketplace
He puts demand first and adds supply only as far as it helps customers. Whoever aggregates demand in an industry wins, because a restaurant, electrician or driver will always accept customers delivered at a profitable rate. Founders over-rotate on supply because it is the product early on.
Past some point extra drivers no longer cut wait times, and Amazon's pet store hits a similar limit with sellers. Settle trade-offs from the customer's perspective.
He changes core marketplace incentives lightly, like a gardener. Commission has no simple price curve: a higher commission can fund returns and faster shipping for buyers, while supply will sign up at nearly any commission if it helps them transact. A change in one place can show up two months later elsewhere.
He contrasts marketplaces as gardening with SaaS as construction, and says he learned this the hard way several times. Airbnb had roughly 100 people on pricing recommendations alone. Be most careful with incentives that are already working.
He picks expansions by adjacency and network reinforcement, not market size. Marketplaces sit in huge markets, so any of ten nearby opportunities would bend the curve if it worked. Instacart fits convenience stores better than traditional retail because frequency and speed match.
Uber Eats reuses the same drivers, and the same customer wants rides and meals.
He keeps go-to-market from running ahead of product. Rivals race to liquidity with heavy incentive spending, but the winner delivers an excellent end-to-end experience first, even to fewer customers.
A small flywheel with retaining cohorts, or a smiling curve, gives conviction to invest and is what makes VCs invest, rather than a pile of low quality GMV.
6Where Hockenmaier says marketplaces work and where they don't
He doubts most unbundling theses of horizontal marketplaces. They overweight user experience and underweight scale benefits in LTV and network effects. Thumbtack tracked hundreds of niche rivals in a spreadsheet and very few gained traction, because it could upsell customers into a thousand services.
Niches that work are high frequency or high value and self-contained: Airbnb against Craigslist, StockX and GOAT in sneakers, RigUp for blue collar work.
He checks fragmentation and ticket size before backing a B2B marketplace. Fragmentation is the share of volume done by the top 5% of suppliers. If 10 companies do 80 percent of volume, they have the leverage and pay low commission.
High dollars per transaction invite disintermediation: a ride pays two or three dollars of commission, while a materials order can be worth tens of thousands to millions. He says this is usually why a B2B marketplace fails.
He says marketplaces should add software only to solve real customer pain. First confirm the marketplace is liquid and high performing, then look for a deep problem, such as integrating with a supplier's back office or accounting. Marketplace commission is often 10 to 20 percent against about 2 to 3 percent for SaaS.
Going from marketplace to SaaS is easier than the reverse, because SaaS has no demand side relationships. The aim is a stickier product, not a few more points of margin.
He expects creativity in supply to decide which marketplaces survive. Where customers want the same experience every time, the market consolidates and leaves the marketplace model. Where they want suppliers to bring new things, as with Etsy, Amazon, Steam and Faire, the marketplace stays.
He expects ride sharing to consolidate once autonomous vehicles arrive. Opendoor buys the house, so he calls it e-commerce rather than a marketplace.
Developing a growth model + marketplace growth strategy | Dan Hockenmaier
Dan Hockenmaier explains how to build a growth model that shows which levers matter, then covers how to measure, grow and expand a marketplace. He has built 20 to 30 growth models and worked on more marketplaces than almost anyone, and his Thumbtack model shifted resources from top of funnel to retention.
Hockenmaier joined Thumbtack at about 30 people and helped scale it. He stayed a little over three years, and the business grew roughly tenfold in that time. He now leads strategy and analytics at Faire.
He learned marketplace growth by running it, first at Thumbtack and then through consulting and Faire.
He built growth models for dozens of startups through his consulting firm. His firm, Basis One, helped dozens of startups build growth models and growth strategies. He has built 20 to 30 growth models himself.
He understood Faire's problem because he had run an e-commerce business. He knew the pain of line sheets, PDFs and wrong pricing from dealing with a hundred suppliers.
He says he would not have seen why Convoy would work, because he knows nothing about trucking. He thinks this is why B2B marketplaces are rarer: fewer founders live the problem.
He builds the model himself from first principles, never from a template. A growth model is a spreadsheet linking acquisition, retention and monetization to show where growth is sensitive. Transactional businesses add transactions per month, AOV and COGS, and marketplaces add supply acquisition and retention.
The learning is the point, so a template undercuts it. He says: I think 50 percent of the value you get from it is simply building the model. He suggests starting with a smart analyst or finance partner and iterating for quarters, so the gap between what teams ship and what the model predicts builds intuition.
He uses the model to compare opportunities, not to forecast. In annual planning an analyst runs each team's projected metric gains through the model, so gains in different metrics become comparable in one currency.
With 10 to 20 pods, planning can be zero-based. Many assumptions are in play, so the output varies a lot, which is why it suits opportunity assessment better than prediction.
He keeps one simple company model and gives each team a mini model. Each product pod or go-to-market team models the inputs to its own North Star metric. A pod's strategy doc should include that mini model.
He advises against stitching a master model for a marketplace, since stacked assumptions make it junk in, junk out. Teams managing a tension, like marketplace quality or credit underwriting, model the trade-off rather than a funnel.
Linking viral and paid loops shows why payback period beats LTV to CAC. With virality, the coefficient drives growth speed. With paid marketing, contribution margin is reinvested to buy the next customer.
The faster money comes back, the faster the next customer can be bought, so payback period predicts paid growth better.
Thumbtack's first model moved resources from top of funnel to retention. Built with the finance team, it showed extreme sensitivity to the repeat rate of new customers, which set LTV and what Thumbtack could pay to acquire them.
Almost all traffic came from targeted SEM and SEO for one specific job, so customers were hard to upsell. One team had shipped hundreds of conversion experiments, but little had gone into cross-selling customers into other categories.
He says retention is won in the core product, not in emails and pushes. Retention is the sum of the whole product experience, and the biggest wins come from inflecting the early user experience. In a marketplace the key lever is often depth of supply.
At Airbnb, analysis showed that moving retention one percent would hit all the goals, but it proved very hard to do.
He looks for first-week customers having a bad experience they shouldn't. A new driver may earn little through bad luck, then conclude the platform pays three dollars an hour and never return. Uber and Lyft compete to guarantee first-week or first-month earnings.
By month 3, 6 or 12 a customer has formed a firm opinion, while early on there is still room to teach why the product is valuable.
He distrusts making other users copy what best users do. Our best users do X, so make others do X almost never works, because something unique about those customers drives both the behavior and the retention.
He has rarely managed to move users from the worse bucket into the better one.
He waits on resurrection campaigns until earlier funnel work is exhausted. A big pool of churned users makes even 1% back look huge, but they already tried the product and decided against it.
He treats liquidity as the top metric, measured in what customers care about. Liquidity is how reliably people can do what they came to do: a four to five minute wait for a ride, or search-to-fill conversion in commerce. Until a marketplace is liquid, nothing else matters.
That is why he advises narrowing to one geography or category first. He also tracks GMV, unit economics and share of wallet. Instacart and Uber lost money on every order or ride early on, so understanding the components matters more than the starting level.
He takes depth over breadth and tracks share of wallet. Share of wallet is the share of a buyer's spend, or a seller's business, that goes through the marketplace. At Faire it is the share of a retailer's shelf that came from Faire.
A high share makes multi-homing less likely. Both sides rarely reach a high share at once, so pick one leverage point. Given 10% more customers or 10% more of current customers' wallets, he takes the wallets.
He prices supply into customer acquisition with a dual-sided ROI equation. True CAC is the rider's CAC plus the driver's CAC times the driver to rider ratio. With one driver per ten riders, that means the rider's CAC plus ten times a driver's CAC. Compare it to LTV to get payback, then push acquisition to your payback threshold.
He has grown less focused on balance metrics such as buyer to seller ratio. He still watches externalities the equation can't capture, such as disillusioned drivers who switch to Lyft and complain on social media.
He puts demand first and adds supply only as far as it helps customers. Whoever aggregates demand in an industry wins, because a restaurant, electrician or driver will always accept customers delivered at a profitable rate. Founders over-rotate on supply because it is the product early on.
Past some point extra drivers no longer cut wait times, and Amazon's pet store hits a similar limit with sellers. Settle trade-offs from the customer's perspective.
He changes core marketplace incentives lightly, like a gardener. Commission has no simple price curve: a higher commission can fund returns and faster shipping for buyers, while supply will sign up at nearly any commission if it helps them transact. A change in one place can show up two months later elsewhere.
He contrasts marketplaces as gardening with SaaS as construction, and says he learned this the hard way several times. Airbnb had roughly 100 people on pricing recommendations alone. Be most careful with incentives that are already working.
He picks expansions by adjacency and network reinforcement, not market size. Marketplaces sit in huge markets, so any of ten nearby opportunities would bend the curve if it worked. Instacart fits convenience stores better than traditional retail because frequency and speed match.
Uber Eats reuses the same drivers, and the same customer wants rides and meals.
He keeps go-to-market from running ahead of product. Rivals race to liquidity with heavy incentive spending, but the winner delivers an excellent end-to-end experience first, even to fewer customers.
A small flywheel with retaining cohorts, or a smiling curve, gives conviction to invest and is what makes VCs invest, rather than a pile of low quality GMV.
He doubts most unbundling theses of horizontal marketplaces. They overweight user experience and underweight scale benefits in LTV and network effects. Thumbtack tracked hundreds of niche rivals in a spreadsheet and very few gained traction, because it could upsell customers into a thousand services.
Niches that work are high frequency or high value and self-contained: Airbnb against Craigslist, StockX and GOAT in sneakers, RigUp for blue collar work.
He checks fragmentation and ticket size before backing a B2B marketplace. Fragmentation is the share of volume done by the top 5% of suppliers. If 10 companies do 80 percent of volume, they have the leverage and pay low commission.
High dollars per transaction invite disintermediation: a ride pays two or three dollars of commission, while a materials order can be worth tens of thousands to millions. He says this is usually why a B2B marketplace fails.
He says marketplaces should add software only to solve real customer pain. First confirm the marketplace is liquid and high performing, then look for a deep problem, such as integrating with a supplier's back office or accounting. Marketplace commission is often 10 to 20 percent against about 2 to 3 percent for SaaS.
Going from marketplace to SaaS is easier than the reverse, because SaaS has no demand side relationships. The aim is a stickier product, not a few more points of margin.
He expects creativity in supply to decide which marketplaces survive. Where customers want the same experience every time, the market consolidates and leaves the marketplace model. Where they want suppliers to bring new things, as with Etsy, Amazon, Steam and Faire, the marketplace stays.
He expects ride sharing to consolidate once autonomous vehicles arrive. Opendoor buys the house, so he calls it e-commerce rather than a marketplace.