Every time a seasoned practitioner shares concrete tactics and heuristics, I listen. Dan Hockenmaier has worked on multiple successful marketplaces from Faire, to Thumbtack and have also contributed to Reforge materials, and has thus battle-won insights to share. He has also published foundational frameworks such as Racecar Growth Model and how to focus on customer acquisition channels. This is my attempt to solidify my understanding.
Building a Growth Model
- Half the value is in building it, gives you a conceptual understanding of how the entire business fits together
- It makes business concepts concrete and hard to fake
- It's not meant to be precise, but directionally accurate
- It is not a forecasting tool, but a prioritisation and bet-sizing tool
- Simplest is SaaS businesses, start from acquisition channels, retention, monetisation
- For transactional businesses, stack on transactions per month, AOV, unit economics, COGS
- Marketplace is on top of that, modelling demand side and how that affects supply side; this is where it gets non-linear
- Payback period is a better measure of paid marketing performance than LTV:CAC because the speed at which we get back the money to acquire a new customer has more bearing on the eventual growth outcomes than raw multiples (Personal note: can be easily modelled)
- Marketplaces are sensitive to a lot of assumptions, and there can be false precision, so just use a high-level model to get a feel for the levers, and have each product pod, GTM-team build their own mini-model, which you then stitch back up to a master model
- There are 2 forms of pods with their own models, the 1st is growth teams with simple funnel analytics, the 2nd is teams that have to manage inherent tensions, like marketplace quality teams
- For managing allocation of manpower across pods, you need a common currency in the growth model every project can be weighed against
Un-intuitive Insights from Growth Models
- Growth is much more sensitive to retention that most of us intuit because there's a lot of interaction between a healthy retained customer base and everything else we care about (referral, content generation, generating con. margins)
- Small percentage moves in retention are therefore highly impactful, but they are also notoriously hard to move because it is a culmination of the whole product experience, and also emanates directly from product value which is hard to change
- Retention is rarely solved at the point of churn. It requires deeply understanding what matters to customers and what is the current customer experience. In a marketplace context, it is often depth of supply, and interaction they have with supply. So you should be working on core product levers than growth levers.
- From Dan's experience, it's easier to move retention early in the customer journey, because by the end, these users have already formed their minds. Extreme version of this is churned users, they have already decided to leave, it's hard to convince them otherwise; therefore resurrection should be a low-priority project
- Focus on inflecting the early user experience, specifically on reducing variability in user experience. eg. for Uber, they try very hard to ensure that no drivers go below a certain earning in the first few weeks. Huge competitive focus with Lyft over first-week and first-month earning.
- At Thumbtack. they cracked their primary growth channel, which is targeted SEM/SEO. Basically the whole team was optimising that. The growth model showed that repeat purchase rates was also very influential, and thus there was a misallocation of resources with respect to what grows the product. They shifted efforts from top of funnel to deeper down
Nature of marketplace businesses
- Hard to start but immensely valuable after the flywheel is built
- Perfect for VC
- Usually, CACs for businesses go up as they reach the next marginal customer, but for marketplaces, it's the inverse due to improving liquidity
- Every business question is more complex with marketplaces
- In the beginning, don't get stuck on marketplace theory, 90% of what an early-stage marketplace needs is the same as any other business, like growth, profit and retention
- Marketplaces have more 2nd, or 3rd-order effects. eg. pricing isn't just the price elasticity curve, but also how it affects the interactions between both sides of the marketplace
Health Metrics of marketplaces
- Obvious ones to start with: GMV, AOV, unit economics (most start negative),
- Liquidity (key function of a marketplace, but need to contextualise it in a dimension that your user cares about. For Uber, it's wait time. For ecommerce, it's search to fail metrics. Until you have liquid marketplaces, nothing else matters. Cut scope until you are liquid)
- Share of wallet: how much they spend on your marketplace vs alternatives. For sellers, how much biz they do on your platform vs others. It's inverse to multi-tenanting. Hard to do on both sides of the marketplace, so need to pick leverage point.
- Between growing GMV 10% by getting 10% more customers or getting 10% more of current customers' wallet share, latter is more valuable due to defensibility of that deeper relationship. Depth over breadth every time for marketplaces.
Supply or Demand?
- Both are important
- On average, founders overemphasize supply and underappreciate demand
- Supply is disproportionately important in the early days, and usually use the product more deeply - there's more product surface area
- But ultimately, demand is the only thing that matters. If you aggregate all the demand in your industry, suppliers are always gonna say yes to you, because it's incremental business and they are rational
- To get that demand though, you need to acquire supply that is hard to acquire. So the upshot is prioritise customer experience, and if it's to spend time acquiring supply so they're happy, do that. Focus on the constraint to great customer experience.
- But be calibrated in how you acquire supply, there is a threshold beyond which it doesn't impact demand. eg. for Uber, beyond a threshold of supply, wait times don't go down. Or on Amazon, past some amount of supply, you are not incrementally improving the buyer's experience
- Pure marketplace balance metrics aren't that practical; the thing that matters is the ROi equation for acquiring supply and demand that internalises the marketplace dynamics
- For a transaction, you need to factor in the CAC of that customer, CAC of the supply for that customer to purchase, which is based on a ratio between the 2.
- Once you have that, you can ignore balance and push acquisition all the way out to payback period you are comfortable with on either side
- 1 thing to watch out for is externalities like too little demand for supplies or vice versa such that they switch to competitors, complain on social media
- Don't just look at TAM, because most markets are big enough
- Prioritise by adjacency and accentuation of the current network effects