Developing a growth model + marketplace growth strategy | Dan Hockenmaier (Faire, Thumbtack, Reforge)

Dan Hockenmaier (00:00:00):
One lesson I learned the hard way a bunch of times on this is that if you think about running a marketplace, you're basically like a gardener. You have to have a very light touch. If you're building a SaaS business, you're a construction worker, you're building the product and the features and selling it, and it's this very linear thing. For a marketplace, you're like messing with this ecosystem that you don't actually really understand how it works. And sometimes you might do something over here which drives this long-term effect two months later, and then you're going to be pulling your hair out two months later trying to figure out what you did over here that made that thing happen. So I think that the main advice is like to tread lightly. When you're messing with the core incentives or mechanisms of a marketplace, be very careful, particularly if you've got something that's working on playing with those variables.

Dan Hockenmaier (00:00:45):
Welcome to Lenny's podcast. I'm Lenny and my goal here is to help you get better at the craft of building and growing products. I interview world class product leaders and growth experts to learn from their hard won experiences building and scaling today's most successful companies. Today my guest is Dan Hockenmaier. I venture to say that Dan has worked on more marketplace startups than anyone else in the world, including helping scale Thumbtack in the early days, currently working at Faire where he is head of strategy and analytics and, through his consulting business, Faces One, where he's helped dozens of startups figure out their growth models and growth strategies.

Dan Hockenmaier (00:01:20):
Dan hasn't shared a ton of his insights and experiences publicly, so I was really excited to chat with him and to dig into all the things that go into building a marketplace business along with coming up with your growth model. This episode gets very deep into the weeds and so, if you're working on a growth strategy or you're building a marketplace business, this episode is for you. So with that, I bring you Dan Hockenmaier.

Dan Hockenmaier (00:01:45):
I'm excited to chat with my friend, John Cutler, from Podcast Sponsor Amplitude. Hey John.

John Cutler (00:01:49):
Hey Lenny. Excited to be here.

Lenny (00:01:51):
John, give us a behind the scenes at Amplitude. When most people think of Amplitude, they think of product analytics, but now you're getting into experimentation and even just launched a CDP. What's the thought process there?

John Cutler (00:02:02):
Well, we've always thought of Amplitude as being about supporting the full product loop, think collect data, inform [inaudible 00:02:08] ship experiments and learn. That's the heart of growth to us. So the big aha was seen how many customers were using Amplitude to analyze experiments, use segments for outreach and send data to other destinations. Experiment and CDP came out of listening to and observing our customers.

Lenny (00:02:23):
Supporting growth and learning has always been Amplitude's core focus, right?

John Cutler (00:02:27):
Yeah. So Amplitude tries to meet customers where they are. We just launched starter templates and have a great scholarship program for startups. There's never been a more important time for growth.

Lenny (00:02:36):
Absolutely agree. Thanks for joining us, John, and head to amplitude.com to get started.

Lenny (00:02:43):
Hey, Ashley, head of marketing and flat file. How many B2B SaaS companies would you estimate need to import CSV files from their customers?

Ashley (00:02:51):
At least 40%.

Lenny (00:02:52):
And how many of them screw that up and what happens when they do?

Ashley (00:02:56):
Well, based on our data, about a third of people will consider switching to another company after just one bad experience during onboarding. So if your CSV importer doesn't work right, which is super common considering customer files are chock full of unexpected data and formatting, they'll leave.

Lenny (00:03:15):
I am 0% surprised to hear that. I've consistently seen that improving onboarding is one of the highest leverage opportunities for both signup conversion and increasing long-term retention. Getting people to your aha moment more quickly and reliably is so incredibly important.

Ashley (00:03:30):
Totally. It's incredible to see how our customers like Square, Spotify and Zuora are able to grow their businesses on top of flat file, this because flawless data onboarding acts like a catalyst to get them and their customers where they need to go faster.

Lenny (00:03:47):
If you'd like to learn more or get started, check out Flatfile at flatfile.com/lenny.

Lenny (00:03:56):
Dan Hockenmaier, welcome to the podcast.

Dan Hockenmaier (00:03:59):
It's great to be here.

Lenny (00:04:01):
It's great to have you here. So we are actual real life friends and we've collaborated on a number of writing projects, including the race car growth framework and a whole thing on consumer growth strategy. But there's a couple topics that we've never actually dug deep into and that you haven't written about. So I'm really excited to dig into two specific things in our chat today.

Lenny (00:04:23):
One is growth models and two is just marketplace growth strategy and all things around marketplace growth. But before we get into all that, can you just give us a 55 second background on all of the wonderful things that you've done in your career?

Dan Hockenmaier (00:04:36):
Yeah, absolutely. So I feel very fortunate being able to work on a bunch of consumer and marketplace businesses for a long time. I started in consulting at BCG and then in private equity. I don't think I learned much about how to actually run a business at those places, but I did learn a lot about how to think about them. I think my real education on this began at Thumbtack. So I joined when there were about 30 people and we just had to figure it out. I was there for a little over three years and we more than 10Xed the business in that time.

Dan Hockenmaier (00:05:02):
From there I built a strategy consulting firm where we worked with a bunch of the top gross stage marketplaces on a range of topics, and ultimately that firm was acquired by Faire where I am today. So I lead the strategy analytics team at Faire. I think this is probably the most fun I've ever had in my career. It's just the incredible mix of the team. The business is really strong and I love the customers we're serving. So it's a marketplace connecting local retailers and independent brands, and I just think that's a really fun group of customers to build for.

Lenn (00:05:31):
Awesome. And we're going to chat a bit about the stuff that you do there and marketplaces. But before we get into marketplaces, I wanted to start with our first topic, which is around growth models. So people may have heard this term, this general idea. Just to set the stage, can you just describe what is a growth model and why is it useful to think about your business through the lens of a growth model?

Dan Hockenmaier (00:05:54):
So I think it's useful in many contexts. If I apply it to the current work that I'm doing, the strategy analytics team at Faire does a bunch of work to help our other teams make better decisions. So we're typically diving really deep into a bunch of topics across the business. I think it's really easy to make that work kind of go off the rails or go too deep unless you have a conceptual understanding of how the whole business that it comes together. And that's how I think about a growth model, so the analytical representation of how the business grows and it's typically built in a spreadsheet which has a really nice feature of being very hard to fake. You can talk about a business conceptually, but when you actually have to get it to line up and link in a model, it's very hard to not force yourself to understand how the business works. So I think it's very valuable for that.

Dan Hockenmaier (00:06:37):
I think 50% of the value you get from it is simply building the model. It forces you to understand it and then you get this artifact which you can use to understand how to weigh different opportunities or understand the benefit of working on different things. I think importantly it is not a forecasting tool, so it's not going to replace what your finance team is building to project the business. In fact, often the output can be highly variable because you're playing with lots of these assumptions, but it is a great tool for kind of opportunity assessment for the business.

Lenny (00:07:05):
Awesome. So maybe a simple way to think about it, just summarizing, is it's essentially a formula for your business that often lives in Excel that kind of summarizes and puts together all the things that can drive your business. Is that a simple way [inaudible 00:07:19]?

Dan Hockenmaier (00:07:18):
Yeah. That's exactly right.

Lenny (00:07:20):
Sweet. So we're going to go through examples of these growth models and how you've thought about this in actual potential formulas for companies. But first, just broadly, how would someone approach figuring this out for themselves? How would you build a growth model for your own business?

Dan Hockenmaier (00:07:35):
So if you think about some of the basic building blocks and probably the simplest use case would be a SaaS business, there's really three components that you need to build for that. One is to understand your acquisition channels. Are you looking at paid marketing or sales or viral kind of customer referrals? And for each of those you have some different assumptions around traffic or spend or conversion rate, those kind of things. So that's the first section. Second would be retention. So at what rate are these customers activating? And then have some kind of basic monthly retention curve. So how long are they staying around? What's the survival rate in each of these? And those kind of stack over time. And then you have monetization. So in the simplest example, they might be paying you some monthly or annual fee and so that translates in monetization.

Dan Hockenmaier (00:08:13):
So actually if you're modeling a relatively simple SaaS business, those are the only core building blocks you need and you add a lot more complexity on it based on your kind of individual business that you're building. If you then are trying to build this for more of a transactional business, you need to layer on the way that your current kind of retained customers start to transact. So how many transactions per month? What's the AOV? And then you're also going to typically need to build unit economics because those businesses often have higher costs. And so you're going to be thinking about cogs or other major costs that you build into it.

Dan Hockenmaier (00:08:39):
And then one click beyond that would be to build this for marketplace business. So now what we've talked about is mostly modeling like the demand side of a business. So now you'd also need to think about supply acquisition and retention and how these two sides interact. So as we add supply, what's going to happen to demand? But those are the basic pieces. I think you can start pretty linear like acquiring customers, they're activating, retaining and generating contribution margin typically would be the output. But where it gets really interesting is we start making it non-linear. So the most basic example of this would be virality. Your existing customers are referring new customers and those go on to refer new customers. Based on that coefficient, it has a lot to do with how fast your business grows.

Dan Hockenmaier (00:09:17):
And similarly with paid marketing, as you generate contribution margin, you can reinvest that and grow and actually if you link those two up explicitly, it makes it really clear why thinking about something like payback period is a much better measure of paid marketing performance than LTV to CAC because the speed at which you get enough money back to then go acquire another customer has much more bearing on how fast your business can grow than just the raw kind of [inaudible 00:09:39]. This is where it gets really interesting where you get some variables to play with.

Lenny (00:09:43):
Okay. So let's unpack a bit of the stuff you just talked about. There's a whole decade of knowledge that you just I think collapsed into a couple of minutes. So just to spend a little more time there. So the core three kind of variables to a SaaS business, if you're thinking about the growth model of a SaaS business, for example, you said acquisition channels, where's traffic coming from, and that's essentially how much traffic are you getting, how is it converting and things like that.

Dan Hockenmaier (00:10:07):
Exactly.

Lenny (00:10:07):
And then there's retention and then there's monetization and multiplying all that together you end up with here's how much revenue you're making as a business. Is that roughly how to think about it?

Dan Hockenmaier (00:10:14):
Yeah, it's roughly right. And those three building blocks are true for here in most businesses. Most of what we're talking about is then nuance on top of that that makes the business unique.

Lenny (00:10:22):
Cool. So if someone was just starting the spreadsheet for their SaaS business, and we'll talk about even more examples, but if someone was like, "Here, I'm going to try to figure out my growth model," it's create a row for acquisition channels and traffic you're getting, then a row for roughly your retention rate and then how much you're making per customer. Very high level, is that to think about this?

Dan Hockenmaier (00:10:42):
Yeah, exactly.

Lenny (00:10:43):
Okay, sweet. And then marketplaces, I think you said that it's transaction over time and average order value. So it's like how many people are buying stuff, how much are they paying each time? And then unit economics per transaction, how much they're making profit per [inaudible 00:10:58].

Dan Hockenmaier (00:10:57):
Yeah, exactly. And the critical distinction there is for a SaaS business, your marginal costs are often very low. And so it's not an important part to model. But for transactional business, you typically have very high cost or something else that you need to model so [inaudible 00:11:09].

Lenny (00:11:08):
Right, because you're just taking a cut and you're not selling software. And then the acquisition channels, retention, monetization, marketplace have those three things just broadly and then plus these additional two elements.

Dan Hockenmaier (00:11:19):
Yeah. I think one warning I would give ... so I built these for many businesses. I built them for Thumbtack, at Faire. Many of the companies I worked with, at my consulting firm, they break down in a few places. As soon as you start stacking assumptions, you're highly sensitive to how many assumptions do you have to make and do you know how to make that assumption. And marketplaces kind of create complexity on both of those because for the first piece you're modeling both sides of the business. There's a lot of assumptions. And second, there's a few pieces which are very hard to understand how they work. So the interaction between supply and demand is a big one. Take Amazon as an example. There's a category manager at Amazon who's running their pets business and he decides that he wants to go add a bunch more pet supply to the business. So he's going to go find pet food brands, pet toy brands. Those businesses will generate a bunch of revenue on the Amazon marketplace.

Dan Hockenmaier (00:12:03):
But how much of that was actually incremental? Maybe there's a bunch of existing pet supply that customers would've bought anyway. And does having a bunch more supply create a happier customer who retains longer and drives those cohorts up? These things are very hard to decide, something we've spent a lot of time thinking about in various marketplace businesses. And so if you're not careful, you can have a junk in, junk out problem with marketplaces. So the thing that I have gravitated to more as I've done more of this is one very basic high level conceptual model, so that's like the building blocks you talked about, which as simply as possible describes how the whole system works. And so you start to get a feel for which levers are important.

Dan Hockenmaier (00:12:39):
And then for each area of the business, so each product pod, each go to market team, they should have their own kind of mini model which describes the piece of the business they're working on. So that team typically has a north star and they should know what are all the inputs that drive that and a little model to articulate how that works. And you'll never kind of stitch this up into one master thing. I think that's a very difficult task. But you have the dual benefit of understanding how the business works broadly and then understanding if you zoom in on this piece, what are the levers i need to be pulling to move my metric.

Lenny (00:13:07):
To make that last point a little more concrete, what's an example of that? What's a team that would have their little model? I imagine every team has their own understanding of how the lever works, but what's an example?

Dan Hockenmaier (00:13:17):
I think there's probably two core archetypes for this. So if you have something that's more like a growth team, it's a little bit simpler. They're typically managing some kind of funnel and they can understand do I want to work on driving more traffic, more conversion, more retention. Typically, that's a somewhat linear relationship.

Dan Hockenmaier (00:13:32):
But then you have all these teams that are really managing some tension in the business, which is totally different than a funnel. So say if you have a marketplace quality team, what they care about is driving some standard for quality for the suppliers on a marketplace, but there's not a linear kind of relationship between working on that kind of problem. So if I let on a bunch of new supply to a marketplace, probably the first thing that happens is our GMV or our revenue goes up because we have all these new suppliers which can transact demand, but if those are on average lower quality, it's going to degrade the kind of customer experience and reduce retention over time. And so the model that they're trying to build is how to manage that tension.

Dan Hockenmaier (00:14:08):
Similarly in a FinTech business or many marketplaces now have FinTech elements, you're often underwriting the transaction. That team is thinking about the tension between extending more credit and driving higher spend versus defaults on the other side, like what's the contribution margin maximizing point at which we could offer credit. And so their models are going to look pretty different than what a growth team is managing.

Lenny (00:14:29):
Got it. So you mentioned this idea of archetypes and I was going to ask you when you think about developing a growth model for a company and with Basis One, the company that you ran before Faire, you basically developed these things for startups and how many companies would you say you helped through this process and helped develop growth models for just to [inaudible 00:14:47]?

Dan Hockenmaier (00:14:47):
We probably built 20 or 30 these in my time at Basis One.

Lenny (00:14:49):
Awesome. It's probably more than anyone out there. So I think that's a context to set. You've had more experience doing this than maybe anyone else out there, but we won't compare.

Lenny (00:14:58):
Are there archetypes/templates/tools that you have found useful to think about, I'm coming in fresh to think about the growth model for a company. What have you found helpful to get started to lay the groundwork?

Dan Hockenmaier (00:15:11):
One thing I'll say is going back to 50% of the value being figuring it out, that actually negates the value of templates in some way. You kind of want to build it up for yourself from first principles to understand how the business works. So the more painful the process of building it is, probably the more you're learning. But I do think there's been a lot written on this that you can find online and I think the building blocks I talked about are a useful starting place for how to put these together. I also think Reforge, which is a product in growth school, which we both know has done I think probably the most work to build this into a discipline. And so if you want to go really deep on this, my top recommendation would be to start with Reforge.

Lenny (00:15:45):
Okay. So getting even more concrete, you've built a lot of these growth models. What's an actual example of growth model you've built and more specifically what have you learned from the experience of building that growth model at either, I guess, I was going to say a fictional company but let's go with a real company?

Dan Hockenmaier (00:16:00):
One of the immediate things that you see when you build these is that your growth is much more sensitive to customer retention than you can ever intuit because there's a lot of interaction between having a healthy retained customer base and everything else you care about, which is the rate at which they're referring new customers, generating content, generating contribution margin. And so it quickly makes it clear that actually getting a smaller percentage gain on retention is often much more valuable than making a bigger change in some other area. And as a result you may be misallocating your product and growth resources in a pretty significant way.

Lenny (00:16:35):
So your point there is really important that the more you work on this stuff, the more you will learn what actually is movable and the expected ROI on investments. For founders that are just getting started on this sort of stuff, do you have any just general advice or guidance on okay, retention is probably going to be very hard to move, don't expect that metric even though it's your point that it's often the biggest lever? Any just general guidance of here's where you probably will see impact, maybe you should now count on a ton. How do you think about that?

Dan Hockenmaier (00:17:04):
I think one tactical piece of advice, I think the best way to start this is to find a smart analyst or smart finance person often is the right type of person to partner with and just start building it. So you may have some intuition around this for the core operating model the business is running. This is a little bit different. You need to start building in more variables but I would start there and start iterating on the model. And I think this process of where you have leverage to move it is very hard to intuit. So I think you just have to start and there's a feedback loop between what kind of gains the team is putting up and how that impacts your model and over the course of multiple quarters of iterating on this, you build much more intuition for what works.

Lenny (00:17:41):
Awesome. So say that somebody has been listening to this episode, maybe they've heard it, they've listened to it three times now and they're just like, "Hey, I got a model that kind of feels like the beginnings of something." What do you do once you have a growth model? How does it actually inform what you do as a startup?

Dan Hockenmaier (00:17:56):
It's very helpful input into a quarterly or annual planning process. It depends a little bit on the stage of the company, but say you're at the stage where you have maybe 10 or 20 product pods which are allocated across various parts of the business and you're going through annual planning. Often you want to do a zero based accounting exercise where we say we want to from the ground up decide how those people should spend their time and there's some kind of pod allocation exercise where we're deciding which places those go to. The most difficult thing about making that kind of effort is developing a common currency by which you can trade off their efforts. So this team is saying they can move this metric by X and this team is moving this metric by Y. I have no way to make those two things comparable. The growth model is the function that lets you do that.

Dan Hockenmaier (00:18:36):
And so you can have this analyst or this finance person we talked about who is operating this model work with the product managers to run those scenarios through this model and generate it into a common currency. So now we have a spreadsheet that says these are the things we could work on. This is roughly how it will impact our short or long-term growth. Now you have the ability to make much better kind of allocation decisions. So that would be at the most macro level. And I think the more micro one is for an individual product pod, I have this north star goal, which levers should I be pulling? They should be using their mini models to make that assessment. If you look at the strategy doc for a product group or team, I think that having a model like this should be a core part of that because that way they can articulate what it is they actually [inaudible 00:19:18].

Lenny (00:19:18):
Have you found a model doing this exercise, coming up with a growth model for a company or a startup you've worked with, radically shifted the way they think? Is there an example that comes to mind of, "Oh wow, we came up with this," and they're like they didn't even know this was a huge lever and that changed the way they approached growth.

Dan Hockenmaier (00:19:32):
I mean the first time this was very eye-opening to me and probably the first one I built was the one we built at Thumbtack in partnership with our finance team, and it made it so immediately obvious that we were exceptionally sensitive to the repeat rate of new customers. If you think about Thumbtack, we offered a thousand categories. This is a local services marketplace for people to hire electricians or plumbers or wedding planners. Almost all the traffic came from a very targeted SEM or SEO on a specific thing. So they hired [inaudible 00:20:01].

Dan Hockenmaier (00:20:01):
Initially, it was very hard for us to then go upsell you into something else, but the rate at which we did that made all the difference because it radically changed the LTV of that customer, which then fed back into how much we could go pay to acquire new customers. And so we had a team that was primarily focused on optimizing that initial flow. So SEO and conversion rate. We shipped hundreds of experiments to increase our conversion rate. We had done much less on the life cycle piece like how do we cross sell you into these other things. Building this model helped us internalize that we needed to shift a bunch of resources from top of funnel to deeper in the stack and that ultimately let us build a much better customer journey.

Lenny (00:20:36):
Awesome. Your point about retention being one of the biggest levers reminds me at Airbnb there's like a data dive once where a data scientist found that same conclusion. "Man, if we move retention by 1%, we're going to hit all our goals." And we tried and it was just very hard. And I'm curious, how often do you find that you can actually meaningfully move retention or is there anything that you've found to be effective in increasing retention for many of the companies that you've worked at?

Dan Hockenmaier (00:21:04):
Retention is a tough measure to work on because it is the culmination of the entire product experience. Whether people come back or not has to do with everything that they experienced along the way. And so I think that the primary advice is actually if you were trying to remove a retention style metric, rarely should you go right to the source and send them more email or push notification or something like that. It's really actually about understanding what is the customer experience and what matters to them. And in a marketplace context, it's often depth of supply or the interaction they have with supply. So you actually should be working more on core product levers than you are on, call it, growth product levers to move retention. So it's a deep understanding of that customer journey and where you actually have the opportunity to improve it.

Lenny (00:21:43):
If you can share any idea or any example comes to mind, what's one of the biggest successes you've seen in increasing retention?

Dan Hockenmaier (00:21:49):
Very often I think the biggest wins in retention come from inflecting the early user experience. So if you're at three, six, twelve months out, usually a customer has formed a pretty strong opinion about whether or not they like this product and they have a pretty good understanding of it. But in that very first experience, you have a lot of opportunity to teach them about why this is a valuable product and kind of prove to them that it's valuable. So the experiments that have been most effective have been typically focused on very early life cycle, and one really valuable lens is to look for variability in that experience. So for the first week or month, which customers are having a bad experience but shouldn't be.

Dan Hockenmaier (00:22:27):
So if you take Lyft or Uber, for example, an Uber driver signs up for the service. Some of them just by luck of the draw are going to make a pretty bad hourly rate because a customer canceled on them or they were just in a low density area, and there was some problem with their experience. That driver doesn't know that's not how it works. They might think, "I just make $3 an hour on this platform," and they're never going to come back. And so if you can target streamlining that experience or homogenizing that experience, it's typically very helpful. This is why you see both of those companies dueling it out to guarantee the highest first first week or first month earning. What they're trying to do is prove to you that this is what the experience is going to be like longer term, and you pull up all those below average first experiences to average and drive much better retention curves going forward.

Lenny (00:23:12):
Amazing. Such a great example and story. It makes me think about this work that I imagine every company ends up doing, which is what are leading indicators of future retention. I'm curious if you've had success doing that. I find there's always these obvious things that you can't do much about to increase retention down the road and there's always this idea of building some ML model that's like anticipating retention based on some behaviors that someone takes. Have you had any experience, any success with those sorts of investments?

Dan Hockenmaier (00:23:41):
I think one of the most common analytical failure modes is this pattern, which is our best users do X, so why can't we make other users do that same thing and then drive future attention. It almost never works that way because there's something unique about that customer and their experience which is driving it. And so these correlational exercises, I put very little weight in because I've rarely been able to move Bucket B into the better Bucket A, for example. So I think it's much more about understanding where the real drivers of value are, how to create that really good first round experience to prove to them.

Lenny (00:24:12):
Awesome. And it's interesting that onboarding comes up a lot on these conversations, just the power of onboarding and how much effect that has down the road.

Lenny (00:24:20):
To your point, people often want to increase retention by focusing on people about to turn and this is a great reminder that your biggest lever is early on when they're just experiencing for the first time.

Dan Hockenmaier (00:24:30):
So yeah, the flip side is absolutely true, right? Working on early user experience is highly impactful. You often see product teams say we should work on resurrection because we have this huge pool of users that is churned out. If we get just 1% of those people back, it's going to be such a big lift. The problem is that pool of users is the group of people who has tried the product and decided they don't want to use it and so it's very hard to convince them otherwise. It's usually much higher leverage to focus on new users and, as a result, typically you want to wait to spin up resurrection efforts until you've exhausted some of these earlier funnel efforts.

Lenny (00:25:02):
So I want to transition a bit to talking about marketplaces. I honestly can't think of anyone that's worked on more marketplace companies than you. Maybe Jeff Jordan at a16z who's invested in Airbnb and worked at eBay and OpenTable and these companies. But I feel you have so much insight into how to grow and run a marketplace and so I'm really excited to dig in to stuff here.

Lenny (00:25:23):
My first question, just why are you excited about marketplaces? What gets you continuously interested in working on marketplaces and why are they such interesting and good businesses?

Dan Hockenmaier (00:25:32):
Why is a marketplace a good business? The first thing is it's like a perfect fit for the venture model, which is why everybody's so interested in them, right? They're very hard to get started, super capital intensive to get started, but once they're rolling, they're very hard to stop. You get this kind of compounding defensibility and gains that makes them very hard to stop. And, as a result, as a marketplace grows, you see these crazy things in the metrics which you don't see for most other businesses. So typically, as you acquire more and more customers, you're acquiring the marginally good fit customer over time and your CACs go up, your LTVs go down, it gets harder and harder.

Dan Hockenmaier (00:26:08):
Marketplace is actually the inverse. The supply liquidity is improving, the experience is improving. So often actually as you see later cohorts in marketplaces, CAC goes down, LTV goes up. You see this crazy inversion where the business gets better and better over time. And so that's I think one of the reasons they're such great businesses. I think to your question on why they're fun to work on, I think everything is just harder. Every question is more complex and so that's made them I think really fun to work on for me.

Lenny (00:26:37):
And that's interesting coming back to your first point about marketplaces are hard to start, but once they're started they become cheaper to grow and you build moats and work effects. It feels like you as a advisor and person that works on marketplaces, the same thing happens. The fact that you've worked at so many marketplaces makes it more interesting and fun I imagine because you've seen so much of this and so many people haven't experienced this, and so there's almost a Dan Hockenmaier network effect.

Dan Hockenmaier (00:27:02):
I appreciate that. I mean I think there's a few folks in this community who've worked with a bunch of marketplace, and you start to see the same topics over and over. So it's really fun to riff with that group on these topics.

Lenny (00:27:12):
This episode is brought to you by Eppo. Eppo is a next generation A/B testing platform built by Airbnb alums for modern growth teams. Companies like Netlify, Contentful and Cameo rely on Eppo to power their experiments. Wherever you work, running experiments is increasingly essential, but there are no commercial tools that integrate with a modern growth team stack. This leads to waste of time building internal tools or trying to run your experiments through a clunky marketing tool. When I was at Airbnb, one of the things that I loved about our experimentation platform was being able to easily slice results by device, by country, and by user stage.

Lenny (00:27:49):
Eppo does all that and more delivering results quickly, avoiding annoying prolonged analytics cycles and helping you easily get to the root cause of any issue you discover. Eppo lets you go beyond basic click through metrics and instead use your north star metrics like activation, retention, subscriptions and payments. And Eppo supports tests on the front end, the back end, email marketing and even machine learning clients. Check out Eppo at geteppo.com, geteppo.com and 10X your experiment velocity.

Lenny (00:28:22):
Okay, so we're thinking about marketplaces. One of the biggest questions people ask about marketplaces is how do you know if they're going well? What are health metrics? What are KPIs that you find are most helpful to think about when evaluating the health of a marketplace?

Dan Hockenmaier (00:28:36):
I can give you a couple of pretty obvious ones or basic ones and a couple that are a little bit more nuanced. So that in that first bucket, certainly you need to look at some measure of GMV or transactions. You need something that brings together both the supply and demand side to make sure it's working and that's typically your ultimate north star of which everything else ladders up to.

Dan Hockenmaier (00:28:54):
I think the second would be a deep understanding of unit economics because the dynamic I talked about where they're hard to get started means that most marketplace in the early days have pretty poor or maybe even negative unit economics. Instacart famously was losing the money on every order. Uber was losing money on every drive. And so I think understanding that and the components that lead up to it is a very important part of understanding your marketable. So those would be the two kind of obvious ones.

Dan Hockenmaier (00:29:19):
I think other two that I would look at beyond that, number one is liquidity. And this is a really broad term that people define in a bunch of different ways. The definition I would give is how reliable is the marketplace? If the consumer is looking for something or supplier is looking to sell something, how often can they do that thing that they're trying to do? And ideally you want to express this metric in the form of a dimension or a type of product experience that a customer really cares about. So for Uber or Lyft, wait time is a classic example. As you add more supply, the average wait time for a customer goes down and there's some magic moment around four or five minutes where it really clicks and is now just a much better service than calling a traditional taxi or something else.

Dan Hockenmaier (00:30:01):
For a commerce market, typically it's some form of conversion rate or search to fail metrics. So if I go look for this thing on Amazon, how often can I find it and convert? And so by articulating what the customer cares about and where this threshold is, you can tell when you have a liquid marketplace and essentially until you have a liquid marketplace, really nothing else matters. So this should be the primary thing you're focused on defining and then building towards, and this is why you hear a lot of advice to marketplaces to basically cut scope down to a specific geography or a specific category so you can focus on generating liquidity in that area before you can then scale it elsewhere. So this is the number one, I think, metric for a marketplace.

Dan Hockenmaier (00:30:41):
And then the last one I'll give is share of wallet. So this is effectively for your buyer, how much of their total spend are you getting on your marketplaces versus the alternatives. For your seller, how much of the business that they're doing is you versus others? So for Uber drivers, if they spend X amount of hours driving per week, how much are you getting versus Lyft or Door Dash or something else? For a retailer on Faire, if they're buying to stock their store, how much of the product on their shelf came from Faire versus from somewhere else? These are very important metrics for us to understand.

Dan Hockenmaier (00:31:11):
One, for the obvious reason that as it goes up, your LTVs go up and you have a much better business. But probably more importantly than that, the higher it is, the less likely a customer is to multi-tenant, meaning use another marketplace or another service. And ideally you want that customer to commit to using just this marketplace and the higher share of wallet is the more likely that is. But I will say that it's very hard typically to get this to happen on both sides of the marketplace simultaneously. So often you have to pick your leverage point, which one do you think you can actually drive very high share of wallet on.

Lenny (00:31:41):
Awesome. Okay. So just to summarize, the metrics that you find to be most helpful in tracking marketplace health, there's these two that are just general business health metrics, GMV and unit economics. Then I think what's most unique to marketplaces, which is liquidity and essentially it's just like how often are people having a good time on both sides of the marketplace? And I love the way you broke it up like for Uber it would be how quickly do you get a car. For most marketplaces, it's just like what percentage of the time do you get something that you want, fill rate basically. And then share a wallet, which to me feels like ... even again going back to the first kind of bucket of just broadly this business going well, do you think about that separately versus just business growth and how much we're making, do you feel like share of wallet is a different category of metric?

Dan Hockenmaier (00:32:30):
I do think share of wallet is different for this reason. If you could tell me we could grow GMV 10% by getting 10% more customers or by getting 10% more of our current customers' wallet, I would take the latter because you now have a deeper relationship with them, which tells you something more about the future retention and defensibility of the marketplace. So I think it's basically a measure of depth rather than breadth. And I will take depth every time in a marketplace.

Lenny (00:32:57):
Awesome. You worked on consumer and B2B marketplaces and so I'm curious, do you find share of wallet is important on both types of marketplaces or is it a lot more important in B2B?

Dan Hockenmaier (00:33:06):
So you typically have some form of business on the supply side of a marketplace. Maybe it's a pseudo business that's effectively a consumer, but you can almost always measure some form of share of wallet on the supply side of a marketplace no matter what. On the consumer side, if you're a B2B marketplace, you typically have a cleaner share of wallet metric, but it's not always the case for consumer businesses.

Lenny (00:33:25):
Sweet. And again, just to clarify share of wallet, it's essentially percentage of spend in, say, a problem space that they're giving to you. So for Faire, it'd be like the retailers, what percentage of their vendors come through Faire?

Dan Hockenmaier (00:33:39):
Exactly. If you look at their shelf in their store, what percentage of that shelf came from Faire versus something else?

Lenny (00:33:45):
Awesome. Another constant question marketplace founders have is whether they should focus on supply or demand. And I know it's never black and white, but do you have any general advice on where to focus?

Dan Hockenmaier (00:34:00):
Yeah. I mean the answer is obviously both to some extent. I think you can't ignore either side. I do think though that on average when you hear advice about where to focus, people over rotate on supply and actually are under focused on demand. And I think there's a couple of reasons for this. One is supply is disproportionately important early on because it is the product. Until you have enough supply, you don't have anything and so you do have to focus on it to a high degree early on. And two is often the supply side of marketplace is using the product more deeply. There's more product surface area and you need more product resources on the supply side.

Dan Hockenmaier (00:34:35):
However, I think this tricks people into thinking that that's the optimization function or you think you should think more about supply. I think ultimately demand is the only thing that matters. If you are successful at aggregating the demand in your industry, you will have the winning marketplace. Because if you go to a supplier, a restaurant or an electrician or a driver and say, "I have this customer for you that I can give to you at a rate that's going to make you money," they're always going to say yes. And so demand is the currency. And so when you think about trade-offs or how to optimize a business, I think taking the perspective of the customer or the demand side is always the right one.

Lenny (00:35:11):
I think there's a really important nuance here, and there's actually a little mini Twitter debate with Bill Gurley I think a while ago where he made the same point that in the end, the most important thing you got to get right is aggregate all the demand. You need to become the place people come to transact in that space. But oftentimes the way you do that is acquire supply that is hard to acquire. And so would you agree with that? Often it's just, yes, prioritize the customer experience, but that may be you need to spend most of your time acquiring supply so that they're happy.

Dan Hockenmaier (00:35:40):
Absolutely. And the culmination of those two points is you only should acquire supply to the extent you understand how it impacts demand. So, for example, if we go back to this liquidity metric, there is some point for a market for Uber where you don't want more supply because you're no longer reducing wait times or doing something that improves the customer experience in a meaningful way. And similarly to the kind of pet store example on Amazon, there's some amount of supply where you're probably no longer incrementally improving the pet buyer's experience and so it's probably not worth investing those dollars. So supply is critically important, but it has to be framed from what is the customer benefit that I'm driving.

Lenny (00:36:17):
And another way to put that is what's the biggest constraint to your marketplace's growth, right?

Dan Hockenmaier (00:36:21):
Right.

Lenny (00:36:22):
While we're on that topic, do you have any just rough heuristic that you use to understand which side is most constrained? This may be a big question that isn't answerable in a short answer, but any thoughts there?

Dan Hockenmaier (00:36:35):
One is I've actually become less and less focused on pure marketplace balance metrics. They're important to monitor, so ratio of buyers to sellers, some of these other things. But actually the thing that matters is can you write an ROI equation for acquiring supply and demand which fully internalizes the marketplace dynamic. So what I mean by this is if you're acquiring a new customer, you need to include the CAC of acquiring that customer, but also the CAC of acquiring the supply for that customer to purchase, which is based on some ratio between the two.

Dan Hockenmaier (00:37:09):
And similarly on the supply side, that business can't make a sale unless you also acquire the customer to transact with them. And so if you have dual-sided ROI equations which are appropriately capturing this dynamic, actually I think you can somewhat ignore marketplace balance and just push your acquisition all the way out to the payback period that you're comfortable with on either side. I think the one exception to this would be are there externalities which you can't capture in this equation. So for example, if you have too little demand for Uber drivers, at some point do they just become disillusioned with the service, switch to Lyft, talk badly about it on social media. You do have to look out for extreme low supply or demand scenarios. But generally my view is build really strong ROI models that account for this and then just push to your threshold.

Lenny (00:37:57):
I like the sound of these ROI models. Do you have any guidance for folks to come up with these models in some way? Or is that a whole master class of its own?

Dan Hockenmaier (00:38:06):
So there's a lot of nuance by business, but the basic formulation is CAC for the side you're focused on. So let's take Uber again, as an example. CAC to acquire a rider and then an additional amount of CAC loaded on for the supply, the drivers that you're acquiring, times the ratio of drivers to supply you're acquiring at that time. So basically, do I need one driver for every 10 passengers? We then take the CAC of that passenger times 10% of a driver. That gives you CAC and then you compare that to the LTV of the customer, and that allows you to calculate payback period.

Dan Hockenmaier (00:38:44):
Now there's a lot of nuance when you get into an actual marketplace because often they're referring other sides of the marketplace or other things are happening. But that's the basic formulation.

Lenny (00:38:53):
Wow. Okay. We should do an actual master class on this formula concoction.

Lenny (00:38:57):
A question I wanted to cover also is I find that for early stage marketplaces, founders sometimes over focus on the theory of marketplaces and how all this stuff that people have put out, including yourself and others, about just how to think about marketplace, all the complexity there. But I find that oftentimes it's simpler just to think of a marketplace like 90 something percent of your success is going to be the same things that any business will have to deal with, growth and profit and retention, all these things. And then there's these additional layers that make a marketplace more complicated. And so just to double click on that last piece, what have you found to be most different about working on a marketplace business versus non-marketplace business?

Dan Hockenmaier (00:39:43):
Yeah, it's a good question. So I think that effectively every decision you make in a marketplace has a second order consequence that you need to think through and maybe third and fourth order consequences at that. Take something like pricing. It's like this is a pretty complicated topic no matter what, but if you're looking at a SaaS business and you're trying to figure out how to price your subscription, theoretically you can draw a curve which says, "As my price goes up, fewer are going to convert," and so just find the optimal point on that curve where we're managing the tangent between more customers versus more revenue per customer.

Dan Hockenmaier (00:40:15):
But if you take a marketplace, typically you're charging commission on the supply side and their sensitivity to that commission is much harder to understand because, theoretically, if they can transact at a rate that makes them money, they'll sign up all the way to that highest possible commission you could charge. The more you charge, the more you can fund benefits for your customers. So if Amazon charges a higher commission, they can fund more returns and faster shipping for their customers. And so what's the right balance between charging more and maybe kind of discouraging supply from signing up to giving more benefits to demand and encouraging them to sign up?

Dan Hockenmaier (00:40:52):
So it's very hard to model that kind of relationship. There's not a simple curve that describes it and so many decisions follow this same pattern. And one lesson I've learned the hard way a bunch of times on this is that if you think about running a marketplace, you're basically a gardener. You have to have a very light touch. If you're building a SaaS business, you're a construction worker, you're building the product and the features and selling it, and it's this very linear thing. For a marketplace, you're messing with this ecosystem that you don't actually really understand how it works. Sometimes you might do something over here which drives this long-term effect two months later and then you're going to be pulling your hair out two months later trying to figure out what you did over here that made that thing happen.

Dan Hockenmaier (00:41:30):
And so I think that the main advice is to tread lightly. When you're messing with the core incentives or mechanisms of a marketplace, be very careful, particularly if you got something that's working on playing with those variables.

Lenny (00:41:42):
I love that metaphor and your point about pricing reminds me ... Your colleague at Faire, Carla Pellicano, she led the pricing recommendations team at Airbnb. It was a team of, I don't know, probably a hundred people that were just dedicated to pricing, figuring out what prices to recommend to hosts, how to get them to adopt these recommendations, building a model to actually come up with the recommendations. And so to your point, pricing is such a complex beast and especially in a marketplace.

Dan Hockenmaier (00:42:09):
Absolutely. And in general, Carla's been such an incredible force in growing our team and helping us think more rigorously about marketplaces. This is one of the things I mentioned at the start that makes Faire so fun is we've got a lot of people like this that are just so fun to riff with on marketplace problems.

Lenny (00:42:23):
There's so many ex-Airbnb people at Faire. It seems to be a magnet for the Airbnb alumni. So whatever you're doing, keep that up.

Lenny (00:42:30):
Another topic that I wanted to chat about is expanding marketplaces and just the idea of thinking about where you expand to with new markets, new verticals, and then also horizontal versus vertical marketplaces. But first, how do you think about the idea of expanding your marketplace once you've got a foothold in a specific area?

Dan Hockenmaier (00:42:48):
I've been fortunate to work on marketplaces that are in these massive, massive industries, which is actually true for a lot of marketplaces because they tend to have winner take all dynamics in really big markets. And so you get these really huge [inaudible 00:42:59].

Dan Hockenmaier (00:42:59):
So the local services industry for Thumbtack or the global wholesale industry for Faire, these are meaningful percentages of global GDP. They're huge markets and, as a result, they're really frustrating to work on and also really fun to work on because you have this thing where there's 10 big opportunities that are just one click away from your core business and they all seem really good ideas to do. So how do you actually prioritize between doing those different things? And one thing I've learned here is that actually beyond a certain point, TAM or the size of the market actually matters very little because these are all big enough that they would dramatically inflect the curve of the business if you make them work.

Dan Hockenmaier (00:43:35):
It's much more relevant to focus on a couple things. One is how adjacent is that to the business as a proxy for can we actually go get it? So if you think about Instacart's setup options, it makes much more sense for them to expand into convenience stores which they have than into traditional retailers because the convenience store looks much more like their current model. The high frequency, shipping speed matters a lot, fulfillment speed matters a lot. And so it's much more likely their current model's going to work there than trying to expand into something else. And that's the right prioritization function for them to think about versus as retail is a slightly bigger market that could be in stores.

Dan Hockenmaier (00:44:09):
And the second thing is are there places that you can accentuate your network effect by expanding into new markets. And what I mean by that is are there places where you can use the same supplier or a consumer has demand for multiple things and so it makes your marketplace stronger versus trying to spin up a new network. So for Uber, it makes all the sense in the world to have Uber Eats because, one, they're the same drivers in many cases, but two, the customer wants rides and meals. And so you automatically have this built-in supply base where if they try to do something that was one click further away from this, it would be much less important to them. And so I think that's the way to prioritize new bets.

Lenny (00:44:46):
That is such an interesting point that basically if you're thinking about the upside of a marketplace, think less about just the total TAM of all the adjacent marketplace opportunities in the markets around them and more about how easily it'll be for you to expand into them even if they're smaller.

Dan Hockenmaier (00:45:01):
Exactly.

Lenny (00:45:02):
Awesome.

Dan Hockenmaier (00:45:04):
There's one other lesson here which I've learned a few times, which is that product is the thing that matters when expanding. So because of this dynamic we talked about where liquidity is so important and there's a race to get there, like the first person, the liquidity wins, you often see this arms race where people will spend a huge amount of money on go to market and incentives to bootstrap the market. And that is an important part of the strategy because it actually does matter who drives liquidity faster.

Dan Hockenmaier (00:45:33):
But I've learned over and over and over that that's actually not the main thing. It's who can deliver an incredible end-to-end customer experience first, even if for a smaller number of customers. Because that's what creates the flame where actually customers are really loving it, retained, talking about it, and you can then expand from there. So the other big learning from expanding a marketplace is don't let go to market get too far ahead of product. You need to keep those two pieces in lockstep as you're expanding.

Lenny (00:46:01):
This touches on a really common piece of advice for marketplaces, which is don't focus on GMV and growth rates and just expansion early on, but instead focus on getting a flywheel going even if it's small, to show that you can make people happy and you can give people what they're looking for. Is there anything you can add there or talk about?

Dan Hockenmaier (00:46:20):
Yeah, I think that's exactly right. And the reason this is the right advice is because everything else follows proving you have a good customer experience. Even if you have a very few customers, if your cohorts look really good, they're retaining or even kind of like the classic smiling curve where you see more in engagement later in the life cycle than you do earlier, that's the thing that gives the company conviction to invest resources against it. That's the reason that VCs are going to want to invest rather than a bunch of low quality GMV in a market.

Lenny (00:46:49):
Awesome. Speaking of VCs, investing and expanding marketplaces, something that I've noticed is a lot of marketplaces try to find a SaaS business to build on top of their marketplace and find some kind of recurring revenue component and then, in reverse, a lot of SaaS businesses look for how do we add a marketplace to what we're doing. I'm curious how often you find that this actually works out and what do you have to get right to add this other type of business model on top of something that's working.

Dan Hockenmaier (00:47:17):
So broadly, I will say I think it's easier for a marketplace to go SaaS than it is the other direction and the reason for that is two things. One is it's a new capability to generate demand, which is fundamentally what a marketplace has to do, and it's a higher value activity. This is why the effective commission of a marketplace, often 10, 15, 20% is much higher than the effective commission of a SaaS business in the 2% to 3% range. So you're just doing much more of the value chain in the marketplace. And second is the marketplace by definition starts with relationships on both sides. But the SaaS business does not have any relationships with the demand side customer. And so they have to acquire a whole new type of demand to make this work.

Dan Hockenmaier (00:47:58):
It's not to say it can't work. There's actually a classic kind of SaaS bootstrap to marketplace playbook. This is what OpenTable did. I think we've actually seen some new examples of companies doing this. One in the healthcare space is Solve. They built some interesting products for healthcare clinics that they're now bootstrapping into marketplace. And so I think it's possible, but I think it's very difficult. And then for a marketplace, the lens you should take is much less about how to drive more monetization, but just how do we create a much better experience for our customers because there's some painful thing that they're doing today that we can build for them instead. And so how do we better integrate with the way that they're running their back office or accounting systems is a classic example of where you can make it much better.

Dan Hockenmaier (00:48:42):
In the process, you're often making their lives easier, but you're also making your product much stickier. Your retention will go up as a result of this. And so I think if you take the lens of what's the customer pain we're solving, you'll be much more effective than how do we get a few more points of margin out of this customer.

Lenny (00:48:57):
If a founder was coming to you and they're like, "Hey Dan, we are a marketplace and we're thinking about adding a SaaS product on top, would you, one, try to discourage them from that? And if, two, that doesn't work, what would you suggest that they focus on most?

Dan Hockenmaier (00:49:12):
I mean I think the first thing is looking at those core metrics we talked about, do they have a really liquid high performance marketplace first. That has to be the optimization function. And before you're there, I don't think you should be thinking about some of these expansion levers.

Dan Hockenmaier (00:49:27):
The second would be show me the customer problem or the reason it's so hard to engage with this marketplace today that we need to build a deep set of tools or products for this customer to solve. And if both of those things are true, then I think maybe it's quite interesting. But I think more often than not, it's better to focus on the core marketplace.

Lenny (00:49:46):
Awesome. Another common question that marketplace founders have is should I go vertical or should I go horizontal? So thinking about eBay as an example, they are very horizontal. You could buy anything you want on eBay. And then there's all these spinoffs that emerged, just classic cars, eBay for classic cars, eBay for guitars. And I'm curious if you have any advice there for either an early stage founder trying to decide should I go horizontal or vertical and/or where do you find the biggest opportunities to slice off a piece of a successful horizontal marketplace?

Dan Hockenmaier (00:50:18):
Yeah, absolutely. And in that eBay example, there are now a few quite successful examples of this like [inaudible 00:50:23] and StockX are two where they carved out the sneaker category and the key insight was you couldn't trust the inventory you were getting on eBay. So there's a lot of work you need to do to verify, and those businesses just did it much better than eBay. Broadly though, I think that we over hyped the idea of unbundling. So I think every six months I'm seeing an article where somebody wrote, "This is the unbundling of Reddit, the unbundling of LinkedIn, the unbundling of Facebook." We're going to take all those blue links that you saw on another site and they're all going to become new businesses. And very rarely that thesis plays out. I think the core logical error in the argument for unbundling is that they over focus on one type of improvement, which is user experience and they under focus on the things that make scale businesses have better economics.

Dan Hockenmaier (00:51:07):
And so to unpack that, if you look at UX, like if you built a LinkedIn just for construction workers or just for architects or just for investment bankers, you could definitely build some set of features that group liked better than the core LinkedIn experience. But then you have to weigh that against all of the benefits of being broader. And so the two big pieces of where you get benefits from scale are in your customer LTV and then I think the network effect you can build. So if you go back to that Thumbtack example, we had a spreadsheet which tracked hundreds of verticalized competitors where somebody would try to pick off the electricians category, the wedding category, the lessons category, and very few of them got traction for the simple reason that we could upsell customers into a thousand things.

Dan Hockenmaier (00:51:48):
And so our customer LTV was always higher and we would always win when we were bidding against those other customers on SEM keywords that were relevant to that category. So it becomes very hard to compete if you're picking off this kind of narrow thing unless you find something which that sub-segment is itself very high frequency or very high dollar value. So Airbnb's an example. They unbundled from Craigslist because they picked off this massive high frequency, high dollar value category [inaudible 00:52:16] but there's not that many of those examples.

Dan Hockenmaier (00:52:18):
And then I think the other source of benefit would be the network effect. So if you go back to LinkedIn, for example, I think actually there's an opportunity and we see some successful businesses picking off now blue collar work. So there's a company called Workrise. I think they used to be called Rigup where they're basically building like a LinkedIn but for blue collar work. And that works really well because it's a huge segment and it's somewhat self-contained.

Dan Hockenmaier (00:52:39):
But for most other things there's a lot of fluidity between the employers and the employees in terms of who wants to transact with one another. And so if you're an investment banker, you don't really want to be on the LinkedIn investment banker because you're probably in the future going to want some other job. So you want to be in the biggest network that's relevant to you and so this is why you can complain about LinkedIn's UI all day, but they have a very strong place in the market because of that network effect. So broadly, I think there are some pretty interesting examples, places where you can unbundle but they're rarer than people think.

Lenny (00:53:10):
That is amazing. There's so much value in what you just shared. So one takeaway I have here is you have an opportunity to unbundle/split off into a vertical marketplace potentially if there's high order value and high frequency. And the third piece is there's almost a self-contained network that doesn't benefit significantly from the rest of the network. For example, I love the Rigup example, like I doubt oil rig operators are on LinkedIn and when something comes around that's like, "Oh, all my buddies are on this thing, I'm going to be on there," you don't need the rest of LinkedIn.

Lenny (00:53:40):
The first piece, though, is interesting. So Airbnb I wouldn't say is high frequency. I'd say it's just very large high order value. And so I wonder if you just need one or the other really in a big way. Really high order value or really high frequency?

Dan Hockenmaier (00:53:51):
This is a good point. Probably what you're solving for is customer LTV and you can get that in multiple places. There's not that many things which are both high frequency and high dollar value so you can do both. I do think if you go to a place that is low frequency, it comes with all kinds of new challenges because, without frequency, customers forget about you. And so what is the hook to get them to come back? Do you have to reacquire traffic? It creates a whole other set of problems. But if you can get it right like in the Airbnb case, it can work really well.

Lenny (00:54:17):
Yeah, Thumbtack is a classic example of how often you need a plumber. And even with the thousands of services that you all had, from what I understand, it's still a struggle to get people to come back often. And remember Thumbtack when they had, "Oh, they have an electrician. Oh yeah, Thumbtack."

Dan Hockenmaier (00:54:30):
That's right. Initially, it was very difficult. I mean the average person hires eight or ten new professionals a year, the average homeowner. And so that's decently high frequency, but it's not food delivery or a ride sharing or something like that.

Lenny (00:54:42):
Coming back to Faire, so Faire is one of the maybe few really successful B2B marketplaces and it's always felt like there's this gap in B2B marketplaces. You always feel like there should be many more. Like why are there so many consumer marketplaces but so few B2B, and I'm curious what's your take there. Do you think there's a rising trend to B2B marketplaces? Do you think this is always going to be a smaller collection? What's your feeling?

Dan Hockenmaier (00:55:07):
So I do think we'll see more of them. Part of the reason we've seen fewer is there are fewer potential founders who understand B2B problems because most of them are consumers and so the consumer use cases are more obvious. So if you take Faire, for example, when I met the founders, which is probably five years ago now, I immediately understood what they were talking about, but only because I had run an e-commerce business in the past and I had the experience of dealing with a hundred suppliers and line sheets and PDFs going back and forth and pricing not being right and just how painful it is to be a retail buyer. And they had a solution which was much better and that clicked.

Dan Hockenmaier (00:55:44):
But had I had the same conversation five or 10 years ago with the team at Convoy like I don't know anything about trucking, I probably wouldn't have understood why that business was going to work. And so I think there's partly that. Just the discovery process takes longer for that reason. But I think the reason we won't see a huge explosion in this area is that B2B also comes with something else, which is much lower fragmentation in many cases and you need fragmentation for a good marketplace. The more concentrated either side of your market is, the more leverage they have, the less likely they are to need you and the less likely they are to be willing to pay a high commission.

Lenny (00:56:20):
You made that point to me once when we were talking about marketplaces years ago, and that's so stuck with me that when evaluating marketplaces in B2B especially, usually the reason it's not going to work is just it's not fragmented enough. And just to double click there, can you explain what that means? What does fragmentation mean in a marketplace context? And then are there any examples of really low fragmentation of this will never work as a marketplace? And then here it's really high and this is why it's working?

Dan Hockenmaier (00:56:45):
Fragmentation is basically just a measure of how many total businesses are there in the space relative to the transaction volume in that space. And if you took the top 5% of suppliers in the space, what percentage of the total volume are they doing? And the higher that percentage is, the less fragmented you are. The challenge that creates for a marketplace is if there's 10 companies in a space that are doing 80% of the volume, it's very important for me to have a relationship with those 10 companies. But those 10 companies are also big enough to have their own sales teams, have their own internal operations. They just need less from a marketplace and, as a result, they're going to be willing to pay less. And you're also probably going to see many more problems with disintermediation, which is when the supplier and the customer go around the marketplace because they can just transact themselves.

Dan Hockenmaier (00:57:27):
One principle to use here is how many total dollars are attached to each transaction in the marketplace. When it goes above a certain amount, it becomes much more attractive to figure out how to go around the marketplace. With ride sharing, for example, the absolute dollars of commission on a ride is $2 or $3. Is it worth it for the driver to figure out how to call the passenger two minutes in advance, go around Uber and pick them up? Maybe. You probably see some of that happening, but usually not.

Dan Hockenmaier (00:57:55):
But now take there's a bunch of people in the kind of material space within B2B marketplaces, you have manufacturers of say beauty products who need to source aerosol cans and all the inputs into making beauty products. There's not that many of these big suppliers and each of their transactions may be tens of thousands of dollars, hundreds of thousands of dollars, millions of dollars. The commission you would charge on that order is too high because a supplier would rather just pick up the phone and call this person and save those tens of thousands of dollars. And so you just run into these kind of fundamental problems where a marketplace doesn't work anymore.

Lenny (00:58:29):
That makes sense. Basically, how much value are you bringing to this market? And if it's not enough where you can charge anything meaningful to run a business is just not going to work. And so that's a really good way of framing it.

Lenny (00:58:41):
Final question around marketplaces and, broadly, and I'll let you go. You've spent a lot of time on marketplaces. You've seen their evolution. You've worked on this maybe for the past decade. Where do you see the future of marketplaces going?

Dan Hockenmaier (00:58:54):
I actually wrote a blog post on this where we charted the commission that a marketplace charges and the year they were founded. And if you put those on the x and y axis, there's this very clear up into the right trend. Newer marketplaces are charging higher commissions and they're doing more work to justify those commissions. So broadly the evolution looks like kind of Marketplace 1.0, which is all they're doing is aggregating demand. So that's Zillow and Home Advisor. They're basically like lead gen and their commission rate is often pretty low. It's like 5%, maybe 10%.

Dan Hockenmaier (00:59:24):
Then you have a managed marketplace like Airbnb or Etsy, which did something really fundamental on top of that, which is generated trust. So they deadhead supply ... You could probably tell me more about what Airbnb did in this space, but they made it a safe transaction and there's a lot of work it takes to make that transaction trustworthy and safe and so they charge a higher commission as a result.

Dan Hockenmaier (00:59:48):
There's then one click beyond that which, for lack of a better term, you could call it a heavily managed marketplace, but now they're typically doing some work in the value chain, which is distinct from just aggregating demand. So DoorDash and Instacart own logistics. They took over logistics and, as a result, DoorDash did a much better job than the previous model of Seamless of being able to bring on a lot more restaurants and make it much more reliable for the customer. As a result, they could charge more to the restaurant. Similarly, Faire, we actually underwrite the transaction. We take the risk. If that transaction falls through or the retailer defaults, Faire eats that. And so we are playing a much more fundamental place in that transaction.

Dan Hockenmaier (01:00:25):
But as you play this out, what happens at the end of this continuum? Ultimately, you're charging a hundred percent commission and you're not a marketplace anymore. So I think, as we think about the future of marketplaces, one important question is which marketplaces are going to tend towards evolving out of the marketplace model altogether and which will stay in marketplace mode in equilibrium? And you see these examples already like people talk about Opendoor as a marketplace, but it's not. It's an e-commerce website which has the highest price points you can imagine because you're buying houses, but there's no supplier on the other side. They've already bought the house so it's just e-commerce. And I think many marketplaces will go that way. And I think the variable to me that matters in deciding which case you're going to have, whether they consolidate or not, is how much creativity there is in the space. So how much do you need the supplier to be coming up with interesting new things for your customers to buy.

Dan Hockenmaier (01:01:18):
What the customer cares about is actually commoditization. They want the same experience every time. Then you're ultimately going to evolve away from marketplaces I think. With ride sharing, basically what I want is a clean car that shows up on time and gets me there every time. So as soon as autonomous vehicles arrive, we're going to fully consolidate that industry and it's not going to be marketplace model any more. On the other end, you have Etsy and Amazon. I think Faire's in this bucket. Steam, the video game marketplace, is in this bucket where the thing you care about is suppliers bringing you amazing creative new things. And that's something that big companies are really bad at doing. So they need the marketplace suppliers to supply this. And so I think those businesses stay in marketplace mode longer term.

Dan Hockenmaier (01:02:02):
And then the middle, I don't exactly know how to call what happens in food delivery. You do want some standardization elements, but you also want the local restaurants and so does DoorDash win or the Cloud Kitchen model win? I think it's a little bit harder to understand, but I do think that's kind of the variable that's determining where the future of marketplaces are going.

Lenny (01:02:19):
It's interesting to think about this event horizon for when a marketplace is no longer a marketplace. Is a simple way to think about that being when you own the supply, you're no longer a marketplace. When you don't own the supply, you are. Is that how you think about that?

Dan Hockenmaier (01:02:34):
Yeah, that's a good mental [inaudible 01:02:37]. Perhaps another way to say owning a supply is when there's no longer a direct transaction between supply and demand. That's what Opendoor took out, for example. You're not transacting with the home seller. You're transacting with Open Door and so that's, in my mind, no longer a marketplace because you also eliminate some of the marketplace mechanics we were talking about a bit.

Lenny (01:02:55):
Awesome. Dan, this has been incredible. I feel like we've achieved our goal of getting really deep into the weeds on growth models and marketplaces. Two final questions for you.

Lenny (01:03:06):
Where can folks find you online if they want to learn more, reach out and how can listeners be useful to you?

Dan Hockenmaier (01:03:11):
Yeah, so I'm on Twitter at Dan Hockenmaier and then LinkedIn. People should feel free to reach out. I think the most useful thing is we're always growing our team at Faire. And so for folks who are interested in this space, I would love to connect with them.

Lenny (01:03:26):
Where do they go to learn more and apply for Faire?

Dan Hockenmaier (01:03:29):
Just faire.com or faire.com/careers.

Lenny (01:03:32):
And that's Faire with an "e" at the end?

Dan Hockenmaier (01:03:34):
That's correct, yes.

Lenny (01:03:36):
Awesome. All right, Dan, thank you for being here.

Dan Hockenmaier (01:03:38):
Thank you so much for the time.

Lenny (01:03:41):
Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcast, 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.