Marketplace Analytics 101 With Colin Gardiner

So what is analytics? I think people use this term a lot and just throw it out there and analytics, yeah, let's do it. But what really is, is like, especially for online products, is really like instrumenting and discovering and interpreting the data that comes from your platform. So things happen in marketplaces. Those are called transactions typically, right? And that creates data and it creates a footprint of things like who bought something from who, or who rented something from that person, or they create a listing. You know, there's all these different types of pieces of data that are created, both from transactions, from listing on the marketplace, from searching on the marketplace. And all of that creates data that needs to be analyzed to make really good business decisions. And so that's what analytics is at its core, is really about instrumenting your marketplace so that you can get very accurate and clean data out of it so you can make great business decisions. That is the end goal of analytics. And so let's go into a little bit further. You know, in terms of philosophy, right? Like, what, like, why are you doing this? Why does it matter? And what do some of the greats think? You know, I think one of my favorite books, you know, kind of the godfather of OKRs, if you're not familiar with OKRs, we can talk about it afterwards. But John Doerr, famous venture capitalist, you know, you had this idea that you have to measure what matters, right? And so that's really what analytics is about is can you measure the things that matter that move your business? And that's what the OKR framework that he came up with is all about. 2, management guru Peter Drucker. If you haven't read his famous books, I highly recommend doing it. But he has a famous quote, what gets measured gets managed. And it's the same thing for any analytics, right? If you measure it, and you see it changing and can impact it over time, you will manage it and you'll make it better. And that's really like, for me, at the core of a lot of analytics is if you don't measure it, you won't change it because it's going to be about your opinion. It's going to be you like testing the wind. You got to measure it so that there's no politics, there's no religion in it. It is just the numbers. And that tells you the truth and makes you look in the mirror. And last but not least, like this is a my contribution to the canon of marketplace analytics is like, you need to make your own data. And it's really about testing big, early and often. And when I say make your own data, what I mean is that if you do not change things, your data will not change, it will stay the same. You know, in lieu of like, you know, some crazy thing happening to you and your data changing, it will. But really what my core philosophy is like, if you're not testing, moving, making big changes, You're not going to have new data that gives you new insights, right? Just because you measure something doesn't actually mean you'll get any insight out of it. And so at the end of the day, human action that changes things creates data that is actionable and gives you insight. All right. So now you need to collect some data. And I'm going quickly here because we've got 25 minutes and I want to make tons of space for questions. You know, marketplace analytics are tricky, right? 'Cause there's a lot of different components. You got multiple sides of the marketplace, you're advertising on different networks and platforms, and you know, you gotta bring all this data together, right? And so how do you do that? What does that look like? What tools do you need? What technology is available? These, you know, especially for first-time founders, this is probably a new area, right? This is an area where you need expertise. And that is, you know, hopefully I can provide you some insight here on what good looks like. But overall, you know, setting up a really good data structure and warehousing is like critical for your success. So for example, like outdoorsy and Airbnb, you know, have to aggregate, collect, and analyze across like a bunch of different dimensions, marketing and sales, transactions and payments, listings and accounts, search, trust and safety. Website and server, the list goes on, right? And those are all different data sources. And how do you bring them together such that they mesh and can help you make actionable insights? You know, for example, I'm sure most of you use Stripe for payments, you know, great, super good data in there. But you know what it doesn't have? It's not attached to your customer on your side with all of their, you know, heuristics about them. Like what does their listing look like? Where is it located? All of those things. And so as you think about, you know, how do you do better and how do you attribute revenue to the right places so you can understand where you need to get more supply or, you know, improve liquidity, you're going to need to attach that ledger data back, right? So that's a good example of like where a lot of people have stumbling blocks to get them to really some of these metrics that we're going to talk about, like LTV and things like that. So it's a lot of moving parts. It is solvable, but I will tell you that large companies struggle with this every day. And just because you struggle with it at the early phase does not mean you're going to not struggle with it at scale. And in fact, as you get bigger, it gets harder because there's more data and there's more users. And it's very, it's a very big problem to solve generally. So, all right. And I can't see the chat, by the way. So just going in full mode here. So what is an analytics stack? I'm not going to spend a ton of time here, but I put a visual here so that, you know, you can see kind of what Drizzly has done. I always like this example, and I know quite a few companies, I won't call out any names, have kind of used very similar structures. And, you know, at the end of the day, what it really is is about taking disparate data sources and putting it into a spot Where you can ingest it and mesh it together, right? And then it can be fed into reporting tools, right? So it can be fed into places where your end users, you as the founder or CEO or executive VP, whoever you are in the company, can look at that data and make decisions on it. And so that's the end goal is how do you like normalize this data, put it in one spot and push it out so that people can use it? And yeah, there's a ton of tools out there for reporting. I won't tell you which ones I like at this moment, but there are a lot of very lightweight kind of SQL-based tools that you can use for early days, which I would recommend. They're relatively less expensive. But then as you get later on, you're going to want to go to very robust solutions that, yes, have more integration time and more expensive, But ultimately give you a lot more standardization and tooling for end users.