📊 Analytics for Cloudflare's Self Serve Business

Welcome, everyone. Thank you for tuning in. It's BI week, and today we're going to be talking about analytics for an exciting and growing part of Cloudflare's business, the self-serve business. We're going to be talking about what the business is all about, how does the product growth team and the business intelligence team work together, and also talk about some interesting analytics that we've been collaborating on to really, you know, help drive business growth. And I'm super excited to discuss all of this with you guys, Philip and Apoorv. So let's start off with some introductions. Philip, do you want to go first? Yeah, sure. Hey, my name is Philip Johnson. I am part of the self-serve or pay-go team, as we call it internally, and I work on growth. So trying to figure out how to grow our self -serve business. Hey, everyone. My name is Apoorv, and I'm a data analyst at Cloudflare. I've been here for almost two years now, and it's been an exciting journey so far. So I work as a data analyst here, and we work collaboratively with different teams here, like Philip on the pay-go team. It's been an exciting journey so far. Awesome. Yeah, I'm Nehit. I'm also a data analyst. I'm a part of the business intelligence team here at Cloudflare. I've been with Cloudflare for about eight months now, and it's been a super exciting journey so far. I love collaborating with all the different teams that we have, and I'm based out of the SFO office. Awesome. So let's jump right into it. Philip, can you start us off with an overview of what exactly our self-serve business is? What is sort of like the self -serve component of it? How does that work? Yeah, definitely. So at Cloudflare, we kind of separate our business into two main groups, right? We have the enterprise or contract business. Those are your big logos who typically come to Cloudflare .com, usually fill out some sort of a lead form and go through a lengthy sales process, at the end of which would sign up for a really custom long-term contract. And then we've got the self-serve side of the business. And so that's basically anyone who wants, you can come to Cloudflare.com, create an account with a credit card, and then select one of our self-serve plans, which range from free plans all the way up to 200 bucks a month. Right. Awesome. So which customers are we hoping will join us and sign up with us as part of this business? Who is this business sort of intended for? Yeah, I think the short answer is anyone who wants to make their online application or website faster and more secure or reliable. But typically, what we see is we see a lot of developers, sole openers, small businesses, startups, things like that, all the way up to Fortune 1000 companies actually can start in kind of our self-serve realm. And some of those folks will end up graduating up into the enterprise or contract business. Awesome. Philip, so we understand our self-serve business a little. So can you explain how this self-serve business model differs from other parts of our business, like enterprise? Yeah, absolutely. So I think the main difference is that Pay-As-You-Go doesn't rely on a sales team to grow, whereas the enterprise team really does, right? So instead of relying on kind of like the super established marketing and sales org that generates leads and those sales team kind of feels those leads and eventually converts those folks into paying accounts, hopefully, we kind of generate our demand through other means. And so what that means is, you know, it could be paid advertising, content, partnerships, working on various different product improvements, and things like that. So the main difference is we don't rely on a large sales team to kind of generate demand for us. Got it. So if I understand, our Pay-As-You-Go business ranges from a simple person to maybe even organizations. So does Cloudflare offer some sort of pricing and plans to category or bucket this set of customers? Yeah, definitely. So we kind of have a hybrid pricing model. And what that means is we have certain plans and products that are built at a flat fee at some interval, usually monthly. And then we also have add -ons and usage-based products that are completely dependent on how much the customer used of that product. So a good example would be Argo, for instance, which is a product that helps speed up your website. But our customers are essentially free to kind of choose one of three base plans. And that's usually how they get started. Those base plans range from, you know, free, which is absolutely nothing. It's getting a ton of value and trying to get as many folks kind of onto the Cloudflare network as we possibly can. A step up from that is the pro plan, which starts at $20 a month. And then the level up from there is the business plan, which is around $200 a month. Then once you have one of those base plans, you don't have to, there's certain plans that don't require you to have one of these base plans. You can select any number of different add-ons like Argo, which I mentioned, or rate limiting, or workers, which are all kind of dependent on how much you use. Nice. So our teams have been working closely together for the past year, partnering on projects to really help fuel the growth of the business. I, for one, have really enjoyed working with the team. You guys are amazing to work with. And getting this fantastic opportunity really to observe a growing business and being able to collaborate with you guys. As a team, we've been able to come up with super interesting insights that we've been able to discover that contribute to the business. So from your perspective, can you tell us a little bit about how this partnership sort of came to be? How do our teams really work together? What's the synergy really like? Yeah, definitely. So I come from a startup background. So believe it or not, this is the largest company I've ever worked for. And prior to working here, you know, we really had to kind of like self-serve our own BI needs. And as a result, you know, I ended up having to learn some SQL and having to do pretty hacky stuff to kind of get answers to super important questions. But one of the great benefits of working at Cloudflare is having access to an awesome BI team with Nahid and Apuv. And just in general, especially for a growth or any type of a product team, you know, data is fundamental. And usually data and analytics is the best way to kind of inform, validate, as well as kind of keep track of our growth strategy as it evolves. And working with you guys and the BI team is the best way to get the data we need to be successful. So that's kind of, I would say, kind of the genesis of the partnership, if you will. Nice. Awesome. And then Apuv, you know, since we're on the same team, can you also share your experience of working with like the amazing sets of business team? How does the partnership work from sort of our end, like the BI end? Yeah, sure. Absolutely. Like I said, I mean, it's been a great collaboration effort so far. I think the good part here is like we are both on the same side of the business. I mean, how it differs from a consultant side of a business is you get a task, you work on something and you move on. It's not here like that. We work on something, we iterate over it, we learn from it, and we further improve it to get more insights and get into a deeper thought process. So I think that is the best part that I like about our job here is that we continuously are working and learning from whatever we develop, be it a dashboard, be it any insights. And we further iterate upon to bring it to a more deeper level and more in-situ so that, you know, ultimate goal is to, you know, help our customers and grow our business. So that's, I think, the overall motive for even it's being a growth strategist or being a data analyst here. Yep. Yeah, no, I definitely agree. I think that continuous improvement that we constantly sort of think about and help to make the product and the business better is definitely one of the best parts about working here. Yeah. Continuing on to that, Philip, what do you think, I mean, what are some of the metrics that we look for to measure the financial health of our business and, you know, why these metrics are really important? Yeah, I mean, I think the really short to the long answer is that ultimately the self-serve team or the PAYGO team is accountable for various different revenue targets. So it's really, really important that we understand the relationship between the things that we're doing and the things that are happening kind of across the funnel as it relates to the dollars and, you know, billings and revenue that we can actually book. And so, for instance, if I really zoom out and think about, you know, how does Cloudflare self-serve grow, you know, you can use a pretty traditional kind of like funnel model to easily explain that, which also ties into the financials, right? So, you know, typically you have the acquisition stays. So actually how, like through which different channels are we acquiring users? That could be through our organic content, like the blog that we write every day, great post that came out this morning, by the way, about APOs, so just check it out, you know, to activation. So how quickly and how long does it take from a user from that particular channel to actually activate and get some sort of value from Cloudflare to, you know, and that can happen in a free capacity or a paid capacity, right? I don't need to pay Cloudflare any money to actually receive any value, which I think is one of the great things about Cloudflare. And then the next phase would be revenue. So how are those users of Cloudflare that could be free actually monetizing into paying Cloudflare a little bit of money? And how does that differ across, you know, all these different factors across that user or business as well as the different marketing channels that they came through, right? So how does a, let's say someone who came through and one of our paid ad channels versus organic content channels kind of behave. And then we look at things like retention. So, you know, after someone signs up and starts paying us some money in exchange for services, how long are they actually going to keep paying us money at that same amount? And then the next one is kind of referral. So out of all those folks who have come into Cloudflare and become customers or users, how often are they kind of referring their friends and contacts into the business? So I think that's how we look at it. And those pretty loosely align with, you know, those key financial KPIs we look at, which I don't know if you want to kind of take the financial side and go through the financial categories. Yeah. And I totally agree. I mean, you mentioned about retention, I think for any SaaS-based line of business, growth and retention are some of the metrics that, you know, any organization would like to measure. So what kind of metrics at Cloudflare's sensor side of business we look for to measure and assess these performances? Yeah. So I think it totally varies across the different, you know, different steps of the funnel. So I think, you know, from an acquisition standpoint, some of the things I might look at, let's say it's a paid ad channel, is, you know, how much money are we spending every month? How many impressions does that spend generate? So how many eyeballs can I actually get in front of for every dollar that I spend? For everyone who sees it, how many clicks do I get? So what's the click-through rate from impression to someone clicking through to the landing page? And then as they get to the landing page, I'm also curious, you know, about things like how long or what's the bounce rate? So how often do people just bounce right off the page? What's the average session duration? You know, clicks to checkout, which is also kind of a form of click-through rate. And then as you get down into the revenue, you know, portion of the funnel, looking at, you know, what is the rate of free plans upgrading to paid plans? As you get into kind of like retention, we're looking at folks who are, you know, how many customers are spending more money with us this month than they did last month or less than they did the month before. And getting into kind of retention or churn is, let's say, let's take a look at all the customers that we signed up in, let's say, January of 2021. And on average, how well are those folks retaining? And so we use, thanks to you guys, like this awesome type of cohort analysis. It allows us to kind of measure how different cohorts, depending on the month or day or week that we acquired them, are actually performing over time. So are we actually getting better at retaining our customers? So I think those are some of the different things that we look at. And obviously there's a ton to dig into there, but yeah, that's a high level. Yeah. Awesome. So yeah, in addition to the growth and retention that Apoorv mentioned, and Philip, you touched upon this a little bit already, is acquisition, right? Which is another key driver of the business. And then that's basically measuring how effectively we're actually bringing new customers onto our business so that they can join us. Like you mentioned, sign up for our free plans and then our paid plans as well. So I want to discuss a little bit about another exciting project that sort of looks at acquisition that we're currently collaborating on, which is marketing acquisition analytics. It's really going to help us see sort of how we acquire our customers online, which channels do they come from? You touched a little bit upon like paid versus organic. And then what is really the efficacy of these channels, right? Like how are customers flowing through them and actually coming onto our site? So can you talk a little bit about why it's important to measure this stuff for our business? Yeah, for sure. So I think like going back to our goals as a team, right? So we have to hit certain growth targets in order to keep the business on track. And also being a team of limited resources, today we're a pretty small new team. We have to be super careful about where we're spending our time, right? And doing these type of deep dive acquisition analytics really helps us figure out where we should be spending our time, right? So let's zoom out and talk about the different types of channels, for instance. So we've got at a high level organic channel, so channels that people come through where Cloudflare is not spending any money to kind of generate that traffic. And then you've got paid channels, which are things like paid ads or podcast ads, television ads, affiliate relationships, stuff like that. And so each one of those channels behaves differently. And one thing that we're really concerned about are like, hey, how can we find kind of the most efficient growth levers, right? So for every unit of effort that our team puts in, like how much comes out the other side? And so that is essentially what we're trying to understand. Not only will that help us kind of optimize the way we allocate resources internally in terms of where we spend our time and developing channels. But it also helps us understand how expensive is it to acquire a customer through channel A versus channel B? And once that customer comes through the channel, how long do they actually stay on board? Because it could be the case that folks from channel A might be a lot cheaper to acquire initially, but maybe they retain a lot more poorly than folks from channel B. So maybe it's actually worth spending the additional money to acquire users from a healthier channel versus an unhealthy channel. But in order to kind of make those decisions and tradeoffs, we really need a ton of data and analysis. And so I think that's really what we're getting at when we're talking about deep acquisition level analytics. Got it. Nice. And then one of the things that you also touched upon was basically how expensive is it, right? And that sort of ties back to ad cost, which is how much we're actually spending on ads. And then that also ties to another metric, which is like what you mentioned, the cost of acquisition, the CAC that we usually measure. So could you talk a little bit about how the ad cost really ties in and what sort of insights does that give us? Yeah, sure. I mean, I think really one of the top level metrics we look at is CAC or customer acquisition cost. And we're really looking for that in healthier channels. We want the cost of acquisition to be less than the lifetime value of that customer. So, like, let's say we're acquiring folks through channel A and we're acquiring them at, like, let's say, $100 a user. If they're never going to spend $100 at Cloudflare, then that means we're losing money on each one of those acquisitions. Right. And so what we're trying to do is, one, figure out which channels we can acquire users through profitably and really quickly find out, you know, which ones warrant further investment or further optimization to try to actually drive that LTV to CAC ratio in the right direction. And then for channels that don't work, it's good to know so that we can kind of turn over that rock and say, hey, this was a good experiment. But now we know this is maybe not the place to look for customers. Let's move on to the next one. And so, you know, channels that show early, early promise, we might actually go out and make hiring decisions or decisions around budgeting around those types of insights. So it's super, super key for us. I think another interesting thing to look at is, you know, I'm talking about lifetime value. So something that we care about is not only, you know, how much money does someone spend on the first day that we acquire them, but, you know, how much money do they spend over the course of their lifetime? And even better is being able to predict to a high degree, you know, within a reasonable degree of certainty, how much money someone is going to, you know, contribute over their lifetime to Cloudflare, which also changes our calculus in terms of how we might decide to bid, you know, on certain ad networks or kind of allocate budget. I know we're going to talk a little bit more about that later. Yes, for sure. Awesome. Yeah. And then one of the aspects that really constantly wow me at Cloudflare, like in the time that I've been here is just the number of products that we have, right? We have a huge breadth of products and every product sort of like a unique purpose. And so as a consequence, not every product behaves the same way from an analytics perspective. So, for example, we have Registrar, which is what people can use to sort of register their domains with us, versus we have usage-based products like you mentioned Argo before. And then we have subscription-based products, which is all part of the PayPal business or the Salesforce business. Can you shed a little light on how the product team sort of handles this from their end of like having so many products? Yeah, I mean, I think it's a challenge, right, to be honest. It's a challenge to understand how people are actually growing at Cloudflare through the product choices they're making. And I think from a product team's perspective, it's a challenge to figure out how to, you know, build an experience that helps the user not only realize, but also realize when there's new value that might be able to be uncovered through other products and services that we offer. So I think probably, I mean, I don't, this is a total guess, but I want to say we have probably over 20 different products that you can pay for. And so in most SaaS companies, you might have a single product with three different tiers. But here, I think it introduces a new level of difficulty in terms of getting, you know, especially really good data around, you know, how does a free customer become a paying customer? And after they're a paying customer, what product usage decisions do they make to kind of expand or contract their business? And how does that differ across various different characteristics of that user, right? So, for instance, do startups take, you know, path A and typically upgrade through certain products versus maybe larger, more established companies? And I think all those different types of relationships are super important for not only, you know, the pay go or self -serve team to understand, but also the product teams to understand because it's going to really influence the way that they decide to build certain products, as well as the overall, you know, Cloudflare experience. And I think even another layer of difficulty on top of that is catering to both our enterprise customers and our self-serve customers, right? Because our enterprise customers have super different needs and a little bit of a different experience than some of our self-serve customers do. So, I think, like, you know, Cloudflare has a really unique challenge ahead of it. But I think we've done an amazing job kind of chipping away at that, at least since I've been here. It's been great to watch how people kind of navigate those unique challenges, for sure. Nice. Totally agree. I mean, the more the products you have, the more challenges comes. But at the same time, there's more learning curve that you get. I think that's the best part of working here, that you constantly have work and more innovations to perform. And I think that's the kind of a collaboration I think we all will agree that we love and we like to do here. I think we've touched upon a lot of points here, like what we have done so far, how we are doing it, and we are able to grow our business. But, Philip, what are some of the factors on some of the future plans that you feel are there in the pipeline that BI can contribute to grow the sense of business even more? Yeah, absolutely. So, going kind of back to our funnel of acquiring customers and getting them to revenue and eventually figuring out how to get them to spend more and retaining them longer. I think we've done the high-level work, right, to understand how is the Cloudflare self -serve business growing. And we spoke with Nahid a little bit more about really understanding at a higher level what's going on in terms of customer acquisition. And so I think there's a lot more that we can actually do to double-click on these other funnel stages to really understand, you know, like what are the main… You know, we have a pretty good understanding of it today, obviously, but, you know, what's the kind of next level of understanding in terms of how to retain customers a little bit better? You know, can we predict with greater accuracy, greater precision when someone is likely to maybe start looking for an alternative option or something like that? So I think that's one area of opportunity is just really kind of double-clicking into those funnels and helping us forecast behavior within those a little bit better. Does that make sense? Yeah, definitely. And I think you touched upon this a little bit before as well, which is like the concept of lifetime value, right? Which is really once they sign up with us, how much are they actually going to spend with us over their lifetime that they're, you know, that they sort of be at Cloudflare? So how does, like, LTV, like you mentioned, is one of the ways we can use it as like the CSE to LTV ratio, which is like how much you're spending versus how much we are actually, you know, they're going to spend with us. Are there any other ways where LTV becomes super useful for like seeing how customers really drive the business? Yeah, definitely. I think the most obvious one to me right now is like, let's talk about paid acquisition. So like I was saying before, you know, on its face, it might seem like we should always acquire customers through a channel, which would cost $5, let's say, to acquire a customer versus a channel that costs $100 to acquire a customer, right? Like, you know, that's 20 times worse. So why wouldn't we just keep acquiring customers through channel A? But I think if you zoom out a little bit, you know, you could really, over time, you might notice super extreme differences between, you know, users who you acquire through a given channel, right? So let's say that $5 customer only ends up spending $20 over the course of their lifetime, right? Versus the customer that we're acquiring at $100, maybe they spend $10,000 over the course of their life. And so if you're really optimizing your behavior, or your decisions around just on its face, how much does it cost to acquire customers, I think it's totally fine when you're starting out. But over time, I think as you mature as a growth organization, the closer you can get to kind of the customer lifetime value and actually changing your calculus based on that, I think the more successful and the faster growth you're going to experience as a business. So I think it's super, super important not only to get a kind of a good idea of how valuable these customers are from a lifetime perspective, as close to the point of acquisition as possible. It's also important to understand, like, what are, you know, how do we, once we acquire those people, how can we actually show them other products that are going to be uniquely valuable to them? And so we can actually realize more LTV over time as well. Got it. Makes sense. And then also, like, constantly running, re-running that calculation, right? So things are going to change over time. And so what's happening in January of 2021 could be totally different than what's happening in, let's say, June of 2021. So, you know, I think this ongoing partnership is going to allow us to kind of constantly iterate, as you said, approve and figure that stuff out, which is great. Yeah, I think knowing our customer base, how much they have spent throughout their lifetime and how they might spend, I think so far we have done a good job. But I think going forward, like including some machine learning modeling to get these insights would be really helpful. We already know, okay, you know, we can define a cohort of customers who will really spend that lifetime value more than enough. And so, you know, that is a profitable business. And I think that would give us really good insights into these businesses and what paid acquisition channels we could, you know, foresee in the future. So I think so far so good. That's awesome. Yeah, I'm definitely looking forward to, you know, our future collaborations. They hold a lot of potential and promise and it's going to be awesome to continue working with you guys. Yeah, totally agree. Awesome. Likewise, we couldn't exist without you. So we appreciate BI very much and we're excited to be a part of BI week this week. So congrats. Awesome. Thank you so much, Philip. Thank you for joining us. Cool. And perfect. And I think with that, I think we are done. That was amazing. Thank you so much, Philip and Apurv, for joining me today. I had a blast talking about the business with you guys. This was awesome. And thank you, everyone, for tuning in. Take care and stay safe out there. Likewise. Thanks, guys. Bye.