Subscribe
Share
Share
Embed
CRAFTED. is a show about great products and the people who make them. Top technologists reveal how they build game-changing products — and how you can, too. Honored twice by The Webby Awards as a top tech podcast, CRAFTED. is hosted by Dan Blumberg, an entrepreneur, product leader, and former public radio host. Listen to CRAFTED. to find out what it really takes to build great products and companies.
[00:00:00] Colin Nederkoorn: Could we just get a level in your voice? What did you have for breakfast today? I had Brian Johnson's nutty pudding. He's spending $2 million a year to extend his life. That's part of the equation.
[00:00:11] Dan Blumberg: That's Colin Nederkoorn, the founder and CEO of Customer.io. And not only is he trying to extend his own life with that awesome nutty pudding, but he's also built a company to make sure your customers don't churn
[00:00:23] Colin Nederkoorn: and die.
[00:00:24] We were so naive. That we just started down the path and we needed to solve these problems as they came up.
[00:00:30] Dan Blumberg: Customer.io is a customer engagement tool that can help you deliver the right message to the right customer at the right time on the right channel. They boast a several hundred million dollars valuation and thousands of customers, including big names, like Notion Onboarding is a key use case, but Collins says sometimes companies try to optimize it too soon.
[00:00:50] Colin Nederkoorn: Try to get people to talk to you. Don't try to nail your onboarding out of the gate. The data is not gonna give you the rich information that you can get from talking to people directly.
[00:01:02] Dan Blumberg: On this episode, we'll talk about great onboarding and retention and the data that powers it all, plus how customer io got its start a dozen years ago, and where they're going now as they embrace a multi-product strategy.
[00:01:15] Colin Nederkoorn: I want things to be real that don't exist.
[00:01:21] Dan Blumberg: Welcome to CRAFTED., a show about great products and the people who make them. I'm Dan Blumberg. I'm a product and growth leader, and on CRAFTED. I'm here to bring you stories of founders, makers, and innovators that reveal how they build game changing products and how you can too.
[00:01:37] Crafted is brought to you in partnership with Docker, which helps developers build, share, run, and verify applications anywhere without environment confirmation or management. More than 20 million developers worldwide use Docker's suite of development tools, services, and automations to accelerate the delivery of secure applications.
[00:01:56] Learn more at Docker.com. And CRAFTED. is produced by Modern Product Minds where my team and I can help you take a new product from zero to one and beyond. We specialize in early stage product discovery, growth, and experimentation. Learn more and sign up for the CRAFTED. newsletter@modernproductminds.com.
[00:02:17] Colin, I'd love to start at the beginning. What was it that got you into product in the first place?
[00:02:22] Colin Nederkoorn: I mean, I, I'm not a good software developer. I think that's part of it, but I want, I want things to be real. Don't exist. And when I graduated from college, I didn't see a good path into tech and so I did something completely different.
[00:02:38] I was in in sales and worked in like the petrochemical industry, but at this point in time I. Um, apple was making their switch to Intel processors and I, I created this competition because I was like, I think that technically it should be possible to run windows on these new Apple machines. But when they came out, I.
[00:03:02] The first versions, I didn't support dual booting windows, and I needed to use windows for my job. And so I created a competition because I was like, I think this should exist. I think it's possible. Let me see if I can incentivize it. And I put in a, I think it was a few hundred bucks of my own money and then raised 13,000 or so in a prize pot.
[00:03:23] And then some people figured it out. They got Windows dual booting on, on the Mac. And at that point I was like, I am in the wrong industry. Clearly. I love technology. Yeah. How do I get into technology? And so I tried to start a company at that point with some friends and moved to Boston and joined a coworking space in Boston called Beta House.
[00:03:46] And my company wasn't going anywhere, but I just started trying to help. Other companies that were in this coworking space, and through that got my first product management job, if you can really call it that. There were a group of rails developers and there was a customer they didn't like dealing with, and so they put me in between them and the customer to get the customer requirements, uhhuh and kind understand what was going on.
[00:04:12] And then, um. Then from that I got my next product management job and just kind of went down the the product management path.
[00:04:20] Brandon Kessler: Very cool.
[00:04:20] Colin Nederkoorn: Realized that I really liked it.
[00:04:22] Dan Blumberg: When you ran that competition about booting up windows on a Mac, was that on challenge post that you ran that or, or No,
[00:04:28] Colin Nederkoorn: that was pre Pre challenge
[00:04:29] Dan Blumberg: post.
[00:04:29] Pre challenge post. Yeah. Okay,
[00:04:31] Colin Nederkoorn: gotcha. I just spun up my own website and posted it to dig.com if you remember that.
[00:04:36] Dan Blumberg: I sure do. Yeah, I mean, web two 2.0. Um, I, I love that. I, I didn't realize that's Brandon, the CEO of Dev post formerly of Challenge Post when you met him and, and got the, uh, head of product job at Challenge Post, I guess at the time.
[00:04:50] I mean, I. He must have flipped when you heard the story that you had been already running your own competitions. Yeah,
[00:04:55] Colin Nederkoorn: I mean, I think that was part of the reason we were both excited to work together,
[00:04:59] Brandon Kessler: but now I started seeing these developers getting together at these amazing hackathons and there was no platform and there was no community, and so that's why I started the company at the time it was.
[00:05:11] It was called Challenge Post, which is to be a platform and community for the skill contest. Mostly engineering, but not just engineering. So that switch,
[00:05:20] Dan Blumberg: how did you identify the opportunity and and how did you come to found customer io? I
[00:05:24] Colin Nederkoorn: was working at Challenge Post now Dev Post, my co-founder was head of engineering there.
[00:05:32] So this is 2011, uh, time period. And we were really curious about the analytics space around this time. You saw companies like Mixpanel and Kissmetrics start to appear, and when we started pitching people our idea about an analytics product, the feedback that we heard was pretty consistent from a few different folks.
[00:05:56] It was, I have tons of tools that let me see what's happening in my product, but what I really want is something that helps me influence people's behavior. And so we. We heard that a few times and then realized, hey, there's, I think there's an opportunity here, and we didn't see any other products in the market that let you instrument like an analytics product and have this massive stream of data coming in, and then decide later what messages you wanted to send to your audience.
[00:06:26] Dan Blumberg: I mean, I remember that era. Well, I was, I was at the New York Times at that time and when I joined the Times I joined in 2012, the Times could send push notifications to every mobile phone all at once or, or not. So we, we, I did a lot of work implementing systems like you described. If user does X and Y but not z and I wanna ask you about the, the not parks.
[00:06:45] I know that part is especially hard. We
[00:06:47] Colin Nederkoorn: were so naive that we just did, we did it. We just started down the path and we started and, and we needed to solve these problems as they came up.
[00:06:59] Dan Blumberg: I'd love to hear more about how you built that first version, you know, signed that first customer to how you knew you really, you know, you're onto something.
[00:07:06] Colin Nederkoorn: So at this time, lean Startup was extremely popular and we. Use that methodology with a lot of our, our early decision making and, and we had a target date of April 1st. For the first day, we were gonna be full-time on the company, and our goal for April 1st was to have five customers paying us $10 a month each.
[00:07:30] And so we got that and they were paying us $10 a month really on the promise that we would deliver a product at some point. What, what did you show them to get them to say yes? I think it was like wire frames and we talked to them and. At some point, we had a JavaScript snippet that they could put on their website and see data streaming into our application.
[00:07:52] But for the first six months, a customer would tell us, Hey, I, I want this set of things to happen based on the data that's streaming in. And we'd look at the data. And either say, yeah, we can make that happen, or, no, we can't. And when we said yes, we could make that happen. My co-founder was behind the scenes writing map, reduce code to identify these needles in a haystack, and we would sort of store that map, reduce code in an admin area.
[00:08:21] And so it was a little bit of smoke and mirrors as a way to validate that the product that we were building was gonna deliver value to our customers.
[00:08:30] Dan Blumberg: I, I should actually just step back for a second as we just got right into it. Could you just describe what is customer io?
[00:08:35] Colin Nederkoorn: When we started, there was no category for what we do, and so people had a hard time putting us in a bucket.
[00:08:41] But today the, the category that we're in is called customer engagement, and the parts of our platform today are a CDP product that helps people collect. Both real time and data from their data warehouse and stream it into our platform, and then send it down to all the tools that they're using, including customer IO journeys, which is our journey orchestration and message automation platform.
[00:09:12] So within Journeys, our customers set up. Rules and triggers for messages that they wanna send over. Email, push, SMS in app. Basically every channel that they're communicating with their customers on, they set up the rules for when messages should be sent out and they can send newsletters. Transactional messages like payment confirmations, receipts, and these, um, trigger based lifecycle messages as well.
[00:09:41] And then we also have a third part of the platform, which is parcel Today it's a code editor for building emails. You can think of it as uh, vs code, but tailored for. Creating emails. Um, and so those are the three parts of our platform. Gotcha.
[00:09:58] Dan Blumberg: And you said you, you've focused at least initially, maybe still to this day on, on B2B SaaS.
[00:10:03] Can you give an example or a few examples of, of how customers are using customer io, you know, the kinds of rules that they're following to send the messages that they
[00:10:12] Colin Nederkoorn: send? Yeah, I mean, I'd say a really common example. For B2B SaaS is onboarding new people into a product. And so Notion is one of our customers.
[00:10:24] And if you, if you get invited to a Notion account at your company, you're gonna get a series of messages from them that introduce you to the product that teach you about. Different parts of the product, and depending on how you've used notion, that experience is gonna be tailored for you. Everyone doesn't get the same set of messages.
[00:10:47] It's, it's all based on the data that notion's collecting about your usage, and they're adjusting the way that people flow through campaigns and the messages that they get. And
[00:10:58] Dan Blumberg: I'm sure if you have not used a certain feature that is something that you know might trigger notification, Hey, have you, have you tried feature X?
[00:11:04] You know you'd love it. Exactly.
[00:11:07] Colin Nederkoorn: Yeah, and this, the tech is like applicable to a lot of different products. We have customers who are in healthcare, in banking, in education as well. I think online, online learning was like a really early use case for us where we saw a lot of companies who had people signing up to learn something and then they would get halfway through a course and then leave.
[00:11:34] Well. If the only tool you have is your product, how do you get people to come back when you have no way to reach out to them? And that's where customer io became really valuable for, for a lot of these companies. 'cause that was the way you got people to come back to the product. I.
[00:11:50] Dan Blumberg: I wanna talk about the rise of, of PLG or, or product led growth.
[00:11:54] I mean, the dream of of PLG is that you have a product that sells itself because of sort of the innate nature of the product. And I'm also interested in just how customer IO and your customers have sort of rode this trend over years and before even, I think they had this, this fancy acronym of PLG that's gotten so hot over the past few years.
[00:12:11] Colin Nederkoorn: Coming from product and engineering backgrounds. When we started, and we were a really small company, the way that we wanted to buy other products was to use them. And so we had a bias towards products that just let you sign up and either have a freemium version of the product or a free trial. And I think both free trial and freemium kind of get lumped into PLG today.
[00:12:34] But I, I'd say freemium is. Is primarily what, what you see as like P-L-G-P-L-G companies and we looked at the market and saw that our, our expectation was there were gonna be a lot more companies over time who didn't sell like B2B SaaS the traditional way where you would never touch the product until you sign the contract.
[00:13:00] You land on a website that gives you these like hard to understand. Explanations of how this product provides value for your business. There's no screenshots of the product. You can't see anything. There's just a form you fill out and you talk to a sales person, and then they probably give you a demo at some point in that process.
[00:13:18] And as a product person and my co-founders and engineer, like we, we wanted to understand things by touching them and using them. Um, and we felt like. The way that people were gonna buy in the future was, was more like that and less like the, the old school B2B SaaS buying process. And so we were really building, building the product, um, building customer io to, to ride that wave and support all of these new companies that were selling the way we were.
[00:13:49] So we were in, in a sense, building this product for other versions of ourselves across all of these other companies. And when we were. Two to five people in the company. We were selling to other two to five person companies. We're now 250 people in the company, and we're selling to companies that are two people just starting out all the way up to, you know, multinational, publicly traded companies with, you know, thousands of employees.
[00:14:17] The thing that ties it all together for us is that. Our customers have a, a product and they're trying to help people get up and running in that product.
[00:14:28] Dan Blumberg: Got it. When a, a potential customer comes to you and they're like, we've gotta improve onboarding, our retention sucks, are there particular things that you ask to understand, you know, where is the problem and do you also recommend ways that they use the platform?
[00:14:42] Or is it very agnostic?
[00:14:43] Colin Nederkoorn: So our decision was to build a very flexible platform. So our customers come to us and whatever they wanna do, they can accomplish in the product. We, we talk about it internally as, as like, dream it, do it, they dream something up and then they can get it done in, in the product.
[00:15:01] That said, in different stages of companies, there's better and worse ways to, to onboard people. I think the, the advice that I would have for folks starting up either with a new product or a new feature in a product is. Try to get people to talk to you. Don't try to nail your onboarding out of the gate because you don't understand enough in order to like really create a, a great onboarding and you can instrument your app all you want, but the data is not gonna give you the rich information that you can get from talking to people directly.
[00:15:40] Just really trigger outreach to people to prompt a response. And one of the things that you've probably seen out in the wild, we were encouraging this early on is like an automated message from the CEO or the product manager of a feature Mm-Hmm. Saying, Hey, notice you started using this feature. I am.
[00:15:59] I'm Colin. I'm the CEO of customer io. You know, it seems like you got stuck here. I would love to know more about your experience. Yeah.
[00:16:07] Dan Blumberg: And it comes from their first name too. You're just making me realize, I, I get the, the emails from Ivan at Notion. Right. That's, that's you guys. Yeah. And what is it about that that, I mean, I think I can guess like it's a, it's a person's name.
[00:16:18] It's in my inbox, my inbox. It's kind of a personal place, you know? It's, it's like that's, that's where Inbox is a very interesting place. You have marketing stuff, you have your friends, you've got everything. Is that, is that really what makes those stand out?
[00:16:29] Colin Nederkoorn: I think so. And it's, there's not an illusion there that.
[00:16:34] Oh, Ivan actually wrote to me, it's you. You probably realize it's an automated message, but it's comfortable and it's familiar enough that if you wanted to respond, that would be okay.
[00:16:49] Dan Blumberg: Yeah, well, it it, but it, it would've when you founded customer io, I mean, I remember working again at the New York Times. I worked on newsletters there, which were all broadcast at the time, and you couldn't hit reply and start a conversation.
[00:17:02] And now, yeah, now they're very, you know, not just the times all over the place. They're substack, they're all very voicey and you can hit reply. In fact, one of the funniest things about having a newsletter is you get people's autoresponders and you can, you like learn a ton just from that alone. Yeah, absolutely.
[00:17:16] Tell me about
[00:17:17] Colin Nederkoorn: some of the biggest challenges, uh, at customer io today. The thing that I'm thinking about a lot is. Man, coming up to a year ago, we made the, the shift from being a single product where we were called customer io and we only had one product, and the product and the company were the same to launching customer io data pipelines, which is our CDP offering, and then having that work alongside customer IO journeys and then.
[00:17:48] We had acquired a company parcel and more recently we, we redesigned the brand. We launched the new brand and new website, and now we have, we sort of. Tell the world, we've got these three products. And so the big, the big challenge for us has been thinking about that shift from a single product to a multi-product company.
[00:18:09] How do users go from one product to another? How do they all work together and how do we price and package the different offerings? Initially, the, the goal that I had was someone could. Enter our ecosystem through any one of these products and use it totally standalone. And what that did was it created a lot of confusion for.
[00:18:37] The sales team and marketing team, and so we're in the process of shifting to talking about ourselves as a, as a platform, and these are pieces of the platform rather than, these are standalone products that happen to work together. I'd say that's the biggest change we're undergoing right now is sort of figuring out all of that.
[00:18:57] Each of the products have. Users and customers of those products who really love them. And so they're serving a need there. And just figuring out how we do that so that users of the platform, people who are using the platform, are using and loving each of the pieces of the platform. 'cause uh, our hypothesis is that when you use them altogether, it's a really powerful combination of tools.
[00:19:22] Dan Blumberg: I'm interested in, uh, you know, you guys are customer io, uh, you haven't changed your name to customer AI just yet. Mm-Hmm. I'm kind of shocked by that. I'm interested in your, your thoughts. I, you've built like a, as you put it to me before, like a rules engine that really works. Uh, and that's great. Honestly, in some ways that's very refreshing.
[00:19:41] Like, do, do you need to do things with
[00:19:43] Colin Nederkoorn: ai? We're doing some stuff already. There's a lot of companies are jumping on the AI bandwagon and just pumping out features, uh, as fast as they can. And I feel some urgency there. I use chat GPT every day and, um, I test like everything that, that that comes out. And I think like we approach things from a very customer centric point of view.
[00:20:07] And when I think about. The jobs that our customers are trying to do in the product, we think about like, how can AI help them do them? And I don't see a world where we totally remove our customers and say, alright, you, you just press a button and AI can do everything that you need to do for. Sending messages to your audience.
[00:20:30] What makes the most sense to me is about augmenting humans and streamlining and saving them a bunch of time. So when we've looked at things like, um, generating content, for example, I think AI can spit out a bunch of content really well, but it's been tricky in my experience to actually get AI to write as well as I can.
[00:20:57] Express me better than I can. And I imagine other folks who have tried it have had the same experience. And so we haven't released a AI based like content generator. And that might be helpful for some folks, but it also creates extra work to then go and edit and remove. And oftentimes it's a lot easier to just start with your thoughts from, from scratch in the in the blank page.
[00:21:21] We have an AI feature out there to help people write sql. And so one of the core principles in our, in our product is that we wanna give our customers a lot of flexibility. And in the data pipelines product, the typical user there is a, is someone on the data team. So they're, they're comfortable with sql, but we're using an LLM to help you write valid sql and that helps save them a bunch of time.
[00:21:47] And so I think there's. Surgical ways we're using AI today, longer term, the way that we're thinking about it is that it should be invisible to our user and not like called out as whatever customer ai, right? Yeah, yeah, yeah. It's, it should just be in, it should be there behind the scenes helping you, giving you superpowers.
[00:22:09] It's executing your will more efficiently and faster than. If you had to like tweak all the dials yourself. Totally.
[00:22:20] Dan Blumberg: Yeah.
[00:22:23] Tell me more about when you were approaching companies, maybe, maybe this is less of a thing now, but in the early days you were like, I want you to send me all of your data because then I can trigger these things when X did not happen. Was that a hard sale for a lot of our customers? It was less
[00:22:38] Colin Nederkoorn: work to Oh, just just like, here, take it
[00:22:40] Dan Blumberg: off.
[00:22:41] Colin Nederkoorn: Yeah, exactly. Um, and a lot of. Other products at the time. So if you were trying to send the amount of data that you were sending to us to Salesforce or HubSpot, like they had like severe restrictions on. How often you can call the API, and we built our infrastructure in a way that receiving data could scale really, really well.
[00:23:05] And so we just said to customers like, Hey, send it all, you know, send as much as you want. And yeah, there there's a certain scale at which we say, all right, maybe not that much. We had companies who had mobile apps that were like, pretty widely distributed and they were sending us every swipe, every tap, every, you know, every interaction that people were making on the screen.
[00:23:26] And every time they do that, they send this massive, massive bundle of JSON data. Mm-Hmm. Um, for every one of those. And it was, that was a lot, but surprisingly it wasn't as. Big of a hurdle and as I said, like people were really familiar with the analytics products of the time and the API that we had was very similar to what they were already used to with analytics products.
[00:23:50] So it was kind of like, Hey, where you're calling this in your code, add another call to customer io.
[00:23:55] Dan Blumberg: Got it. And, and in terms of building data centers and pipelines, that scale, uh, was that a very conscious choice or was that just, you're just building on modern infrastructure and its cloud base and so it, it, it scales.
[00:24:07] I'm, I'm dramatically oversimplifying a thing, but how hard was it to create something that could scale, to take in the, you know, millions and billions of swipes and taps that were being sent your way?
[00:24:18] Colin Nederkoorn: Over the years, the ceiling on what the infrastructure can take keeps going up, and that's really hard.
[00:24:26] I'd say that's, that's probably the biggest challenge at the company is how do you keep raising the ceiling on how big of a customer can we support? How many profiles? Yeah. Across our customer base. We're in the billions of profiles in, in customer io now and, and with. These activity streams. Some of our customers have terabytes of data that they store with us.
[00:24:49] I think that's the biggest challenge with with
[00:24:51] Dan Blumberg: the product. Part of the reason I'm asking these questions is, as AI has risen in importance in the last, you know, year and a half or two, like the robustness of companies, you know, data pipelines and the, and just the, the actual cost, the server costs of, you know, how many tokens am I using today?
[00:25:06] Like, is, you have a lot of experience here, even if it's not in, in the generative AI space. Like, just in terms of managing the, the cost of storing and, you know, processing all this data. So I'm, I'm wondering if there's more advice that you care to share.
[00:25:18] Colin Nederkoorn: Yeah, I mean, I think what we've observed is that people will store a lot of data.
[00:25:22] The majority of it is not use useful, and one of the trends that we've seen as well is, is the move to data warehouses. When we started, a lot of companies didn't have all their data in a data warehouse. Now people are pushing everything into a data warehouse. The majority of it isn't useful, but you can do great things in a data warehouse, like combine multiple data sources and then build a useful set of data that you can.
[00:25:47] That, that's then enriched and basically make it a lot more useful and valuable in generative ai. And that's one of the things that we can help customers with through customer o data pipelines is that you have this like stream of data coming in, you can transform it and then send it downstream to your data warehouse.
[00:26:04] And it definitely takes some work. And the, the more data you have, the harder it is. To massage it and get the useful stuff. And I think it's a big, it's a big challenge that a lot of companies are gonna, are gonna have. 'cause obviously if you're not efficient with your token usage, you'll pay a massive, a massive bill.
[00:26:25] Dan Blumberg: Yeah.
[00:26:26] Colin Nederkoorn: Colin, thank you so much. I really enjoyed this. Thanks,
[00:26:29] Dan Blumberg: Dan. That's Colin Nederkoorn. I'm Dan Blumberg, and this is CRAFTED. CRAFTED. is brought to you in partnership with Docker, which helps developers build, share, run, and verify applications anywhere without environment confirmation or management.
[00:26:45] More than 20 million developers worldwide use Docker suite of development, tools, services, and automations to accelerate the delivery of secure applications. Learn more at Docker.com.
[00:27:00] Special thanks to Artium where I launched CRAFTED. Artium is a next generation software development consultancy that combines elite human craftsmanship and artificial intelligence. See how R Dium can help you build your future@artdium.ai.
And CRAFTED. is produced by Modern Product Minds where my team and I can help you take a new product from zero to one and beyond.
[00:27:23] We specialize in early stage product discovery, growth, and experimentation. Learn more and sign up for the Crafted newsletter at modernproductminds.com. Please rate and review CRAFTED. and share the show with your product making Friends.
[00:27:39] Colin Nederkoorn: There's no right answers here. There's just different ways of doing it.