On Finding Product Market Fit at Segment — Peter Reinhardt (YC S11)
On Finding Product Market Fit at Segment — Peter Reinhardt (YC S11)
1:30 - Segment’s first customers
10:30 - Launching Analytics.js
19:15 - Evan Farrell asks - You mentioned in the SS lecture that you had to totally pivot to Analytics.js to find PMF, is it possible to purely iterate on something people kinda like to find PMF, or should it be clear from the outset if a new idea is something people want?
20:30 - The importance of having a skeptic on your team
28:00 - Idea generation
Craig Cannon [00:00] - Hey , how's it going? This is Craig Cannon and you're listening to Y Combinator 's podcast . Today's episode is with Peter Reinhardt. Peter's co-founder and CEO of Segment . Segment helps companies capture data from every customer touchpoint and send it to the tools where it can be used most effectively. They were part of the YC summer 2011 batch. You can find Peter on Twitter at @reinpk. All right, here we go. The average person probably doesn't know what Segment is. Could you explain?
Peter Reinhardt [00:31] - For sure. Segment helps companies give their customers a better customer experience. We do that by helping them organize all of their internal data about all their interactions with the customer. For example, if you go to the bank, they interact with you at the ATM, at the teller, via a phone call center. They have a web app, a mobile app, they send you emails. They're interacting with you across this huge surface area. They need to be able to coordinate that interaction. They need to know that if you encountered an error on the ATM, the teller needs to be able to say, "I'm so sorry you encountered an error, love to be able to help you." What we do is, we help sort of bridge that gap of having a single record of all those interactions with each customer.
Craig Cannon [01:10] - Because previously companies would build all this in-house or not at all, maybe.
Peter Reinhardt [01:15] - There's sort of two worlds. One is they would build it all in-house, exactly and there'd be a rat's nest of data pipelines from one place to another. The engineering team would spend all their time building these data pipelines rather than actually building things for the customer. That's one world. The other world is one where it really used to be a one-on-one relationship with a bank branch manager, for example, and they might keep the information in a CRM. If you are somewhere would be an optometrist, you go in, they have your past orders, et cetera, but now the world is moving much more to a Warby Parker kind of world where you're not interacting with a person, so a CRM doesn't even make sense. It's not the right technology for understanding what interactions you're having with a customer and instead, it's all these different digital channels. That's where we come in.
Craig Cannon [01:57] - Your first customers, were they large-scale companies like banks or who did you get in the beginning?
Craig Cannon [02:11] - To be fair, it was kind of like the long road 'til overnight success--
Peter Reinhardt [02:15] - Oh, yeah, no, no. Let's not skip the year and a half of dark times.
Craig Cannon [02:18] - Yeah, yeah, we shouldn't totally go over that. Anyway, you blew up overnight on HN.
Peter Reinhardt [02:24] - Once it went live on HN, it blew up overnight. Our first customers were the folks hanging out on Hacker News . It was basically small companies for the most part. Founders who were looking for better ways to instrument their web applications and mobile apps with this sort of analytics tracking.
Peter Reinhardt [02:50] - The strange thing about the open-source library which is sort of this data router. You put in one piece of data, customer data X, and then we turn around and transform it and fan it out to all the different places downstream. What's funny about the library is, if you want to turn on a new tool , if you want to send the data to a new place, you need to recompile the library and redeploy it to your website.
Craig Cannon [03:06] - All right.
Peter Reinhardt [03:23] - No, no, this is completely accidental.
Craig Cannon [03:24] - Really?
Peter Reinhardt [03:33] - We actually applied as a classroom lecture tool . The idea was to give students this button to push to say, "I'm confused." The professor would get this graph over time of how confused the students were. We thought it was a really cool idea. We were college students at the time and we had a bunch of professors who were excited about it at MIT and elsewhere. I'll never forget, in our YC interview we were pitching this and PG was getting pretty excited. Then he turns to Professor Miller from MIT who we had talked to and had given us the original idea for it and says, "Hey, would you use this?" Professor Miller says, "Nah, probably not." We're like, "What the hell?" But we just rolled with the punches and said, "Yeah, well, you know, we talked to like 20 other professors and they were all excited about it."
Peter Reinhardt [04:21] - Right!
Craig Cannon [04:21] - You were building it out right?
Peter Reinhardt [04:23] - We went through the whole YC with this idea. Built it out. Hundreds of thousands of lines of code. Super-complicated classroom lecture tool product . It had presentation view and people could ask questions. It was very complicated. We actually even raised money at Demo Day with this idea. About 600K. Finally we deployed it in the classroom as the fall semester got started after Demo Day. It was just a total disaster. All the students opened their laptops and they all went straight to Facebook . The way we discovered this is we were standing in the black of the classroom and just counting laptop screens. We'd be looking over the shoulders of the students and going one, two, three and we discovered at the beginning of class, about 60% of the students were on Facebook and by the end about 80% were on Facebook . Oh my God.
Peter Reinhardt [05:05] - That's right and the professor , at the beginning of class, had been like, "Can everyone please get out their laptops? We're going to use this new..." And all of sudden, all these students are distracted by Facebook . We had accidentally sort of put in an attention bomb, if you will.
Craig Cannon [05:18] - You hadn't brought on any users during Y... You were in, wait, I wrote it, you were in the summer 2011 batch.
Peter Reinhardt [05:24] - That's right. Summer 2011.
Craig Cannon [05:25] - And you weren't testing during the batch?
Peter Reinhardt [05:28] - We tried testing, but there's not that many classrooms that are in session during the summer, right? The school year starts in September. We had beta tested in a few summer computer science classes at both Stanford and Berkeley, but there was always technical difficulties and other things that sort of prevented us from getting a real sense of what was happening. Oftentimes it was pretty YOLO. I don't know. It's just we didn't really get tests rolling until the fall.
Peter Reinhardt [05:56] - Standing in the back of a BU classroom. It was an anthropology class. And I remember arriving at the 60% and the 80% and we went up and apologized to the professor and walked out.
Craig Cannon [06:08] - At that point, you're just like, "Okay, we got to kind of shut this down or figure out what to do with the money."
Peter Reinhardt [06:13] - That's right. We had just gotten wires for these checks, for this money , literally a week before. We called back all the investors and we were like, "Well, it turns out this is a terrible idea. So what do you want us to do with the money?" Almost all of them said, well, we invested for the team so go find something else.
Craig Cannon [06:30] - Okay, next day.
Peter Reinhardt [06:33] - Next day we were like, "Well what are we going to do?" We realized, we should have been able to figure out some of this analysis by not just standing in the back of the classroom. Like, we should have been able to see some of this digitally. You couldn't see it in the analytics metrics at all. We decided, hey , let's build an analytics tool . Let's build a better web analytics, mobile web app analytics tool to compete with Mixpanel and Google Analytics . The idea was to give really advanced segmentation because we also wanted to understand how some computer science classrooms at MIT were using it differently than anthropology classes at BU, et cetera. We couldn't do that analysis in the tools we had. So that was the idea, it was to build an analytics tool . We spent about a year building the infrastructure necessary to do the analytics and really were not succeeding in getting any customers during that timeframe.
Craig Cannon [07:20] - Did you have a whole onboarding process? Was it a landing page? What did it look like?
Peter Reinhardt [07:24] - Oh yeah, we had a landing page. I was going on little sales trips. I was meeting with people trying to--
Craig Cannon [07:28] - Oh yeah, so you were trying.
Peter Reinhardt [07:29] - Oh yeah, yeah, we were trying. But it was not going well. Basically what would happen is people would say, "Well, I already have this other analytics tool installed so, yeah, actually I'm not that interested."
Peter Reinhardt [07:43] - No, no, no.
Craig Cannon [07:46] - Were you at parity?
Peter Reinhardt [07:48] - We were at parity in certain dimensions. We were exceeded in other dimensions. But they just weren't the dimensions that mattered, apparently.
Peter Reinhardt [08:01] - Well, so, pause. This is just the analytics tool which we realized, in December 2012, was, failed. We were well over a year into the analytics tool . We're now have tried two ideas . Both have failed. Neither is clearly going to work at all. We realize that we're screwing up. We decide to have Office Hours with YC again. We come back, walk around the little cul-de-sac by YC with PG. He sort of comes to a stop and says, "So just to be clear, you spent half a million dollars and you have nothing to show for it." Kind of a gulping moment. Yeah, I guess that's true. It was a good sort of come-to-Jesus moment though. Pause there, and then rewind all the way back to the first week of YC . It was in that first week that we had been like, well we should have analytics tools on our classroom lecture tool . So we Googled "analytics" and we found Google Analytics , Mixpanel and Kissmetrics and we're looking at them and we're like, we don't which one of these things we should use. They're kind of all similar but Google Analytics is a little more marketing-y, Kissmetrics is a little more revenue-y, Mixpanel is a little more product-y in terms of the sorts of insights they can give you. At the end of the day though, they all collect the same data. They all collect, basically, who is the customer and what are they doing? We decided to write this little tiny abstraction that could send data to all three. It was like 50 lines of code among the hundreds of thousands for this classroom lecture tool
Peter Reinhardt [09:22] - and then we decided, we'll just send it to all three, we'll look at all the tools and we'll just pick the one that we need. It'll give us a lot of optionality, basically, for free.
Craig Cannon [09:28] - Yup.
Peter Reinhardt [09:28] - And then we forgot about it for like four months. Four months later, we cleaned it up a little bit more. Four months later it cleans up a little bit more. As I mentioned, now we're struggling, at that point we're struggling with this question of, "Well, I already have Mixpanel installed, so I don't really want to use your analytics tool." My co-founder, Ilya, has this idea. He's like, "Remember that little library we wrote that sort of abstracted away the differences between these tools and let us send data to all three. What if we added ourselves as the fourth service that it could send data to and then every time someone has that objection, we hit them back with an open-source library that they can use to send to us and them." That seems like a clever growth hack. It gets us around this problem . So we did that, cleaned it up, open-sourced it. And people started replying like, "Oh, this library's great, we'd love to use it." Couple weeks later we'd follow up and be like, "Hey, we saw you're using the open-source library. But all you have to do is copy/paste our API key so that you can use our analytics tool . Could you just copy/paste the API key?" "No, like, eh." We started just feeling like there's some traction on this little routing library which was maybe up to like 25 stars on GitHub or something like that. Not much. But some people are issuing pull requests and it's the first time we've ever felt sort of like pull from the market, if you will. It was, we weren't just pushing a boulder uphill. It was a little different, but subtle.
Peter Reinhardt [10:42] - We had this conversation with Paul Graham the next day. We sit down. This is our second time with this. We're like, "Okay, we have a hundred K left in the bank. What's our final shot?" My co-founder Ian is like, "You know what? I think there's a big business behind analytics.js," which is this routing library. I was like, "That is, literally, the worst idea I've ever heard." I'm like, "It's 500 lines of code," I did groan a little bit, "It's 500 lines of code and it's already open-source. I have no idea how you build a business around that." We fought about it all day long. It was the four of us. I was the most skeptical. It was just literally like brutal fighting. I went home and I was trying to figure out how to kill the idea. Awake half the night. Finally figured it out. Came in the next day. I was like, "All right, guys, here's what we're going to do. We're going to build a beautiful landing page. Really going to pitch the value of this analytics.js open-source library. And it'll have a sign-up form at the bottom so that we can get people to sort of express interest. We'll put it up on Hacker News and we'll see what happens." I was thinking this'll just totally kill it.
Craig Cannon [11:38] - Right.
Peter Reinhardt [11:56] - Yup.
Peter Reinhardt [12:08] - That's right. That's an apt description.
Craig Cannon [12:08] - Had you experienced--
Peter Reinhardt [12:10] - That's an apt description. I had never experienced it before. It feels very much like losing control. Previously you're building a thing and you roll it out and you're building a thing and you're pushing it out and all of a sudden you put a thing out there and people start running away with it, and using it in ways that you didn't necessarily expect and you're sort of like, "Wait, wait, wait, wait, wait. It's just in beta, like, stop, stop. Like we need to fix these other thing 'cause otherwise..." It's like this feeling of losing control and almost like the market is dictating to you now what the rules of the road are and what needs to get built. It's a very different feeling.
Craig Cannon [12:41] - Would you differentiate that from overwhelming demand for one particular feature versus we're just going to take this and use it however we want but there's a ton of demand there? Would you separate those two things?
Peter Reinhardt [12:58] - Not really. There's people who always want more features. But the thing that flipped was, people would previously tell us they wanted a feature, but not use it. Whereas now, people were using it and they would want a second feature. It's a super important distinction. I think a lot of founders get caught in this sort of like all the death spiral of user feedback where they keep going and showing someone their product and asking them for feedback . They give them some feedback about how they could make it better, but they don't use it and then they bring it back with those fixes and they ask if this is better and it's just like the death spiral where it never gets anywhere. But once someone starts using it, they'll have more requests and that just means they're going to pay you more over time .
Peter Reinhardt [13:40] - Yeah, you really can't miss it.
Craig Cannon [13:44] - This was now six years ago? Five or six years ago?
Peter Reinhardt [13:48] - Almost six years ago, yeah.
Craig Cannon [13:49] - Right.
Peter Reinhardt [13:49] - 5 1/2, yeah.
Craig Cannon [13:50] - And you're still feeling the same way?
Craig Cannon [13:56] - Like what?
Peter Reinhardt [13:58] - About two years in, we discovered that all of our most valuable customers were sending their data to an s3 bucket, which is, basically, that they were keeping log files of the raw data, which was a little weird because, typically, what people were using the data for was to send it to an analytics tool , and an email marketing tool , and a CRM and a help desk, like places where a business person is driving value. Log files is a little different. It's a little weird. It's unclear what the use case is. We went on this sales trip to New York, myself and our first salesperson, Raff, and we met with five customers that were using this s3 bucket and we just asked them like, "Oh, what are you doing with the s3 bucket?" The first customer was like, "Well, we have a data engineering team that's taking the data out of the bucket and converting it into CSV files and then they're uploading it to our data warehouse which is a Redshift cluster." Basically they were using it as the initial input into a ETL pipeline. I'm like, "Oh, that's interesting, but, meh." Went to the next meeting. Second customer's like, "Well, we have a data engineering team who's taking the data out of s3, converting it into CSV," and we're like, "Are you kidding me? Like, okay, drop the eh," that's interesting. And then the third, fourth and fifth ones all said exactly the same thing and that was the point at which I started becoming a conspiracy theorist because it seemed like some pre-meeting had happened--
Craig Cannon [15:11] - That's funny.
Craig Cannon [15:20] - And that was a huge thing for...?
Peter Reinhardt [15:22] - It was huge, yeah.
Peter Reinhardt [15:25] - Yeah, it was very explosive. We'd grown revenue from zero to 2 1/2 million in the first year. Then we launched this Redshift connector and the next year we went from 2 1/2 to 10.
Craig Cannon [15:39] - But people weren't--
Peter Reinhardt [15:39] - So it was explosive there.
Craig Cannon [15:40] - asking you for that.
Peter Reinhardt [15:41] - That's right, it was one step too far for them to realize that we could do it easily. Their mentality, I think, was that, "Oh, Segment is a way that I integrate marketing tools and so a data warehouse isn't a marketing tool . It's a BI tool . Surely Segment can't integrate that." It just didn't click. So we had to go find by asking.
Peter Reinhardt [16:05] - Our go-to-market model is primarily through engineers . We talk a lot about sort of the way that we've built our infrastructure over time . We obviously process a lot of data, so there's a lot of interesting infrastructure problems . Now we're processing hundreds of thousands of user actions per second. There's a lot of data going through there. We were right about that. That's generally interesting to the Hacker News and engineering crowd. Typically an engineer is the one that brings us in. Sometimes a really technical product manager but it's someone who's like, "Yeah, this is going to solve this weird rat's nest data pipeline problem that I've got."
Craig Cannon [16:40] - Whoa, and how much of it is open-source still?
Peter Reinhardt [16:43] - A good portion of it is open-source but most of the value that we deliver is actually by running the hosted version.
Craig Cannon [17:17] - Right before you guys launched on HN, or this like small tiny micro launch or whatever it might be? Were there other avenues that you were considering pursuing? In that debate, in the day before, we're you thinking about other stuff?
Peter Reinhardt [17:36] - The debate was whether to build out the full product and then test for product-market fit by trying to sell it to people versus this super, super lightweight MVP landing page that we would put on Hacker News to see if there interest in the concept. What drove us towards the super, super lightweight test was actually the fact that there was a skeptical divide among the founders. Since the founders couldn't agree , the only way to answer the question was to go to customers ASAP and get an answer.
Craig Cannon [18:06] - Okay. It's tough. The launching early thing is always a challenge because there have been instances where people are like, "Eh, we'll just launch this early," but because they were like 10% off of what that product ought to be or they're not very good at communicating it, they never really get the feedback that they need. Like how do you kind of balance that out? Like, "Uh, this is kind of like fully-formed enough or we're communicating it clearly enough that we can launch it." How do you determine that?
Peter Reinhardt [18:34] - Usually that test is so cheap to run that it's worth running even if you decide that it was inconclusive and you should go a step deeper. But I also think that the way you framed it is actually an excuse that a lot of founders use for not doing the cheap early test when, in fact, they should.
Peter Reinhardt [19:04] - The really big product market moments for every company are pretty unmistakable. The drop by founders have called it stepping on a landmine. I really don't think you can mistake it. It really happens in a way that you lose control. It's very obvious. Every metric goes haywire. People are talking about it a lot. It's not mistakable for like, "Well, you know this person said that it looked valuable and was really exciting and blah, blah, blah." If they're not using it like, it's not there.
Craig Cannon [19:34] - Yeah, okay, so there's a question from Twitter . It's clear that a bunch of people have watched your lecture. This is first one, from Evan Farrell. He asks, "You mentioned in the startup school lecture that you had to pivot to analytics.js to find proper market fit. Is it possible to purely iterate on something like that to find product market fit or should it be clear from the outset if a new idea is something people want?"
Peter Reinhardt [20:00] - There's two versions of this. I would say the Airbnb version of product-market fit is much more iterative. They struggled for years and years and made slight iterations and iterations and iterations and finally it caught on and obviously they're a runaway success. My feeling is that that's extremely rare. Again, this is really dangerous place to be because you can stay in this iterative mode for years. It is unlikely that the iterations are going to get you to a good place. I remember very clearly early on being really inspired by the Airbnb story and it being a logical reason to why we should keep plugging away at a bad idea. And I think we abused the Airbnb story to just keep stringing ourselves along on a bad idea. I would be very, very careful of following the Airbnb example. I don't know many other companies that hit product-market fit that way.
Craig Cannon [20:52] - Right, how long do you give an idea at this point?
Peter Reinhardt [20:56] - I actually don't think that it's quite the right frame to think about it in terms of how long to give an idea. What you want is someone, either yourself or someone else on the founding team who's a skeptic. Someone who is going to have enough context with whatever the specific idea is and whatever the sort of regime or market you're in. Someone who's skeptical who will question and push for the fastest reasonable test . In other words, if you have an optimist and a skeptic and they both agree on what a valid test is, then I think you actually will end up with a good test . But if you have three optimists in a room who all agree on what a good test is, I don't believe that that's a good test . Ditto for skeptics, but yeah.
Craig Cannon [21:35] - Have you recruited skeptics or did you just kind of luck into that?
Peter Reinhardt [22:03] - That's right, yeah.
Craig Cannon [22:04] - How interesting.
Craig Cannon [22:14] - Right.
Peter Reinhardt [22:15] - That's what you want. You want someone who's going to be pushing it and you're like, well, how, hmm, yeah, how would we and then, who's willing to collaborate with you on how you should reasonably test whether those things are the case? The sorts of tests that we have run, for example, with this mindset recently and even in the past year, were, should we switch from a technical buyer to a marketing buyer? Unclear how to test that. Well, so, this early skeptic, who's amazing, her name is Diana, she was like, well, "I'm just going to go to a conference with marketers and I'm going to try pitching a bunch of marketers. I just flew to Florida and pitched a bunch of marketers." I came back, she's like, "Nope. Not a good idea." I ran my own set of tests. The hacky ways to test these things are very valuable and it comes from having skeptics and different perspectives of people willing to go tests those things.
Craig Cannon [23:02] - I imagine these tests from skeptics occur on a, maybe daily, but probably at least a monthly basis in terms of you guys working on your products.
Peter Reinhardt [23:13] - Yeah, I'd say it's maybe more on a per idea basis. If we're going to launch a new product then it's really helpful to have a skeptical perspective of here's why this might not actually be a good idea.
Craig Cannon [23:22] - Do you rely more heavily on data or actual customer interaction like in there?
Craig Cannon [23:30] - Okay, because this is the thing that bugs me more is when people are just putting up landing pages left and right and thinking that they can like kind of, I forget, what did Seibel call it, what do they do at Disney? Oh yeah, imagineer your way towards this winding down this path to finding it and it fact it's
Craig Cannon [23:52] - Yeah, they're just scared.
Peter Reinhardt [23:54] - Yeah and, when you actually go and talk to a customer, if you have that conversation in the right way, you'll learn a thousand times more from that conversation than you will from, than from putting up a landing page. Ultimately, we learned a lot more from talking to our customers after the Hacker News landing page than we did from the landing page itself.
Craig Cannon [24:12] - Yeah, totally. What are your tactics when you're talking to customers?
Peter Reinhardt [24:17] - I'd say the main thing is, most founders are not familiar with how a sales process is actually run and you basically want to run a sales process. The sort of typical founder motion with running a sales process is they come in and they say, "Okay, I'm going to give you a demo and I'm going to give you this really shiny polished pitch." Then the customer decides at the end of that pitch whether they're interested or not. That's not actually how good sales works at all. The way good sales works is, you do qualification up front. You have some method of understanding the customer's problem better than they understand it themselves and then you do the computation in your head as to whether your product is a fit for their problem . There's a lot of methodologies for this that the methodology that we use at Segment is one called MEDDIC, M-E-D-D-I-C. It's literally just a list of sales qualification criteria. This is what sales reps do. If a sales rep comes back and they're like, "We're going to close this deal," the sales manager says, "Okay, well let's go through M, metrics , what are the metrics by which this company is going to judge whether or not the product works for them? And if the sales rep can't answer metrics , economic buyer, decision maker, decision process, the identified pain and doesn't have a champion, if they can't have all six of those things, there's no deal. When you're searching for product-market fit , you can just go through all of those things by asking the customer a ton of questions. Then you can grade whether or not
Craig Cannon [25:37] - Well, it kind of ties into this skeptic versus the optimist idea, right?
Peter Reinhardt [25:42] - That's right.
Peter Reinhardt [25:56] - That's right and a sales qualification criteria is a way of almost putting the skeptic out as a structured process that enforces some level of skepticism.
Craig Cannon [26:04] - It's so dangerous when you're the opti... Because I fall into this camp for the most part. When you get good at sales, you can kind of sell many people on almost anything. But, if that product doesn't exist yet, it's very easy to just kind of mold it in the way when you're reading someone, you're like, "Oh, I can totally kind of want it to be like this so I'm going to kind of go down this path." But then when you actually show them the product and like you said, they won't even install it, then you see the world as it really is.
Peter Reinhardt [26:35] - Yup.
Craig Cannon [26:35] - And that's the thing. Are you still having these conversations with people personally?
Peter Reinhardt [26:42] - Sometimes, yeah, for sure.
Craig Cannon [26:44] - Yeah? All right. Because this is one of the things that people, I think, in large part, because they're influenced by your startup school talk, they have so many questions about it. Benjamin Liam asks, "How do you find that you have the right messaging around explaining your product?"
Peter Reinhardt [27:02] - Oh man, this is super hard. I'm not even the right person to ask about this. Actually, Diana, who I mentioned before, and our VP, Marketing , Holly, are the two people who have really refined our messaging over the years and we're always trying to refine it. I don't know how you know that you have the right messaging. You know that whatever messaging, You can sort of test whether alternate messaging is going to work and you can do that qualitatively in interviews with customers. You can try explaining it one way and see if their eyes light up. You can try explaining it in another way and just sort of see what resonates. A really talented early salesperson will also have this sort of pattern in their habit of how they pitch that they'll always be testing different ways of explaining the product . That was definitely true for the first salesperson that we hired. He was fabulous and just constantly experimenting with different ways of doing it. I don't know if there's, You never know if you have the best messaging but you are constantly searching and testing for different ways of explaining it.
Craig Cannon [28:19] - Then we should talk about idea generation because that seems more important than these minor deviations.
Peter Reinhardt [28:26] - Yup.
Craig Cannon [28:26] - Where do you begin?
Peter Reinhardt [28:42] - That's right.
Craig Cannon [28:43] - and then the individual features.
Peter Reinhardt [28:44] - That's right. And not just the individual features. You might have entirely new products
Craig Cannon [28:48] - That's true.
Peter Reinhardt [28:48] - that come along. But those are much easier. The problem with the first product and product market fit is that you can move the product and you can move the market because it's fit between these two things. What's unclear, and they move in some crazy multi-dimensional space , so the issue is to get them to both match up. You can always move either one and in different conversations. In one conversation you might shift the product and you might, in a different conversation, realize you need to shift the market. That's super tricky. I don't think there's a repeatable way to do that. You just have to go very, very deep into a particular market and understand the problems that people have in that market--
Craig Cannon [29:30] - So do you have--
Peter Reinhardt [29:30] - much better than anyone else.
Craig Cannon [29:31] - a particular process for idea generation? Or you just, you get into something and you're like, man, this is--
Peter Reinhardt [29:37] - You just go super deep. For that first idea, you just have to go super deep. You just have to understand the market and the ecosystem and the customers upside down and backwards, better than they do themselves.
Peter Reinhardt [29:53] - You'd have to start with something that was interesting to you, personally, and then you'd go dig in a deep direction. It becomes more repeatable when you are finding a second product . At that point, you've mostly locked the market side, right, because you already have a buyer, you already have a go-to-market motion. You already have an area of interest, which, for us, is these sort of data pipelines and data infrastructure, customer data infrastructure. Then it's much easier because you know exactly who you need to go to and you know, roughly, the type of questions that you need to ask. And then you can run a process which is a much deeper x-ray of the customer than you might be comfortable with. At least it was much deeper than I was comfortable with when we first got started. As an example, we recently were testing product-market fit for a product that we're going to announce at our user conference in September and its now in beta and it's doing really well. But, the initial way that we were sort of testing fit there, we would go in and say like, "Oh, hey , do you have a problem with data cleanliness?" And the person would be like, "Oh, yeah, yeah, totally. That's one of our big problems." We'd be like, "Okay, cool, cool," and like, check, data cleanliness. That is, and we might ask, like two or three or more questions but like that was sort of the depth. But the actual level of question needs to be like, okay well like, what do you mean data, like,
Peter Reinhardt [31:11] - "Do you currently invest in data cleanliness at all?" And they're like, "Oh, well, yeah. We actually, we have a team of like six people who do data QA all the time." And we're like, "Oh, well, those data QA people, where are they based?" "Oh, they're based in LA." "Oh, interesting, so they have real salaries and they're not just like overseas. They're like real like US salaries." "Like, yeah, yeah. Okay, so, what," like "80K a year, 100K a year?" They're like, "Yeah, that's about right." So like, "Okay, so you're spending like 750K a year for this data QA team and like, tell me more about that process. What are they QAing exactly?" "Oh, well, they're clicking this button in the app and," "Well, which button?" And then, are they, like "in the, What do they do when they find a bug?" And so we would ask, literally, 45 minutes of questions like this. Now we actually understand their problem and we understand what they're doing. We understand where their cost centers are. We understand how the team, And then we're like, "Oh, well, what if had a product that did X," which is exactly what they just explained to us for the previous 45 minutes. And they're like, "That would be amazing." So like, okay, now we, now this is, That was both sales qualification and discovery which is a standard sales process but now it's being used for product development.
Peter Reinhardt [32:18] - Totally.
Craig Cannon [32:19] - And the customers will tell you.
Peter Reinhardt [32:21] - Yeah, especially if you take the champion part of MEDDIC, the last one, the C, and you just start by asking them like, "What's your vision for X thing that you do?" Someone will tell you and you're like, "Oh, our mission is similar because that's why you got the meeting in the first place." That person is instantly aligned with you. They'll talk for 45 minutes about their problems before you have to tell them anything.
Craig Cannon [32:41] - Yeah, I think that's one of the things that most people don't realize that like, many of the best salespeople don't talk that much.
Craig Cannon [32:53] - Yeah, interesting. I wonder what the correlation is between sales and skepticism. It's probably pretty high. And people who are questioning things and like can see the angle.
Peter Reinhardt [33:03] - Yup.
Peter Reinhardt [33:11] - Thanks Danny.
Peter Reinhardt [33:22] - We have four values at Segment that we're quite dedicated to. The first is karma, which is we want to have a positive impact on the world. That manifests itself in a bunch of ways. One of those ways is we really care about the customer getting value out of our entire process. You'll notice that all of our marketing materials, for example, are often highly educational. We have a really high bar for what a piece of educational material looks like. Even in the sales process, we want to be helpful. If we're not the right fit, we'll tell you and sort of like refer you to the right places. Separately, we really care about doing the right thing by the end user. This is still within karma. And that's from like a data privacy perspective. We're very interested in helping companies understand all of their own first party data, so all their interactions with their own customers within their four walls. We're super uninterested in helping companies data broker data between different companies sketchily. We call it data gossip. It's gross. We don't want anything to do with it. There are plenty of other companies out there that do stuff like that. It's going to go away and die eventually. That's karma. We care a lot about that. The second one is tribe, which is that's when we're all there to support each other, we'll all there to accomplish the same thing. And so, what we expect is that, and what we value is, that people really support each other, both when they may be struggling with something but also giving them credit.
Peter Reinhardt [34:48] - Really try to reward folks who are willing to go the extra mile to give credit when it may be hard. It could be giving credit across teams or up several levels or whatever. That's really something that we value. The third is drive. Much more self-apparent. We like to get shit done. We value people who are getting shit done. And the fourth is focus, which is not just sort of the ability to sit down and get stuff done but, or power through something, but actually thinking carefully about prioritization. We've done a lot of research around how to make the office an environment where you can actually focus. You can check out our blog. We've written about, sort of, sound decibel levels that we've measured around the office and how we've mitigated that.
Craig Cannon [35:27] - That did pretty well. That piece did pretty well, right?
Peter Reinhardt [35:29] - Yeah, and it was a surprising result for us to discover the different parts of that office had very different sound levels that were not correlated with people talking but were just correlated with the sort of acoustic shape of the office and just moving people around into different places helped a lot, depending on how much noise they were willing to tolerate and sort of needed in their role. Anyway, those are those four values . They are literally the things that we value and so we push them into all of the places where you would expect what you value to have an impact. It's who gets highlighted at all-hands. We have a citrus prize, which is someone who's living all the values . Promotions, hiring . We have a strict interview in the hiring process. We have performance reviews.
Craig Cannon [36:08] - What do you mean a strict interview?
Peter Reinhardt [36:11] - Sorry, we have a culture interview where we literally have these four values and specific ways that we're going to test for them. When we run performance reviews, the performance review is literally the four values . Are you strong? This is what we value and therefore it's what we test and measure by and ultimately it's that cycle of giving feedback and measuring by it that is what drives culture to stick.
Peter Reinhardt [36:38] - I don't think so. We learned it. We got to about 25 people before we realized that it was something we should write down. We went off site, the four founders went off site and we tried to synthesize the values out of what it was that we really liked that was already happening, and what it was that we didn't like that we had seen already happening. And not just among the team, but amongst ourselves too. What were we not proud of that we had done and what were we proud of that we had. And that ultimately was what got synthesized into those four values .
Peter Reinhardt [37:19] - Interactions with other people and interactions with partners and customers and things that we were proud of that we wanted to see more of.
Peter Reinhardt [37:35] - GDPR , for those of you who don't know, is a new EU regulation which, basically, gives end users a lot of rights about the data that's collected about them. First off, I think it's an awesome regulation, both as a consumer, but also wearing my Segment hat. It's interesting in that it impacts the entire globe because, if you are storing data about an EU citizen, it doesn't matter what jurisdiction you run your company in, you are still responsible to do that for an EU citizen. The biggest impact broadly on the sort of overall ecosystem is it really negatively impacts third-party data and third-party data brokers because they have no real consent path to the user for sharing and buying and selling their data. Because we help companies purely with their first-party data it's not a existential threat to us in any way and in fact it's something that we're really sort of aligned with for another reason as well, which is, because we're routing the data out to all the different places where people are using it, so we're routing it out to an analytics tool , to an email marketing tool , to a data warehouse, to a CRM, to a help desk, to ad conversion pixels. If that user shows up and says, "Hey, I want you to delete me from your system." Well it's actually like 20 systems for most companies and we're already plugged into those 20 systems. It's actually now a feature of Segment that we can go to those 20 systems and delete whatever user is requesting it
Peter Reinhardt [38:58] - and clean up that record across all those different systems. For us, GDPR one, is like aligned with our values philosophically, two, is actually an exciting new feature and a sort of requirement that we can support and a sort of a value that we can provide to our customers. We're huge fans.
Craig Cannon [39:15] - Nice, that was a was a perfect answer. But, people have been stressed out about it. My friend makes Instapaper and they had a big issue with it.
Craig Cannon [39:29] - Yeah, especially these like little tiny products that are parts of really big companies and they didn't necessarily know and now they're everywhere. Totally. Cool. All right. So, next question . Andrew Pikul asks, "Any advice that you have on asking for more money than you're comfortable asking for?" This was part of your start-up school lecture where, I guess , one of your sales reps was forcing you to ask for more, a lot more.
Peter Reinhardt [39:54] - Yeah, we had a sales advisor who was, well, I got to back up a little bit. We were initially selling our product for $10 a month and $120 a year and we brought on the sales advisor and his first advice was, "Well, you have to ask for $120,000 a year." I was like, "That's a thousand X, that's crazy." We were going to the first sales meeting, me and him, and this was with a company called Xamarin and I've since told not this story, which you found amusing, but now is the CEO of Xamarin and, as we're walking up, our sales advisor says, "Okay, you have to ask for 120K in this meeting." And I was like, that's the most ridiculous thing I've ever heard. I'm not doing it. And he's like, "Well, if you don't do it, then I quit as your sales advisor." I was like, "All right, I guess I'm asking for 120K." So, we go in and we have this demo and everything and he says, "Okay, well what's the price?" And I say, "120K and I turn beet red." And he says, "Well how about 12K a year?" And I said, "Well how about 18?" And he's like, "Okay, fine." From his perspective, he got 85% off. From my perspective, I got 150X and it was a successful negotiation . It's really hard to offend people with price, at least if you're sitting in the same room or on the phone. It's probably not a good idea to share pricing information via email . If you do that, then its really easy for them to hang up. But if you're on a phone call or in person, there's a bit of a social contract to sort of continue engaging. Particularly in person you can recover.
Peter Reinhardt [41:19] - I would encourage you to not be scared of offending someone with a high price. But maybe just start it in person, which is probably the most uncomfortable place to do it, but gives you the most opportunity to recover.
Craig Cannon [41:30] - And the thing is, if it actually matters for your business then that's just what it costs.
Peter Reinhardt [41:33] - You have no other way of assessing the value. And in fact, what will happen is, when they say, "That's crazy," then you say, "Why?" And then they'll explain to you how they actually value the product and then you say, "Okay," and then you value it according to their logic and you ask for that price.
Craig Cannon [41:50] - And how long did it take you? Well, are you charging them 120 now?
Peter Reinhardt [41:56] - For sure, yeah, we have customers that get way more value than that out of it now.
Craig Cannon [41:58] - Yeah, exactly. How many customers did it take you to reach that six figure amount?
Peter Reinhardt [42:05] - A dozen at most.
Peter Reinhardt [42:18] - YC was super helpful. The most impactful thing early on is just Demo Day. You're not going to find a bigger concentration of investors who are excited about investing in startups , creates a compelling event. Structure of the timeline is incredibly helpful for our first round of financing that can easily get strung out and waste a lot of your time . That's the first thing. The second thing, really, is the founder network. There's not only a lot of reasonably high-profile companies now that you can learn from or companies that are sort of farther ahead that you can learn from now, but there's companies at all stages. There's almost always a group of people in your area or in your market that you can learn from and share from. There are tons of little groups that spring up. A group of enterprise founders that are all between like 70 and a hundred people in San Francisco and you can have dinner once every two months and that becomes an incredible support group and sort of way of learning about what's going on.
Craig Cannon [43:19] - Have you stayed in touch with people from your batch?
Peter Reinhardt [43:22] - A few, yeah. Like Zack Sims from Codecademy.
Craig Cannon [43:24] - Oh, right on. Yeah, I've heard of like these informal founder meetups happening quite often.
Peter Reinhardt [43:29] - There's a lot, yeah.
Craig Cannon [43:30] - It seems to be great.
Peter Reinhardt [43:31] - It's a trusted network. There's no replacement for that.
Craig Cannon [43:35] - Yeah, totally. I definitely didn't get that from college. All right, Juan has another question . "In the early stage , what's the thin line between ignoring a customer's suggested feature or moving a customer's requested feature to the core of your application or product?" I think what he's trying to ask is, basically like, at what point do you say like, "Hey, this customer is requiring or asking for this feature and we have to kind of hold the line because we don't want to become a custom dev shop?"
Peter Reinhardt [44:06] - I see.
Craig Cannon [44:06] - So, should we integrate this or tell them to find someone else?
Peter Reinhardt [44:11] - The best defense against that is, having a clear product vision for where your product is going to go long-term . And if you have a clear product vision for where it's going long-term , it's a very simple question of, is this thing in that picture long-term or not? And if it is in that picture long-term , then you can prioritize it to be sooner or later depending on whether a customer is going to pay you for it or not and, if it's not, then it's not and you probably shouldn't build it.
Craig Cannon [44:38] - Right, that's like the infamous customers you don't want scenario, where you just have to let them go.
Peter Reinhardt [44:44] - Yes, so I guess the important thing is, imagine the entire timeline of everything you're ever going to build. Feel free to move things around. We do this all the time . We move things around based on what customers actually want because it's a reasonable signal of what's actually more important. But I wouldn't add major things or move major things from it just based on one customer.
Peter Reinhardt [45:08] - Oh my gosh, so many. A huge bucket or a huge area of learning for me, as weird as it is, is finance . I came from an aerospace engineering background. We were doing software engineering for the first couple years and, so you just are completely unprepared for the like business side of things. I've learned a tremendous amount about finance as we've raised money and learned to manage a business with a P&L and all those things. Not that you should rush into it, but it's a huge area that can be leveraged.
Peter Reinhardt [45:48] - I actually think we did a reasonably good job of that. We hired a part-time CFO around the time that we raised our series A, so we were at about a million in revenue and we raised a $15 million series A. We had had a bookkeeper up until that point but, we were like, "Uh, ee, ee, I feel like we should have someone like, point us as to what we should be doing with the money." Maybe have ke a plan or a model or something?" That was definitely the right time to hire a part-time CFO and Jeff Burkland was super impactful over the years.
Craig Cannon [46:16] - What have been the other big, important hires for you that have made like a huge difference?
Peter Reinhardt [46:22] - Well, every member of the exec team--
Craig Cannon [46:23] - Everyone was huge.
Peter Reinhardt [46:23] - A whole bunch of people but, advisors, maybe, as an interesting category?
Craig Cannon [46:26] - Sure.
Peter Reinhardt [46:26] - Part-time CFO , I think, is in that bucket. We had an HR advisor who was really impactful. We have invested more in HR than most startups of our size and I think that was the right thing. A lot of startups , like Uber , for example, do not and end up with paying really high prices for this later on.
Craig Cannon [46:43] - It's a challenge, right, because if you go around and start Googling like, should I hire a CMO, should I hire a CFO , should I hire X Y and Z? You can always find someone strongly advocating for any particular role. But the challenge is you only have so much money and so much time and you can only find so many great people. Where and when do you decide to hire those like optimal people for this stage in your company, right? Just kind of curious if there were any big turning point moments for you.
Peter Reinhardt [47:15] - It was a huge turning point around 10 million in revenue when we hired the first two sort of execs to the team. One was our VP of engineering and the other was our VP of people. They were the first people who had previously been managers. Our VP of engineering had managed a team of 150 at Dropbox . We went from literally zero management experience aside from what had been picked up along the way growing from zero to 50 people, to having someone who really knew it or two people who really knew what they were doing. That was hugely impactful and we should have figured out a way to do that earlier. 50 people and 10 million in revenue or whatever it was was way too late for that.
Craig Cannon [48:01] - Nice.
Peter Reinhardt [48:01] - There's a lot of cool things happening there. I'm always blown away by this like breadth of things that are happening in the batch. This batch there's a really exciting company building an in-space rocket engine and another that was doing industrial inspections by drone. I just can't imagine a world where we continue to have people in harnesses hanging off of wind turbines. I can't imagine that that continues for a long time so, that seems like an obvious market opportunity.
Craig Cannon [48:29] - So flying things.
Peter Reinhardt [48:31] - Well, I have a background in aerospace engineering so it shouldn't be too surprising.
Peter Reinhardt [48:53] - Cool, thank you.