Episodes

The Strategy of Kyle Wild on How to Build Analytics into Business

Kyle Wild

Kyle is the co-founder and CEO of Keen IO, which helps app developers build custom analytics & data science features directly into their mobile apps and web dashboards. They help you build analytics into your business.

TOPIC KYLE WILD COVERS

  • Find out what is Keen IO does
  • His thoughts on Keen IO
  • How to solve the data talent crisis
  • How to use APIs
  • His strategy on how to build analytics into business
  • His favorite quotes
  • How does the Keynotes differentiate itself from KISSmetrics or Mixpanel
  • His best advice for somebody trying to grow a startup
  • His focus is more on data, strategy, and education
  • And a whole lot more

LINKS & RESOURCES

WATCH THE INTERVIEW

READ THE TRANSCRIPTION

Bronson: Welcome to another episode of Growth Hacker TV. I’m Bronson Taylor and today I have Kyle Wild with us. Kyle, thanks for coming on the program.

Kyle: Thanks for having me.

Bronson: And I have to say, your last name might be the best last name ever.

Kyle: Right.

Bronson: But Kyle Wild. So you are the CEO and founder of Keen IO. Tell us about that. What is keen? IO.

Kyle: So keen IO is a developer platform that lets software engineers do really complicated data science work really easily.

Bronson: Perfect. Now, I saw online that you said that you’re trying to do for analytics what Amazon Web Services did for the data center. What do you mean by that?

Kyle: Well, the field of it’s I mean by by analogy. So the field of analytics and data science has this really complicated skill set that’s really difficult. And right now, it’s actually the hardest thing to hire in technology is the data infrastructure engineer and a data scientist. So we thought back to the days before Amazon Web Services, like when I when I got my first kind of web business, when I was a kid in like 1998, 99, I had this.

Bronson: The dark ages.

Kyle: Yeah. Pages where I had to think about servers, everything from servers and how big is my Internet connection and what’s the co-location center like all the way to like my code and what’s the web website design like? What’s the content like and everything in between. And, you know, there’s the the data center engineer and the system administrator, that kind of whole skill set used to be part of the cost of doing business when it came to building an online company. Mm hmm. Mm hmm. Now, a lot of that’s gone away, and at least the upfront costs and the talent cost has gone away. And Amazon Web Services puts that behind an API so that you can just any any developer can spin up, you know, 500 servers if they want, if they’re willing to pay as they go. Mm hmm. Mm hmm. And, you know, we looked at it like that. That would be the right solution for how to solve the data talent crisis. So data talent crisis is really, really big. It’s really real. I mean, you should look at any of the people on my teams. LinkedIn inbox is just full of recruiters trying to get data people. Yeah, I think there’s a better solution than throwing more people at the problem. And the solution is giving developers tools to build whatever they want in the data. Data like in data space, I guess.

Bronson: Why? Why is it a better solution? Talk us through the pros of that of that, you know, API approach instead of just throwing more data scientists at it.

Kyle: I mean, one of them’s really cause, you know, even if there were enough data scientists on the talent market, it’s it’s not cheap to have these people on staff all the time. And another and a piece of that cost is that it’s front loaded cost. So, you know, it’s front loaded value, but it’s ongoing overhead. So most of the work that you can most of the value you get out of the data science team happens. You know, in the first few months that they’re working with your company to build up the infrastructure, the practices, and get the stakeholders all the data stuff they need. And then it’s kind of on autopilot. We think that, you know, the solution of throwing a bunch of data scientists at the problem is way too expensive. And that’s if the people around on the talent market. But really they’re just not. Yeah I mean they can helpers have most of the tools they need to be data people. I mean software is just this thing where you write code and move data around. That’s what some software already is. Data oriented. Mm hmm. And software engineers have they’re all they’re usually pretty quantitative. They know how to use APIs. They already have the skill sets to be data people. And the tools are just difficult. And it takes a long time to figure out how to use some of the tools available. And we’re here we’re here to try to simplify that whole world down to the things that we think everyone can understand.

Bronson: Yeah. So just so we have a even a clearer view of what it is you’re kind of replacing what a typical data scientist do in a startup. If you were to hire one or a team of them, what kind of numbers are there crunching? What kinds of reports are they giving, what kinds of things they do day to day?

Kyle: I mean, some of the big things are really not that complicated. Things like AB tests, you know, test where you say, I’ve got content that looks like this and content looks like that. I’m going to put them against each other and see what performs better. So like my landing page might have, you know, ten versions that have different copy and I want to see which one converts better. That kind of stuff is pretty simple. It’s like a two step conversion analysis and then you’ve got more complicated stuff, like somebody, you know, came from a certain ad that we posted on Google, and then they landed on the site and they made an account, and then two months later, did they pay or did they not? That kind of analysis like a long funnel analysis. Those are kind of two of the most common cases and really it’s kind of a matter of what they do all day. So they answer the first part of your question. What they do is they like learn how to manage like a ten cluster, a ten server Hadoop cluster, and how to build players on top of Hadoop and write distributed functional code and like really like pretty esoteric stuff. Yeah. Just just to do like a funnel for election analysis stuff that’s not that hard to understand conceptually. It’s very hard to, to actually implement. So we’re here to fix.

Bronson: So you kind of provide the tools that, you know, give developers the ability to become data scientist because they already have so many of the tools they need and so many of the characteristics and personality traits, you just give them the last little bit. They need to really go deep into data. Would that be accurate?

Kyle: Yep. All right. Yeah. Like we want to make a funnel, like we have an API for funnels. Like we have a we have a JavaScript library for people to make funnel. So like instead of all this, all of those things I just describe that people have to know as data gets big, they just don’t care. The data is big, they just pay more and it’s pretty much.

Bronson: Yeah, makes sense. So what was the initial reason for building this type of platform? Was it just a personal need? You were a developer that needed an API to do this. Was it just you saw a pain in the marketplace and you took advantage of it? What was kind of the impetus for beginning? You know.

Kyle: So one of my favorite quotes is in the Stanford commencement speech when Steve Jobs says like, you can’t connect the dots where you’re from, where you’re standing, but if you look back when you can connect the dots, right? Yeah. It’ll it’ll make. It seemed like this was on purpose when I connect the dots for you. That’s right.

Bronson: That’s life right there.

Kyle: It makes it seem it’s like, oh, wow. Actually, if I’d planned this, it would have made a lot of sense. Uh huh. And but so, I mean, my first job out of college was I was a developer support engineering support for Google Analytics. Mm hmm. And what I did was I help developers who wanted to get more power at Google Analytics, which wasn’t really a developer tool. It’s really more of a tool for people buying ads. Mm hmm. But people who, like I talk to those people all day long and help them with their with their saw and the use cases that people have for analytics, stuff are just widely divergent and they’re all over the place where people value. Then I went on and I did a few startups and at every startup I was building kind of internal. I mean, I’m a data geek, so I was always expected to be a software engineer. I was always trying to figure out what’s going on, what are people doing, what’s working, what’s not like segment, the user base, all that kind of stuff. Yeah, I did that across a few startups and by the third time I was like, okay, you know, I’m pretty decent at this now. Mm hmm. Why do I have to keep building this stuff from scratch? Like, what tool would have worked out of the box? Yeah. And, you know, my co-founders are both from the Foursquare platform team, and this is the Salesforce developer ecosystem. And between the two of them, they spent ten or 11 years there. Mm hmm. So I talked to them about and I was like, I think the key problem is that everybody out there that builds analytic stuff is trying to build for a use case. They’re trying to build a specific product for a specific use case. Mm hmm. Which is fine. But then there are inevitably, there are all these users who want to use it for other things and want to have their own use cases. So we’re like, what if we just took down the sort of level of abstraction just a little bit and said, Build your own? Like, what if you could just build your own Google Analytics in, you know, ten or 15 API calls? Gotcha. So it takes a little bit more work to get going. But now if you want an additional kind of feature, you can build it because you own it.

Bronson: Yeah. Well, that kind of leads into my next question is how does Keynote differentiate itself from like KISSmetrics or Mixpanel? And I think you just answered it, but I want to hear you kind of do it in the context of those companies, you know?

Kyle: Yeah. I mean, the key distinction is that they’re built for a use case. So, you know.

Bronson: So what does that use? Case? Tell us what they’re built for. And so they show us the robustness of scenario.

Kyle: So the use case for your panel is like you go you you send them data. Mm hmm. You write and you write code in your app, you send the event data. So events that are happening inside your app, and then you go type in Mixpanel dot com and you log in and you click around and build charts and stuff and that’s it. And on top of that, they have a really great use case that’s for marketing automation. So now you can like set up triggers that will automatically, you know, send alerts to users or to your sales team about different things that are happening with users. And it’s a very good product. It’s a product. Mm hmm. What we do is, like, if you log into our site, you don’t see any of those fancy tools. You see, like developer credentials, API documentation. But the difference is that with, you know, with some effort, you can build tools like Mixpanel or KISSmetrics on our platform internally owned. So you know, basically if they, if those tools are really good for one or two use cases, we’re kind of this tool that’s built for. All the other ones or even just more tailored versions of those use cases. So a good, for instance, is like one of the, one of the key when we so I keep saying like, oh, it’s all these many, many uses, but we do segment the world a little bit and try to think about how do these how are these things falling into buckets? Mm hmm. One of the things we talk about is like, okay, let’s say that you just wanted to have the exact same functionality, but you want it to be on your own website. Mm hmm. So you’re assessing your SAS business, you know, and your, uh, your customers log into your website and their businesses themselves, and they want to go to their admin panel and see analytics. You can’t exactly just say, Oh, here’s the account credentials for our master account for KISSmetrics. You can’t do that. So a lot of our customers are kind of this we call it like the two level analytics problem. They want to do data science and analytics across all their customers, but they also want to provide a subset of that reporting to each customer. Yeah. And that’s that’s one of the things that we’re that we’re tailor built for. You know, we have security features to allow people to do that in a way that segments one customer’s data from another, but also they can zoom out and look at the whole thing. That’s one of those things where we didn’t really anticipate that use case, but it started to emerge and now we’ve built features to support it.

Bronson: Yeah. Well, I mean, I don’t know where else I would go to actually do that, you know what I mean? I can go and get data for my personal web app, but I can’t then give that to my clients in a in a third, fourth, later down. So I don’t know what the other solution would be except for what you did, which is doing it in-house over and over, and then realizing there’s got to be a better way. So it makes sense why people are using it that way.

Kyle: It’s in-house, it’s over and over. You know, like look at things like Flurry. Flurry has like eight or nine really nice charts for mobile, both for people, mobile app makers to study user behavior. Mm hmm. What if you want the temperature, you know, you want you want an 11th chart or you want each of those charts to have a dropdown where you can search by gender. You know, it doesn’t have that. So you’re just stuck. It’s like, Welcome to Better Tool doesn’t do it. So we can’t actually do any kind of ask any data, data questions about like gender or age or whatever, whatever the random thing is that puts people over the top and then they come to us.

Bronson: Yeah, absolutely. I was actually studying Flurry this week to try to get my head around oh. And seeing what they do and actually help me understand what you guys do because I saw the differences. You know, I could see what it means. It almost reminds me of like the difference in Twitter and like app dot net where it’s like you have one that’s the product and you have the other. That’s the real developer tools to rebuild a Twitter if you want or rebuild anything else if you want. Is that a fair comparison, you think?

Kyle: Yeah, I think so. I mean, I think that’s a very solid comparison. Another one that that we’ve kicked around is like, you’ve got Tumblr. Mm hmm. Which is good to go make a blog, and then you’ve got WordPress where you can make all kinds of stuff. You can add whatever you want to it. It’s a little bit harder to get started. It’s built for developers. Yeah. You can do a whole lot more with WordPress than you can with Tumblr. And one of the things you could do with WordPress is build Tumblr if you wanted to.

Bronson: Yeah, that’s great. So, yeah, I can’t leave the my next question, which is what kinds of companies should really consider using Qiniso because like you said, it’s a little bit harder to get started. So this is going to scare away some kinds of companies, but it’s going to make other companies get very excited. Who is your customer here? Ideally, what kinds of verticals or what kinds of DNA do they have?

Kyle: So so the the verticals are really all over the map, but the we segment in other ways. So we’re focused on companies that have between, call it, three and 20 developers, you know, not just one or two people. You know, they probably have a little bit of funding, although we we’re happy to support people that are in accelerators and hackathons. We have a free a free tier for that. But our sweet spot, really, they have a handful of people up to maybe 20 developers, probably more like 3 to 10, to be honest. Mm hmm. And they, you know, the company size can skew all over the place, like we’re working with a large online publisher who has six developers, but they’re a very large company because they’re not a software company. Yeah. And then we’ve got, like, software companies, when they have six developers, you know, they’re, they’re obviously much earlier stage and in their DNA, they’ve got you know, they want to get more out of data. You know, they’re they’re kind of chronically unhappy. Like when we see people complain or hear people complaining about the analytics tools they’ve used before, it’s easy to throw stones at them, but those tools are great at what they do. The real the reality is that this customers are just more demanding. Yeah. So we skew toward definitely more demanding. And people like people that are kind of unhappy or wish there were more they can do with whatever analytics tool they might be using all the way up to, you know, people who have already realized this is a problem, that that building in-house is the only alternative, and they’ve started trying to build in-house. Sometimes they’ve they’ve spun up a database like MongoDB or Postgres to try to build an analytics and try to try to build the kind of stuff we built. Mm hmm. And they they run into problems. Either that developer left and wasn’t there just isn’t there to support it anymore, or it ran into scale issues, which is pretty common. This stuff’s not easy to build at scale. Mm hmm. Or, you know, they just realized at the very beginning that there must be some way to do this and get value, like next week, instead of having to wait for a six month build out of of a big data backend. And those companies, I mean, so yeah, we skew toward I mean, we focus very much on the kind of technical. A component of a company. A lot of times it’s developer teams. Sometimes they’re people who have data engineers, and then engineers are just they just don’t want to build the event data thing to handle all these all this event data and historical stuff. Yeah, they’d rather do data science and ask good questions that can drive the business now. And they want to get they want to get more leverage the same way. Like I used to build servers as a kid, but I, I sure don’t do that now. Yeah, I, I, A.W. as a new software, I use Rackspace and I don’t want to manually put hard drives into computers, even though I know some of them actually know very well what we’re doing. And they ask us very few questions about like our actual data stack and they ask us about how we’re queuing and like what’s our scalable database back kind of they ask us to kind of questions. I’m like, Oh, you could do this. You just don’t want to. And that’s kind of.

Bronson: Yeah. You allow people to focus on the higher order things, the more important things for their particular business, because they’re not an analytics business, they just need the analytics to drive their business. So so it makes sense that you provide a forum and I was looking at your website, it seems like that you guys are more than willing to actually provide the service of actually setting up key now for some of your customers and actually building out these dashboards and these panels. Tell us about that a little bit. What does that process typically look like when they bring you in to actually implement beyond just providing the API for their developers?

Kyle: So I mean, we do a few things in the in the kind of like professional services side. We I mean, I’ll, I’ll talk a little bit about why and then I’ll dig in. So the reason here is that, you know, you look at a product like Google Analytics or Flurry, one of the things that should stand out as a is if you’re evaluating those products is that they’re free. And that tells you something. Either it’s either it’s an unsustainable business or they’re making their money in some other way, you know. So Google Analytics is really generally a component that helps AdWords sell better. And, you know, tourism is an intelligence company, so they want you to instrument your app so that they can study the data themselves and build good research and do ad optimization stuff for their customers. So, you know, the incentives aren’t aligned unless you’re paying. Mm hmm. Now, at the same time, we’re in a position where our incentives are very much aligned because they’re like, we get more, we get we basically as our customers grow, they pay a small business of actually making our customers better at what they do. Yeah. To the extent that, you know, we can. So we’ve got, we’ll give away like a day at a certain tier. We give away one day of data science consulting or we’ll give away a half day at a, at a smaller tier. We’re we’re really doing that because we want them to have a better integration and get more value out of their data. And there are some customers who, you know, who think this is going to take longer. And sometimes it does. And a lot of times we have to focus more on data, strategy and education. So educating by giving them the vocabulary as a team to talk about about this stuff and helping them with strategy. Like we’re really digging into their business. And so it’s it’s really kind of some of the most fun work we do is like sitting out people and talking. We’re like talking about where, where’s the data coming from now? What do we have and what can we do? You know, where’s the 8020 kind of trade off on on what kind of integration we can do to start getting business value immediately? What should we do? So should we be focusing on like marketing data or on usage data or what, you know, that kind of thing on the strategy level. So there’s data, strategy and education. And then what you what you asked directly about was kind of implementation consultant. So we’re very, very cautious about this. Like we don’t want to become a dev shop. We want to partner with those kinds of companies, but we want to focus on our platform. That being said, like some of the best things that some of the best learnings we get come from when we’re using our tools are using our product ourselves. So what we’ll do in the integration site is we’ll pair program with people, which means we will deploy somebody to sit next to their, their, their developer and like work through the problems together. That way we know it’s still partially about education and coaching them alone. And we’re we’re getting they’re getting value because they feel comfortable that they know the best practices are definitely being followed. The person in the room. Yeah. And on our end, we know that we’re not an outsourced developer software that they’re that they’re using to just do that. We have partnerships with development firms and consulting companies that do that kind of work. We focus really on teaching people to fish. We don’t want to get in the business of fishing.

Bronson: Yeah, absolutely. It makes a lot of sense. You talked about how you’re you’re cautious because you don’t want to become a dev shop. And so the reality is that most people in the world will never be able to sit down with your team. So here’s what I want to do for a minute. I want to do like a little more consulting. Okay. Now, obviously, I know that every situation is 100% different, that, you know, the things that are different are more severe than the things that are the same. But putting that aside for a second, I want to ask some questions as if I’m coming in to your data scientist and you’re trying to help me figure out my business. All right.

Kyle: So what? I have to give a little bit of a disclaimer. I’m not on the team at this point. I’ll give it a shot. No, no, it’s fine.

Bronson: It’s fine. My questions are high level because I can’t get into the actual specifics because they’re so different. So let’s do a little more consulting. What? Data. Should a startup generally be recording what kinds of data happens on a site or happens in their ecosystem? That really needs to be captured.

Kyle: And it’s a very general question. Yeah, I’m a data not so we’ll give.

Bronson: Us general answers to some of the possibilities.

Kyle: But but like a website, you know, the very minimum is you need to capture every page view. Mm hmm. Mm hmm. And to the extent that you can, each page view should be captured alongside all of the user information that you have. So if somebody lands on my site from a Google ad, let’s say something like on my site from a link, you know, from from goes hacker to write up that information about where they came from needs to follow them forever. Okay. So that’s that’s kind of what we call, you know, capturing the state, capturing everything about the the user, even if they don’t have an account yet. Mm hmm. We do know something about it. We know what’s what page they landed on. Might have been some blog post that we wrote about, you know, iPhone data collection or something. Mm hmm. We know where they came from. It might’ve been somebody that cross-posted that blog post. Mm hmm. We need to know that stuff forever so we can attribute their future behavior. You know, let’s say you’ve got pageviews, but you’ve also got sign ups. Mm hmm. You’ve got purchases. Uh, I need to be able to attribute those purchases all the way back to where they came from and what landing page they hit. So we can point to, like, wow, this blog post doesn’t get as much traffic as the others, but 4% of people who went through this blog post end up paying within a month. Like, that’s pretty awesome. Let’s like drive traffic to that. And that’s just with three kinds of events, you know, like landing on a page, looking at any pages and paying. Mm hmm. You can do a lot with just those. Mm hmm. But the reality is, like, the more general answer is any, like what qualifies as an event? It’s kind of a big question. Mm hmm. For us. Well, what we like to tell people is anything that your team on the business side and the developer side agree makes sense. Like anything that you if you both speak the same language on something like everybody understands page view, that’s probably an event that you should be reporting a click. The business people don’t know how they’re going to analyze. Click. Mm hmm. Even though it’s easy for click to be, it’s very easy to to, you know, record that as a developer. It’s probably not the right, you know, level of granularity to be recording. Yeah. So, see, I mean, it’s case by case. A lot of times we have it deployed and we just have to think on the fly and we have kind of heuristics for how to think about this. But there’s no one size fits all answer, which is part of why we exist.

Bronson: No, I like what you just said, because you actually give a great insight, which is it’s where the suits and the devs agree. Right. When they agree on what an event is, you probably have hit upon something because the developers know, you know logically why that matters in their system and the business people know how they’re going to use it to inform their future decisions. So in that Venn diagram, when those two things overlap, you have something very valuable that’s happening on the website and that could be different for every startup. But now at least there’s a way to kind of get to that. So I like what you said there. Are there any kind of things that they should be doing with that data that every startup should be doing? You know, you hear a lot about, you know, cohort analysis and things like that. Now, should all startups be running those kinds of analysis on the data, or is there any other things you think they should be doing, this data that maybe doesn’t get as much, you know, buzz or hype? I mean, what do you think about that?

Kyle: I have a lot of thoughts on that. I try not to make any decrees like all startups should be doing the following. Like there’s a lot of stuff that almost all startups should be doing that we’re not even doing it. We’re a data science company in our own site. We’re not doing everything that we want to be doing. And some of it’s because, you know, sample size, basically, you know, there are a few kinds of analysis that I think are definitely under appreciated. And, you know, number one, easily number one is funnels. Conversion funnels are awesome. And like we have we have cases of people using conversion funnels for crazy divergent things. You know, conversion funnel just means somebody did did step one. How many of those people did? What percentage of people did step two and what percentage of those people did step three? Mm hmm. So, for instance, we have a partnership with SendGrid, which is an email API. Yeah. And, you know, SendGrid generates all these events that you can dump into our system with like a little one click integration. And those events are like delivered an email, email got opened, email got clicked, you know. People should be making funnels that start there and and continue on to landed on the website made an account of that or logged in to the website all the way through to paying so you can actually attribute an email campaign to, you know, what happened in your business. Are they not doing that? The reason for that is that it was actually impossible before. I mean, you could either build in a house, but there isn’t a tool that can do that besides ours because we have a very generic kind of concept of a funnel and people use our use that stuff for things like a server boots up and then a server hits a memory limit within it within 3 hours. Like that’s totally crazy. It’s not necessarily the people oriented. I think that kind of analysis is really powerful and I think being able to filter on the cohort is really, really crucial. So, you know, being able to say that same thing. And then about the email campaign, you know, email got sent open, you know, clicked, landed on site, you know, paid money. That’s a five step kind of funnel. Being able to take make that funnel and then segmenting on which email campaign or which which, you know, what was the sign up bucket for this user? These users who came through the big PR push in November are these people who came through ad buys in January like that kind of segmentation gives you a lot more insight than just kind of a broad stroke, broad view of, you know, how’s the funnel look? That’s the difference between vanity and data. Vanity and data sciences is pretty crucial. I think a lot of people want to just feel a lot of people out there in tech aren’t data oriented. They don’t have the DNA, but they maybe they have investors or they read blogs that get them into thinking they need to do it. We need to do analytics. So there’s like a box that they’re trying to check. Yeah, there’s too much of that thinking. Yeah, exactly. Sort of. I’m trying to get knowledge. I’m trying to actually scientifically figure things out so I can drive my business with that stuff, with that insight. And, you know, you’re not doing cohort or deep segmentation analysis. You’re probably just sort of, you know, it’s vanity. You’re dissatisfied.

Bronson: No, absolutely. We actually just had Alistair Kroll on the show, one of the authors of Lean Analytics. Yeah, we talked about that a lot. He went through the Vandy Analytics and you know why they don’t matter. And so, you know, at this point, you know, refer back to that interview because you need to watch that again as well because it was exactly what you’re saying. You guys are 100% agreement.

Kyle: But I’ve seen a couple of his talk. Yeah, I.

Bronson: Think that books are really going to help kind of the industry, honestly. I think it’s going to set kind of a new level of understanding for analytics of what matters, which I think will help. I Oh, because the more you know, there’s an understanding of what you guys do and why it’s not vain and how you actually support companies is better. It can be for you guys as well. Let me ask you a question. You get to see a lot of these different companies. You see the ones who are involved with Bain Analytics. You see the ones who are just doing awesome, recording the right things, understand the data, making good business decisions based off of it. What are the companies that succeed at Analytics have in common? What what threads do you see that runs through them, if there is any? I don’t know.

Kyle: So that’s a good question, something I have to think about a little bit. You know, one of the things we like to talk about is kind of there’s like a continuum. You know, people talk about data driven, right? So my hand, right. Okay. So data driven over here like as if that’s sort of the Holy Grail and then data blind is over here. We actually look for something right in the middle, which is we call data informed, which is, you know, not not purely intuitive like an apple, but not totally mechanical like like a Google. A lot of times, I mean, we think that they’re the right way to run a business is really informed by data, but not necessarily paralyzed by it. Yeah, there are companies where, you know, there’s conflict on a decision and nobody has authority to make a decision. So it’s okay. Well, let’s just run an experiment. It takes like a month to collect the data. And then meanwhile, the competitors have already implemented one and failed implement the other and great they’ve made the decision. Yeah. In two weeks, you know. So I think, I think. Obviously now we’re talking about Apple and Google. These are two of the most successful tech companies ever. So tells you that there’s probably not a right answer. We succeed, you know, because of or despite where you are in the spectrum. But I think that, you know, if they just get this data is going to really contribute. You should you should sit still somewhere in the middle. There should be a sort of data informed approach.

Bronson: No, I think it’s a very powerful answer because, I mean, you’re an analytics company pulling away from the purely data driven side of that continuum. So I think that has power when you say it particularly, you know that there’s something to that for sure. And we’re at Google. So, I mean, you understand what the other side looks like as well. So I think it’s a great answer.

Kyle: Yeah. I mean, we’re make it we’re making things. We’re making products for people. That’s what that’s what that’s what you know. That’s what we all do. And, you know, there’s no algorithm for poetry, right? There’s something I always say. It’s like sometimes it feels like there’s companies out there. They’re like Vulcans from Star Trek trying to trying to reason their way into, like, creating, like, a beautiful song. You know, it doesn’t always work. What you can do is you can you can throw five songs at the wall and then test them. It’s like, you know, the music industry in L.A. has definitely demonstrated that you can you can use data, you know, to help select three, that you have time. But, you know, we’re definitely I might be the worst designer’s ever saying that, like, you shouldn’t be that strict about your kind of your analytics focus. And I think the Lean Startup Movement has done a lot to create awareness, but it’s also done a lot to make people feel like there’s a, there’s a, there’s a right way, like there’s a syllabus for actually making a great product. And there if there were, there’d be a lot more great products.

Bronson: And that’s right. There’s a lot of people that follow lean methodology that fail miserably because there’s there’s poetry and like you said, you can’t you can’t just make that. So I think you’re absolutely right. Let’s talk about Keynesian a little bit and the growth of Kenai itself. What have been your goals? Primary channels for customer acquisition? What have been the channels that you really get your customers through?

Kyle: So, you know, we’re we’re we run a very lean shop. I mean, we have a brand that’s way bigger than our sort of age and funding would would indicate. So, you know, we do it through guerilla. Like, we’re just totally online gorillas. Like we, you know, we’re hustling on chat rooms. We’re answering people’s questions. We’re making sure that the people like, you know, Maven first marketing kind of like the tipping point stuff. You know, we take that to heart, ensure that the people who really know what they’re doing know about us. It’s been pretty key. Our biggest channels, our blog by far, and you know, our blog. Again, it’s like going back to the data versus intuition. Like we take a very long approach on, on, you know, marketing. You’ve got on marketing, there’s another continuum. It’s like quantitative marketing, which is all about ad buy and like outbound and CRM and all this stuff put through rates. And then you’ve got branding on the other side, which is like that’s, that’s been around way longer than data science. So which is much more sort of how will this make people feel in general about us? And like, what’s the changes? There’s math somewhere changes like the likelihood that a random person won’t make a net promoter score and stuff. But you can’t you can’t run a split test on your net promoter score. You know, you can’t say. Let’s, let’s, let’s go there and let’s go there and be nice and see which one works. So you actually have to just make a gut home. We took a very long approach on that and it’s definitely paying dividends now. But it’s you know, we’ve been a company for 14 months and it’s slow going. It takes a lot of faith to do that. Like we’re not writing blog posts all of out. Like if you go to our blog, a lot of our blog posts about startup culture, about work culture, about, you know, social issues and tech that we’re passionate about and like. And then, you know, probably 15% of our blog posts have something to do with data, so and so the analytics really, it’s those other ones that drive a lot of traffic. So a lot of it’s like a position. One of things we think we didn’t really like. A lot of people talk about content marketing and blog based inbound marketing. They’re really powerful. But then they turn around and they they delegated the delegate being like, okay, now let’s hire a full time blogger so the team can focus on the real stuff. We don’t have a full time blogger, but everyone on the team has written a blog post with at least 5000, I’m pretty sure at least 10,000 uniques. So, you know, giving people a voice, creating an authentic voice as a brand and then giving each individual an authentic voice of their own, letting a voice that has worked really well for us, you know? So, you know, like my blog posts are going to be a little bit professorial, a little, little while short. You know, Dan and Mike over our CTOs blog posts tend to be way more narrative. A great storyteller. Our biggest blogger, Michelle, who’s one of our engineers, talks about like really personal narratives and extrapolating kind of like meaning and heuristics and stuff. So like if you go Google right now for start up negotiation, I’m pretty sure. So number one under just a blog post about that. Wow. Which is not exactly speaking directly to the tech, not like our our target is the technologist in a startup or the technologist in any company with between three and 20 developers. Yeah, they’re not speaking directly to them all the time. But you know in the long term, like so many of our. So many times we’ve had a conversation with like, oh, actually, yeah, we found you first, you know, with your demo day post back in June. And then, you know, then we saw this other thing on Hacker News and like found a tweet that went viral and know how we made an account. And it’s some of that stuff is hard to measure. It’s hard to attribute actually. Some of it’s impossible to attribute well, but we see it working. So, I mean, I think that there’s tons of leverage in that. That’s something that the big companies can’t do. They can’t they can’t build a product and company brands that are super laser focused, you know, like Google Reader can’t have a brand that does its brand is kind of dead now and you can start to build a lot of trust in the product that’s, you know, is just some product of some bigger things. So like as a startup, we have to use our advantages. This is a strength. And one of the weaknesses is we don’t have a huge marketing budget, don’t have a huge PR arm. It’s also a strength because it’s like, okay, what are we going to do under these conditions of constraint?

Bronson: Yeah, absolutely. You mentioned the word guerilla. You guys are just scrappy, you know, besides the blog, because obviously that’s a huge inbound piece of marketing for you guys. What’s been like the single best guerilla tactic that you’ve tried that just worked, maybe surprisingly worked, or you know that it had some blip on the radar for you guys.

Kyle: Sorry. You’re breaking up a little bit there.

Bronson: Sorry. I said. Yeah, you mentioned that you guys can hear me okay now.

Kyle: Did you ask that again? Sorry.

Bronson: Yeah, I’ll ask if.

Kyle: You can hear.

Bronson: Yeah, I can hear you. Can you hear me? Can you hear me?

Kyle: I can hear you. You’re. You’re jumping around a little bit, I think. My Internet. Give me 1/2 just to make sure.

Bronson: If you want, you can hang up and call me back and then it usually fixes it.

Kyle: I just need to check one thing to see how my. Okay. Actually looks pretty good. Can you hear me now? Yeah, I can hear you. Well, I just need to run a little diagnostic here on the Wi-Fi. It looks fine. Had a hiccup? Yeah, you might ask.

Bronson: Yeah. If you’re streaming Netflix, maybe you could just pause for a minute. No kidding.

Kyle: I mean, I just got the Spotify in, like, perfect detail, right?

Bronson: Yeah, exactly right. Now, I ask you again, we’ll start with that question. So you mention the word guerrilla. You guys are scrappy. What’s been besides the blog? One of the guerrilla tactics that’s really worked for you guys or you tried something a little bit subversive, a little bit unusual, and it worked a little bit. Any examples of those?

Kyle: Yeah. In fact, I saw one of them. There’s a really great marketer, this guy named Saul Colt. And if you haven’t found this stuff, you’ve got to find the stuff. He’s really all up. So we’re a Techstars company, and we went to the Techstars founder, Con, which is a conference for the Techstars company founders in Boston back in November and saw Colt was one of the keynote speakers. He gave a talk about about word of mouth marketing, which is another way of saying like guerilla marketing. Mm hmm. A lot of ways. And one of the things he said in his talk was try 100 things is like focus on, like, little things, you know, like that where you’re encouraging failure. You’re not putting everything behind a single tactic and you’re it’s forcing you to come up with quantity instead of quality ideas for things. And that really inspired, inspired us. So one of my team members and me, Michelle, we wrote lots and lots of ideas down and came up with all the stuff that we are still implementing rather random ones whenever we whenever we get the time. And one of the things that we did was so I wrote this tweet on our account. We were kind of playing around we’ve been playing around Twitter marketing like just, you know, Twitter marketing because it’s a new ad channel. And I found that new ad channels usually are undervalued. So that’s that’s a tip like go buy Twitter ads, you’re going to buy any ads is super undervalued right now. And what we did was we wrote this tweet that says, like, don’t click this tweet, because if you do, we’ll get charged. Instead, just type this in your browser team dealio, which is like a just a total joke. It’s like something that we knew people might see and rather than click it, they might type in the URL and match. Our Twitter won’t charge us if they don’t click it. But so the tactic worked, I guess. Well, it’s hard and you can attribute it. It’s hard.

Bronson: To measure because they type to do.

Kyle: Something to type in the URL. But the interesting thing is that that tweet got retweeted a lot like 100 plus times and that drove a lot of traffic that we could attribute in those. That’s all free. So that happened. Then I wrote a post which is like just a blog post that it ended up on the front page of our news for like 6 hours. And the blog post was the story I’m telling you right now I’ve seen. So I’ll call to that conference and then being inspired by them. And then one of the things we tried was writing this tweet. So we spent like $100. I don’t know, I was recently $100 on promoting this tweet and got just crazy traffic, like, oh, I wish there are a numbers in the post, I don’t really know. But so that works super well, just little. That’s a little bit subversive. It’s kind of like a lot of times people would respond and be like, this seems a lot like spam or whatever. And then they reply a little later and they’d be like, But touche, I checked out your site and I think I’m going to be a customer. Yeah. So that was definitely a little guerilla, you know, a little like we sponsored the Techstars party. We were one of the chief sponsors there. You know, Techstars has some big companies that spend a lot more and have a lot more funding and stuff. You’re one of the lowest tier sponsors. Mm hmm. What we did is we showed up, like, an hour and a half early for this is at South by Southwest. Two weeks ago, showed up like an hour plus early for the party so we could stake a claim on this really awesome booth. Like, we got, like, a, like a coffee table and a few couches, and we just, like, stacked up our, like, multicolored flashing, like, cups. Uh, the red, red, green, blue is kind of our color theme. So we stacked with these a pyramid of these cups, maybe like 100 of these, like, pint glasses. And everybody felt like we were like the anchor sponsor because we got so much attention, and everybody that was walking in had to walk past us and see this multicolored pyramid, you know? So that’s like physical guerilla. Mm hmm. But really, you know, it’s about trying as many things you can do as you can. And honestly. Even if you have a big marketing budget, it’s healthy to pretend you don’t. Hmm. Hmm. Or just slash it for six months and see what happens. Because I feel like without the necessity to be highly leveraged, people won’t. You know, it’s not about being smart or clever. It’s about. It’s about necessity. Like necessity is the mother of invention.

Bronson: So we can quote the same thing with stranger.

Kyle: You’re kind of your. Your ability to spend money. Then you’ll figure out ways to get a lot more leverage. Yeah. And bunch of. Bunch of more online tactics that, you know, we could dig into.

Bronson: Yeah. Tell us more. We have some ready. I mean, you know, this is good. Just helps us brainstorm and good ideas as well.

Kyle: Yeah. I mean, one of them, you know, one of them we just discontinued because somebody wrote us an email that said it was a little shady and would actually make a little shady, some stuff.

Bronson: Well, now I’m really interested in what it is it does.

Kyle: Storytelling. It’s all about tension release. Right? So all this stuff.

Bronson: No, no, you can’t. You can’t.

Kyle: Oh, there is a site I’ll type it in. Really great site. I’m sure you know what you guys do. It’s good to know about Cosmopolitan. Mm hmm. F0ll or WGN dot com, followed by this this awesome dude, miles. And what follows and does is lets you set up campaigns and basically set up search campaigns for tweets. And it will, on behalf of your account and will favorite tweets that match those searches. So we did this for a couple months. We got really good growth out of it. Like we were paying I mean, the prices have gone up now, but we’re paying like, I don’t know, like 20 bucks a month. And we were getting more than $20 a day. So you’re talking like 1% or less of what it would cost to buy Twitter ads to get that kind of exposure. And it’s all super relevant stuff because we’re favoring things like somebody mentions Google Analytics and the negative sentiment, just stuff like that. That’s perfect. We got a lot of growth out of that. Super, super cheap. I highly recommend it. I wrote Maximus. I mean, wrote in there like, oh, it seems a little spammy. Like, you know, I worry about ethics that are coming out and we’re like, Oh, huh. That’s a good point. Like, we don’t we? That’s really dangerous for our kind of long term brand perspective. Mm hmm. So we we just continued using it, but I highly recommend using it when you’re getting going.

Bronson: And to me, that’s not in the gray at all.

Kyle: Yeah, well, I was about it. He’s like, look, like this. This is what? This is natural is what happens like, yeah, really good for people getting started trying to get a boost of growth and like if they get to a point where the brand is big enough that, you know, they’re worried about the sort of image of using a tool like that, they just continue. So like, yeah, it makes sense, you know, which is very, very candid from the creator of this tool. Mm hmm. And and we were one of the first paying customers. You know, it’s only been a few months, but it has grown far. Jun’s definitely grown quite a bit. Yeah, like, that’s a really cool, really cool hack. Yeah. You know, little things. Like, even just. You know. Engaging people on Twitter that have reached just this is Maven. This is this is maven first marketing you know like people who have reach in your in your segment. If you if you. I mean, it could be looked at like you’re using those people’s reach. But on the other hand, it’s I think it’s authentic if you really believe that they and their audience need to know about what you do. Yeah. In which case, you’re really just. You’re trying to get a megaphone. That that and the megaphone needs you. Mm hmm. So engaging people that really that know this stuff, whatever your market target market is. So, like, if you, you know, if you’re like, a wedding photographer and you’re not engaging the dot.com, you know, which is a which is kind of that we’re working publication for for weddings. Like you’re doing a disservice to your customers and you’re not you know, you’re not hustling enough. Yeah. And so, you know, engaging those people on Twitter can be really valuable. Like one of the real secrets. Serious. One of the one of the secrets. I don’t why it’s a secret of the Internet right now is that, you know, Facebook is like a reunion, right? Mm hmm. You know, all the people when they add you on Facebook. Right. Like you’re going back to Twitter is a cocktail party and Twitter is a cocktail party 24 hours a day. You reach out to people on Twitter. If you if you send me a random email, I feel somewhat violated. Yeah. If you send me a random LinkedIn connect, I feel like stock. Mm hmm. Facebook connect. I feel really sucked. If you send me a random tweet, I feel flattered. It’s like somebody came up to me at a cocktail party.

Bronson: Well, that’s how you connected. I believe.

Kyle: It is. How you. How we connect. Yeah. Like, that’s one of the real secrets to growth, is that, like, Twitter has really democratized access, and that’s something that we should all be using. You know, it doesn’t really you know, it doesn’t really matter where you live anymore that much. It’s really about, you know, knowing who to talk to and having an authentic message, having something that they might be interested in to talk about. Yeah. And, like, you can just reach out. Like, I reached out to people who are literal or very close to literal billionaires and had gotten coffee. Yeah, because I wasn’t afraid to be like, Hey, I see you’re in town. Like, Let’s grab coffee. I look at your opinion on such and such and then like, you know, I’m not saying that’ll always work, but I’m sure they get flooded, but like it can’t hurt. And that’s kind of that’s unleveraged. That’s like direct, kind of direct hustle. Like it’s not leveraged unless you consider that they might be amazing and they might know the customer or the investor that’s going to get your company over the hump. Yeah. And if you’re going to recruit allies who have a lot of access, then, you know, then you’ve done your whole job in a lot of ways, like you did. It was recruit allies. You have a lot of access. Now you’re Switzerland, you’re safe. Yeah. So that’s I mean, just using Twitter in general, just basically there’s no real tactic. Message somebody with something interesting and try to meet up.

Bronson: Yeah, well, it makes a lot of sense as you watch kind of the growth of Teneo. Has anything surprised you about it? Because I’m sure you came in with some preconceived notions like this is how we’ll grow, this is who’s going to be talking about us. This is the way the chart is going to look. Has anything been surprising about watching people grow?

Kyle: You know? I mean, a lot of things have been sort of surprising. And sometimes even though we have preconceived notions, we have preconceived notions like we’re very strategic people and we talk a lot about strategy and a lot of theory about what should work. We all play board games and like hardcore games and stuff, right? And we had plans that even when those plans work the way we thought, they were super surprising. Like it’s been super surprising that, that our like, I mean, our blog for strategy kind of works like that’s awesome. I was I was in India like a month ago and I sat down for dinner or for lunch with like a table there, like five people, six including me. And somebody mentioned something about how a friend had told them to follow me on Twitter because I tweet about my culture, including a lot of and like work culture. And then somebody else said something about how he had he found us originally because of the blog posts that we wrote like a week earlier about our iPhone session, tracking systems and data, stuff like how to track people’s sessions in an iPhone at a model, a session, an iPhone app visit. And then so else was like, Yeah, you know, I like how your blog doesn’t only cover that stuff like this. Like your logo looks really great in the blog, you designer. I’ve been posting something about how we designed your logo and I was like, I’m in Mumbai right now. These are just random entrepreneurs in India who know who we are. Oh, that’s awesome. Like, we didn’t even exist like 14 months ago, and we certainly nobody knew who we are a year ago. So just the fact that it’s working alone is super surprising, even though it shouldn’t be surprising, like that was the plan. It’s just like you got to pinch yourself and it’s working, especially whenever you know, it took faith to keep going with it. When it didn’t, you didn’t know how to organize like, oh, that blog was got 40,000 uniques. But like we were the signups, you know, like the kind of short term outcome oriented thinking is, is valuable sometimes totally. Right. But sometimes, you know, faith can faith in your strategy can really help. Yeah. Like surprises, things that didn’t work the way we expected. I think that geography has really melted away in a lot of ways. Like approach is since it’s not verticals and it’s not geographic. Over four is like thin horizontal like developers and companies with between three and 20 developers. That’s very hot. So it’s been surprising how how geographically distributed. So like that Mumbai instance, just one of so many like I’m on my recent trip to the Middle East and India to do tech ambassadorship. Like, I just met people that know us, you know, people in cities around the US, you’re more likely to know us. For instance, I think I don’t have the data for this person, but I. It feels like you’re more likely to know us if you live, you know, like we’re a San Francisco based company. But it seems like people in, like Austin, downtown Austin are more likely to know about us than people in the valley 30 miles south of here, like in Palo Alto, which is kind of it’s interesting, like the, you know, the hipster first kind of sector. It’s really it’s really playing out. Like if I wear our keen eye a hoodie or like I’ve got a t shirt on that.

Bronson: Is like that.

Kyle: But if I wear that thing around in like, you know, in will say like one of the tech strips in Brooklyn, it’s so much more like it has really stopped me than I am if I’m in Mountain View. So it’s it’s been surprising that it had we haven’t spread like a virus from San Francisco out, which spread all over the place, which is in more in cities. And we’re obviously as we go up market, we’re probably minutes, you know, from downtown Chicago to the huge companies in Chicago and in the suburbs or from San Francisco is like skinny jeans and series eight companies that like the Valley was like big companies like Cisco and Citrix and stuff. So that’s how we go. What market is we radiate out from cities. But it’s been interesting to to see that, you know, the growth when you do Internet marketing geography just really melts away like the cultural ties are tighter between two city hubs than they are between the city and its own suburb. Yeah.

Bronson: Now, that’s great insight. A couple last questions here. Are there any strategies for growth that you haven’t yet tried but that you’re considering? They’re like, oh, we want to throw a bunch of money on pay per click or we want to try this other thing. What’s kind of on the horizon? Do you think you’ll eventually test out a little bit?

Kyle: Oh, man. I mean.

Bronson: Everything, right?

Kyle: So many, so many. Some of them are back pocket and we’re super transparent, usually after the fact.

Bronson: No, it’s like we’ll tell us what you can tell us.

Kyle: But, you know, so a few things like we’re really digging into education as a form of marketing. We education feels good. You learn a lot more by teaching than you do by sitting in a class anyway. So, you know, we’re doing it, you know, sort of on a scalable way now teaching small classes, teaching groups like, you know, we teach it like Hack Reactor, which is a bootcamp we’re involved with Cool Develop it teaching classes like Analytics for hackers that comes about that we’ve taught analytics trackers at a school in Denver which is a. One of our investors Galvanize runs this awesome co-working space in school that we’re thinking in more over the next few months to build. In fact, we have a column of 3 hours to talk about the class material to build a more long form kind of thing. Because, you know, this developer bootcamp movement is really important. We can help, we think a lot of people think can help with the unemployment or unemployment crisis that we have. People are dropping out of their undergrad, going to three months ago boot camp and getting a good job like after three months that’s awesome as the first two have been four years of student wound up. Yeah but we think that data science can really help help those people. Fine, fine. So you know we’re doing that in a non scalable way now. So time is actually growth hacking. It’s more like revving up the engine and kind of designing, trying to figure out what’s the material that works. But we’re going to be launching an online education thing, kind of like code academy, but for data science and analytics.

Bronson: Well, you got to come back when you launch that and so you can tell us all about it.

Kyle: Yeah. Got to love. I mean we we really think like and the reason that we think that that’ll help us with our growth is that the more people know about this stuff, more familiar with it, more concepts that they know, the better taste they’ll have in products and they’ll probably end up choosing us. Yeah. Which is a very long term approach. Yeah. So that’s, I mean, we’ll see how that works. Like education as marketing. I mean, it’s seems like it works with growing the Ruby community. Mm hmm. It wasn’t a business growing it. You know, it’s like the education and things like Try Ruby, how can you hack? You know, why is pointing guy to Ruby. These are things that happen, you know, between five and seven years ago that we kickstarted in the beginning because they made a bunch of people who are not developers feel developers all of a sudden now you know, rather than investing in changing people’s attitudes and tools. Voice, you’re just creating a generation of people who have a certain kind of taste around tools. Yeah. So, you know, that’s. That’s a hack that. It’s going to be ongoing and I don’t know. I feel pretty confident about that. That’s a 2013.

Bronson: Yeah, absolutely. Well, I think, you know, education is marketing is kind of an extension of what you’ve already learned from your blog. I mean, people don’t go to your blog because they’re not learning something. They go because they’re learning something. They’re seeing the world in a new way. They’re understanding something they didn’t understand. So I think it’s going to be almost an extension of what you’ve already kind of proven works, you know, in some sense. This has been a great interview. Have one last question for you. What’s the best advice that you have for anyone that’s trying to grow a startup right now? Because you provide the tools to help people grow and you yourself are in a startup trying to grow. So what’s the best advice you have for somebody trying to grow a startup?

Kyle: Best advice or best advice? I mean. Never stop studying, never stop trying stuff, you know, like, you know, I mean, I spent the weekend reading a couple of books, Antifragile and to sell as human because I don’t feel like I know enough. You know, never like, seriously, there’s no such thing as being too good at. GROSS So even if things look good, don’t get complacent. Like always push the curve up more, up, more, up to like. I think that’s pretty key. And I think I think don’t delegate marketing, don’t delegate growth. You know, like if you’re a CEO, I’m a CEO. My job is growth. I mean, everybody on the team is job is growth to some extent. But a lot of companies would be like, oh, yeah, you know, I’m focused on this thing. We’ve got a growth person over here, we’ve got our blogger, we’ve got a marketer. You know, that kind of when you’re small and you you compartmentalize people like that and compartmentalize functions like that, you give up the biggest advantage we have, which is that like I have like a dev department and a marketing department and a PR department all in my head and I’ll have that meeting. So like I, you know, I just like, oh, I need a tool that’s going to be the it’s going to do this. I’ll just make the boss like. Mm hmm. That’s the key advantage we have is how fast we are. And we’re fast because we’re just interdisciplinary, so we can’t give that up as we startups. That’s our key advantage. We have to use it. So I would say, because that’s the real advice, don’t, you know, don’t compartmentalize and delegate. This stuff is just too important. It needs to be all across your company.

Bronson: Absolutely. That’s awesome advice to end on. Kyle, thank you so much for coming on the program. You’re a wealth of information.

Kyle: Thanks for having me.

Are you an entrepreneur who is trying to grow a startup?

Get the strategies, motivation, and in-depth interview with all the details every week!