Moving Crunchbase From a Freemium to Subscription-Based Model

Jager McConnell, Chief Executive Officer @ Crunchbase; Jordan French, Executive Editor & Co-Founder @ Grit Daily

Ascent Conference 2019

Jordan French [00:00:04] For the uninitiated, Jager, now they’ve heard your name a few times. There are few, though, who don’t fully know Crunchbase in its present form. What’s in most of our minds is what it used to be, which is just a database, but it’s a lot more than that.

Jager McConnell [00:00:19] Yeah, that’s right. Actually, I was just curious. Just show of hands. How many of you don’t know what Crunchbase is, just so I know. Awesome. Well, thank you for coming. Listen, I’ve got a few. I like that so much. Crunchbase is we try to be the best private company database in the world and we are building applications on top of that. So we think of ourselves actually as an application company and not as a data company as we’re trying to figure out how do prospector’s like maybe you’re a entrepreneur trying to find your next investor, maybe you’re a salesperson trying to find your next prospect or investor try to buy your investment. All of those are prospective use cases. So we’re trying to build prospective applications on top of what we would argue is the world’s best data platform to build applications on top of.

Jordan French [00:01:04] And especially for some of the crew who are uninitiated, don’t don’t really even know the name. Let’s let’s if we could just establish size and even some of the volume, like how many people are visiting.

Jager McConnell [00:01:17] So we’ve got about 60 million people using Crunchbase today. About 55 percent of them are international. About half of our companies in space are international. So it’s a it’s a global thing. And then if you look at, like, the percentages of used cases, you’ll see about a quarter sales people and BD people about a quarter are entrepreneurs, about a quarter of our market researchers, and they’re sort of everyone else. And then there are like investors, job seekers, because, of course, you’d want to find companies to work at. So there’s like a job seeker use case. It’s very it gets the slices get smaller and smaller within that remaining 25 percent.

Jordan French [00:01:54] And we’ll definitely talk about those slices Jager and set them aside for a bit because they specially relate to your move from free to premium to to the to the freemium model. I want I want to go back in history in time a little bit because a lot of people recognize the crunch part of the name, and no one ever gets to ask this question. But if you could sort of dispel or explain some of the history in the relationship The Crunchbase that might help the audience, too.

Jager McConnell [00:02:23] Yeah. So it’s it’s I’m Crunchbase CEO. I joined in September 2015. But Crunchbase has been around since 2007. So how is that possible? What was part of TechCrunch? And so TechCrunch is a journal sort of news source and they are using crunchbase as a sourcing tool to help find companies to write about. So as 100 or so countries was 100 percent user generated content, people would go and say, oh, I want to be written about every public company in the country, sorry, TechCrunch. And that was how it was for years and years. And then eventually TechCrunch got bought by AOL, which got by Vermont, by Verizon, and then they had this thing called Crunch Base. So they valued at zero dollars and they said, we don’t know what to do with this thing. Should we shut it down? Or maybe there’s value here. Maybe we should spin it out. So they started talking about spinning out. I heard about it. Just heard about it, which is what the VC did it and we spun out as a company. So ever since then, we transitioned from an ad based model, which is what we were when we spun out trying to make money from ads. And now I’m an application guy. I used to work at Salesforce, like, let’s make it an application company and build subscription businesses on top of it. That seems to be working a lot better.

Jordan French [00:03:38] Yeah, definitely. So interpret that back to you. Is it was either shut it down, throw it away or let’s just let’s just give it to Jager.

Jager McConnell [00:03:45] Well our give is a strong word. Its its emergence invested capital. They bought a large part of the company Verizon stepped back to, had a much smaller percentage of the company. Then they were able to connect with me and convinced me to be the CEO.

Jordan French [00:04:03] Certainly. And I promise we’ll still get to the premium funnel. But there’s one interesting, almost ironic tacit and that is also that crunch base also does news and is in some way in competition with its former culpably, its former sort of a sister company.

Jager McConnell [00:04:22] Yeah, I mean, it’s something more of a partnership like we with crosspost content. We we are definitely friends. Not even in front of me is like we’re definitely friends. We share they give links back to us. We give links right to them. We’re still on TechCrunch as navigation. So it’s a very it’s a very mutually beneficial relationship, not at all competitive like which is why we do podcasts with them together in harmony, sweet harmony, you know.

Jordan French [00:04:52] A rarity among among perhaps companies these days in a way. But in its evolution, I think it’s obviously quite fascinating for a lot of the audience. Is part of that evolution, though, is a shift also in your. Business model, and it in our minds, countries used to be free. It hasn’t always been that case. But before we even unpack that, why even why move away from the ad stack that you had in the first place? If if that seemed to be working as a revenue model.

Jager McConnell [00:05:20] It wasn’t working. That’s that was the problem. If it was working, they would have wouldn’t have spread it out. So I shouldn’t disclose this, but circle of trust in the Internet and so on. So we were burning about a half million dollars a month when I when we spun out just paying the salaries of the people and making the service go. And we are making a million dollars a year so that math doesn’t really work out really well. So companies go out of business really fast with that sort of model. So so we it was kind of scary, obviously will be spun out. We had about a year and a quarter of runway, so I was like in a year and a quarter, can you build a tool, does application and keep talking about this prospect to launch it, sell enough of it to to raise a Series B before you run out of money and you’ve got to like you’ve got a year and a quarter to do that. And I wanted or it was a little sketchy touch and go because it wasn’t clear we were going to be able to do. And as we building it, of course, naturally, you discover a lot of the things that you didn’t expect that you have to also fix before you launch that it was it was a good time.

Jordan French [00:06:29] And we’re we’re all already sweating. I’m thinking about four hundred and maybe twenty days, maybe a year and a quarter. We we can empathize. It doesn’t seem like necessarily a natural fit for a user base that an audience that was just accessing adding data, making it more valuable to then sort of throw up what seemed like a paywall went into that thought process.

Jager McConnell [00:06:54] Sure. So the the way we think about it is, yes, everything about Crunchbase was free. And you could just you could use it, crawl it, take advantage of it. You could try to do a denial service. There’s a lot of things you could do before we spun out. We had to protect the service. We had to protect the data. So we have now a Ridgewell and a paywall. The the paywall came up first and the paywall was if you want any of the new things that we have built and that’s really these Prostratin towards the monitoring tools, the things that help you find the companies that you need to do your job. That’s a new thing that didn’t exist before we spun out. So we’re going to charge you for that. And that seems like a fair deal. And then with the Ridgewell, which is actually more of a recent thing, honestly, it’s if you are clicking around and browsing, a bunch of the service were to ask you to tell us who you are. And there’s reasons to do that, both from a marketing and selling standpoint for sure, but also user experience. And probably most importantly is it helps us protect our service, because as we get bigger, people call us like crazy, like 300 gigabytes a day of people trying to pull down our data by putting up walls. It helps us identify those abusive users and stop them.

Jordan French [00:08:09] Certainly. And it’s, you know, indisputable. There’s a lot of valuable data in there, especially that, you know, the traffic itself, our user base, you know, speaks to that, you know, look at the scoreboard. But but certainly when you consider when you consider the the the user base Jager and. Apologies, apologies, I was rather.

Jager McConnell [00:08:44] Sort of so good, just just slow you down. I love that.

Jordan French [00:08:48] Yeah, I had a I had a train. I thought, well, let’s let’s let’s turn back to the funnel really quickly. So just to establish the trenches, because in the audience’s mind, when you say Reigel, what you mean is registrational. Yeah. And that’s not necessarily something that people have to pay for, is that right?

Jager McConnell [00:09:05] This is something that so we we track sort of user sentiment and a lot of ways some of that is just through engagement and bounce rate. So how many people come to the site? Are they clicking? Do they leave the site right away? Another thing that we look at as scores US net promoter score for those folks who don’t know. And that’s just a methodology to essentially survey our users and say, is this something that you enjoy? Is would you recommend it to someone else? So we track those things. And as we put up walls and paywalls, we see how people’s sentiment changes. And certainly as the CEO, a lot of people come up to me and say, oh, you put a paywall up. So now I can’t see profiles past the second or third profile. And then I have to explain to them. And this means that we have to we have to do a better job with the product. Hey, that’s actually a free login thing. We’re just asking you to go and register. And we’re like, if you don’t want the marketing email that you’re definitely going to get after that click unsubscribe. Right. You still have a log in and now you’re able to go and use the site almost like it was when we spun out. And that that’s like we need to do that and we just we have to do a better job, communicate. But from our perspective, it’s still 70, which is like ridiculously high salesforce. It was much, much, much lower than that. So we feel pretty good about that. And our engagement is actually going up. So we have many, many when a user comes to country, stood there for a long time, which is great.

Jordan French [00:10:32] Certainly. And I’ll pull on one little thread there, Jäger and that is you mention you could do a better job on something specific. It sounds like that was articulating to the audience when they go to space that that. Right. Well it’s really just an email, a login that’s free. Is that specifically what you’re referring to?

Jager McConnell [00:10:49] This is a I’ve seen screenshots on Twitter, like ohh Crunchbase , has a paywall now. And it’s like, no, not really. It’s actually what you took a screenshot of. If you read it, it says it’s free. Just we need an email address. And again, without it, it’s very hard for us to do a lot of things that make the user experience better. But also it helps us make people not like pulled out our entire site.

Jordan French [00:11:13] Certainly. And and for those in the audience also who you know, media, for example, grittily, we consider a paywall. I think there’s others with various models are starting businesses that are considering paywalls to how do you sort of decide where that that perimeter is jagga on, you know, how much you want to give away and then what should be paid for after.

Jager McConnell [00:11:33] Yeah. So and this is like if you are yourself evaluating if you should do for freemium or paywalls, a critical part of that is experimentation. And we use we have an AB testing framework. It’s actually a service because it’s called split. I oh check it out. It’s pretty cool and allows you to go and do a B testing across your entire application in our case, our site. So another problem that I have with the complaints is it really depends on what cohort you’re in. So there’s different cohorts based on let’s say it’s 20 percent of our traffic is seeing a like it’s the second profile view where you have a Ridgewell for another 12 percent. They don’t have a register at all because that’s how our control group there’s another twenty percent that has it at five, you know, so so by doing that, we can go and try different things and see how the user engagement, a user experience changes based on science since he like what’s what’s going on in each of those different scenarios. We do that on the Ridgewell. We’re doing that with the paywall. We do that with changes of features. We will move things around. All of that are those are all experiments that we’re doing and we’re using the same testing framework so that we can scientifically say this is actually decreasing engagement. Users are unhappy with this in a material way. Roll it back and to roll back. It’s just a switch. So some of the stuff like there’s there’s the most aggressive cohort like where we’re like, will they do this? You know, like that’s like maybe at five percent of our traffic and we have enough traffic where we can have that, but still be a meaningful amount where you might have the price go up really dramatically because we want to go and see if the if there’s a for certain types of users, if they’re willing to pay more and what happens. And there’s another cohort where it’s even less where we can go and see would we be better served having even a lower price. So there’s all sorts of different AB testing splits that we’re doing right now.

Jordan French [00:13:29] Certainly I think we don’t impact that all day. It’s always fascinating, especially the data wonks in the audience. You mentioned pricing.

Jager McConnell [00:13:35] Datawonk’s.

Jordan French [00:13:36] Data wonk’s. This is a real.

Jager McConnell [00:13:38] I’ve never heard of that. Okay.

Jordan French [00:13:39] We can.

Jager McConnell [00:13:40] How many people have heard the term data wonk?

Jordan French [00:13:42] Well, how many are you? That’s a good number. How many are data ones? How many of you are data wonks know?

Jager McConnell [00:13:49] And I saw one coffee cup go up.

Jordan French [00:13:52] So there’s the rare breed like ENTJ personality to learn something new. Awesome. So that’s especially useful advice for those who have the traffic had this go up. You mentioned a cost structure, though, earlier in the conversation when you spun out. There’s people doing things we in our minds can imagine relates to quality of data. But it seems like you’re at an interesting intersection all the time where you have to figure out what the data is worth and the quality of that goes up with the more that you spend. So since you mentioned pricing, how do you figure out what you should charge at a given time for a certain subset for what they’re after?

Jager McConnell [00:14:31] Sure. I mean, it’s there’s a lot of ways of doing that. One is we did we did a lot of focus group testing to sort of see what is of our of our power users in particular, what is the sort of how how OK, are they spending a large amount? And what’s interesting is, you know, there’s users that will pay nothing, you know, in the focus group. And there are some users that were willing to pay a thousand dollars a month because that’s really what if you look at a lot of our competitors, that’s what they’re charging. It’s like, oh, cool. You’re like this other person. I would pay the same as what I pay them if you had these extra features or you have this data. So we had to sort of pick something in between. We actually went on the lower end. It’s three hundred forty dollars a year compared to like ten thousand dollars a year for some of our competitors because again, we were trying to democratize the access. Right. We want to be able to make it available to the people that like maybe can’t afford the ten thousand dollars a year one. So. So you have to sort of balance that. And so that was a big part of it. And then, of course, there was actual experimentation where we go and see how people react to different prices. And then there’s also just the math of the business. Right. It took like people to know this. We spent something like eighteen will spend something like seventeen million dollars to make crunch base be the data set that it is for our users. That’s a lot of money. We need to have a business model that can support that. So if it was a dollar, a user a month, we couldn’t do what we’re doing. So the data quality would decrease. In the end, when I’m looking to 2020, we’re looking to double that investment. And that’s a lot of money. And most of it a lot of it is to go and push into the data and go and make the data even better and do better insights and deeper insights. And we were talking about 30 million dollars plus in costs. Hopefully we can do a good job.

Jordan French [00:16:18] And you mentioned a really specific number of those. Three hundred forty eight dollars. Why not? Why not 347, Jager, 349?

Jager McConnell [00:16:26] Its worthy of an experiment because for some reason we’ve always done this this math this way. And it’s twenty nine dollars for 12 months. So twelve times twenty nine is three hundred forty eight. There’s probably something to be said like should we try four ninety nine or three ninety nine and to ninety nine. I don’t even know when the last time the study was done on the night is better than the zero. I feels like something in the 1920s but that’s what we, that’s why it’s that number.

Jordan French [00:16:53] Yeah. It certainly works as long as it doesn’t. And I want to challenge maybe the point, the underlying assumption of all the cost base also, at least in some of the audiences minds, is our other sort of maybe in more encyclopedic databases like Wikipedia, they have all volunteers. Why why have all this 17 million cost structure at all? Why not have a volunteer user base jagga that moderates that that vets’ that verifies each of the edits so that the integrity of the data remains or stays high.

Jager McConnell [00:17:27] And Wikipedia’s case, I think they have something like 300 or 400 million users. If I’m ever going to have 60 million in a world where we had that that number, maybe it would work. But the reality is that most people who come to crunchbase simply want to consume. They don’t want to contribute. They don’t necessarily even have anything to contribute. Nine times out of ten. You need more data wonks. I need the world’s more data wonks. Really? So so that’s part of it is is we have a consumption based model rather than a sort of contribution model. Again, it used to be one hundred percent user generated content back five years ago, but that seventy million dollars, of course, I was talking about that’s doing things like we have 4000 partnerships. Right. Who’s managing those partnerships? How are you finding more partnerships? That’s with governments. That’s with events, that’s with investors, that’s with accelerators. And that’s going on giving us direct primary source data that no one else has. So that’s an example of costs. We’ve got the five minute warning in the back there. We also have things like data science and engineering going on, doing, crawling and finding the data and absorbing and all of the news that’s in crunch basis. Automatically, Poplin gets all the companies. It makes like that’s hard stuff and that stuff that keeps the date alive, that’s what keeps our users happy. That’s why when they’re prospecting, they know that our data is pretty close up to date. No date is perfect, but ours is pretty damn good compared to some of the competitors. Then, of course, I’ve got our own data team. So that’s, you know, all in 60, 70 people who manually are going and updating that data where the machines can’t do it automatically. All those things add up to cost. And that’s why the price.

Jordan French [00:19:06] Yeah, certainly we could pick that apart probably for a few hours. One more question and I’ll go definitely go Q and A and I’ll drop down and and come to commandeer the mic, get it to some people to ask you questions directly. Jagat I know the ad questions before we started. I want to finish on this though, and that is 60 60 million uniques in that audience every month. You mentioned international these twice earlier in the conversation. Is there a cohort that might even be here in part that’s missing from countries that you’d like to to to be part of that you haven’t quite been able to touch enough yet?

Jager McConnell [00:19:41] No, not from a use case standpoint. So so when you like, we obviously there’s competitors all over the world and we don’t look at all of them. And usually they’ll say that our data is better than ours when we go and actually evaluate our data, they like crushes them. There’s a competitor over in Europe, for instance, that says we have more funding around than anyone else, like we’ve got actually twice as many as they do in Europe.

Jordan French [00:20:02] Or crunches them, Jager.

Jager McConnell [00:20:03] Thank you. So so not really all the use case. Pye’s, if you like, look at a specific country. The use case of of who’s using us now is actually pretty similar globally. However, where we want to improve is the breadth of companies. So in China, for instance, there’s millions of companies that we don’t have in countries that probably should be in countries. The challenge is not just how do you get them, though, but it’s also how do you not completely destroy the user experience for everyone else if there is a lot of Chinese companies. And there so you search you just got all these like Chinese characters and things that just don’t make sense, or the Chinese and then the English blending together, all that stuff is our thing, things that we’re working on that next year.

Jordan French [00:20:46] So it is certainly a blend, but it sounds like you’re more than well, open and welcome to that to that data. I’m going to talk to Mike down. If you could just raise your hand. We can definitely get at least one, perhaps two questions into Jager real quick and just say your name.

Audience 1 [00:21:03] I’m Ivan from generic photos. The question is, how do you divide your time and energy into your team’s resources between actually creating the product, generating the data in developing the website and your efforts to monetize it?

Jager McConnell [00:21:20] Yeah, it’s a really, really good question and it’s a hard one to answer just a few minutes. But fundamentally, it’s if you if you think about our total company, it’s about a third on go to market. So that’s salespeople, marketing people trying to figure out how to talk about it. And it includes like onboarding, experience, retention, all that kind of falls into that some of that bucket. Then there’s a third is the data. So that’s where we’re going and spending a lot of time on just making the data better place to be. So that’s our data platform. APIs are involved there that day to clean up all of that hides and then that other third and then the final third is the application. So that’s why we’re building that’s where you find certainly the monetization piece for us. The blend between the monetization piece and the free stuff is actually it’s the same teams working on it that will change over time. But that’s how it is today.

Jordan French [00:22:13] Certainly. And we’ve time for one more question. I’ll Walkertown, just raise your hand and you can pop one to Jager. There we go. Thanks. And just say your name.

Audinece 2 [00:22:24] Julia McAllister from Def Method. I’m curious about we talked a lot about the monetization of the platform. I’m looking forward to the future. Do you see there being more revenue potential from just providing the raw data and having the data be accurate and correct? Or do you see there being more value in analyzing that data for the end user, kind of the model of don’t make me think and then providing them the insights that they’re looking for is that they don’t have to do the number crunching themselves.

Jager McConnell [00:22:52] Yeah, it’s definitely the former or sorry, the insights of the latter. I think you just said that’s where the the applications are. That’s where our users are. That’s what our users are looking for. When you think about the prospecting use case of I can tell you, hey, entrepeneur, these are the investors you should talk to next, because they’re the ones that should invest in your company or hey, salesperson. These are the companies that you probably care about because they look like to you typically sell to and they just raise money. That’s why we play very, very well. There is still the raw data play is about 40 percent of our revenue today, and that’s the TAM there. The total addressable market is a little too small for us to sort of bet the whole business on how many. Yes, we have about 500 applications using Country Day today in their own applications. I mentioned some like Yahoo! Finance, LinkedIn market. Watch Business Insider, there’s a number of those folks if but there’s only so many of those out there in the world and I wouldn’t want to bet the business model on that. So let’s focus on on adding value to the 60 million people coming Crunchbase.

Jordan French [00:23:52] That’s a great question. Do that’s all the time we have. If you could just join me. Big round of applause for Jager McConnell. Busy man.