skip to Main Content
Back to Insights

Enterprise is Eating the Software World with Ed Anuff (Apigee / Google)


    Have you noticed how many venture-backed software companies seem to be focusing on B2B these days, instead of say, consumer apps, or developer tools? And inside of B2B, many of them are targetting the so-called “Enterprise” segment — companies with thousands of employees, who can afford contracts with annual values in the millions of dollars. What’s so different about the Enterprise? Why are we seeing this flight upmarket? Why is it so hard for some startups to make it in that industry? We discuss with Ed Anuff, formerly VP of Product Strategy at Apigee, which was recently acquired by Google.

    The Enterprise Trend isn’t New, but it has Changed

    Tim Anglade, Executive in Residence at Scale Venture Partners: The thing I wanted to chat with you about was enterprise software ‘cause, you taught me a lot about it. I came into Apigee when we met, with very strong opinions about software, and open source, and there being a right way to kind of build a company that is all software, and I was wrong about most of these things, I can say, and in particular, I think I was very wrong as a lot of people coming from an engineering background can be, thinking about the market for software, and in the past few years, it seems like there’s been a lot of movement, there has always been a lot of enterprise software companies, but it seems like there’s more and more of them, now, and more and more plays that are going to revolve around enterprise pricing of software, and enterprise plays, and I don’t know if it’s me just kinda being more aware of it, if you also feel like it’s been a trend actually, that people have been shifting towards enterprise software.

    “It’s been really fascinating to watch the experimentation that’s happened around how you sell to enterprises. I’m concerned actually: that’s been the only interesting thing to watch.”

    Ed Anuff, Product Strategy at Apigee (now Google Cloud): Well, no, I mean I think that it goes through investment waves. I think that, the interesting thing about enterprise software has been that, it’s really been a 20 year thing, which is that companies have been going and having to, by necessity, go and adopt internet technologies for their businesses. There’s been different waves of it. Way back in the day, every company needed to go and have an e-commerce site, or so everybody went and bought all the app servers, and companies like BEA, that some people may remember, did a lot of business in helping people get their applications online, or their e-commerce sites. My company, Epicentric, when everybody wanted to go and have a web portal, they came and bought software from us. You had people like Vignette going and getting content management and all of that. But that’s ‘cause the companies have always had to capture, that’s one aspect of it, that’s infrastructure. And then you had a lot of other stuff, like the productivity software coming online, and people going and buying the collaboration software, and Microsoft with Sharepoint and all of that. But a bunch of other companies. And the modern day equivalent of that is literally just Box, and Dropbox to an extent, but Box, the latest iteration of collaboration software. And we have all of the database companies going, and doing replacements, so, there’s this constant wave of enterprises, companies need to go and buy software. So, if you’re a start-up, and you’re trying to figure out where’s there going to be a way that you can go and solve a problem, the enterprise market is always really interesting. I think, it’s been really fascinating to watch, has been, kind of the different types of experimentation that’s happened around how you sell to enterprises. First, I’m concerned actually. That’s been the only interesting thing to watch.

    Tim: Right, you are correct, right, because it kind of like it feels like we’re seeing more and more enterprise plays, because people are still kinda coming out of the internet wave, and the mobile wave, and the cloud wave, and kind of everything that the companies have to catch up to. But the one thing that’s really been changing is like the sales models and distribution models in general. How people package that software, maybe that’s why we’re seeing more and more activity. And so, very noticeably, there’s been, I guess like a SaaS wave, where people were doing this kind of subscription pricing, but what else have you seen that’s very interesting?

    SaaS Changed Everything

    “Somebody asked the question what are the skills you need to know to prepare yourself? What should I learn, should I be learning Ruby? His comment was: you should be learning Excel.”

    Ed: Well, the biggest thing that’s changed, obviously, has been the SaaS play, and that’s, I think we all understand that pretty well, at least the basics of it. I think what people don’t fully appreciate has been… So I think most start-ups are really good at understanding kind of the first order, economics, of SaaS. What they don’t understand is how to get scale out of it. You know, it’s really funny, one of my colleagues from Six Apart, Michael Sippey, went on to run Product at Twitter. He did a, he was on a panel at South by Southwest, and this was after, I think, whatever the big collapse in 2008, or whatever. Somebody asked the question what are the skills you need to know to prepare yourself? What should I learn, should I be learning Ruby? His comment was: you should be learning Excel. And, everybody laughed, but he was dead right. Especially when I look at companies focusing on developer services, open source start-ups, I always think that in the back of my mind. You should’ve learned Excel. You know, when I see, and I’m not going to name names, but when I go and see blog posts about the open source business model, how it doesn’t work, and I’m like, you know what, there’s a guy who didn’t learn Excel. Because the reality is that you look at the companies for whom open source business models are working, they just did the math, and that was actually the most interesting lesson I learned many years ago in the enterprise space, was that you actually, your price, was like the last thing that you figured out. You could just backwards derive what your price needed to be from what it was going to take for your business to work. And the companies that did the reverse of that, that went and said oh, I think my software should be worth this much, and then sort of tried to take it from there, they ended up always in the situation where the economics didn’t work, the unit economics of what they were charging for, did not scale up to actually build a business.

    “when I go and see blog posts about the open source business model, how it doesn’t work, I’m like, you know what, there’s a guy who didn’t learn Excel.”

    Tim: I mean, I guess I see what you mean, it’s just like the big kind of complicated part of it is finding that, right? It seems like there’s this, high price point type of sales model, you know, like the whole suits kind of knocking at your door type of sales model that people commonly associate with enterprise sales. And it goes all the way to this box like model, where you assign a lot of seats, a substitution model, and everything inbetween, like what GitHub, I guess, and GitLab are trained to do, which is kind of somewhat different as well.

    Collaboration Software has a Natural Fit in the Enterprise

    Ed: Well, I mean, I think that GitHub and GitLab and so on, those are collaboration companies. You and I have had this conversation before. They have more in common with Sharepoint than they do with, and same, throw Atlassian in there, things like JIRA, so on. They’re more in common with Office365, then they do with developer technology.

    Tim: It’s really about sharing data and the company.

    Ed: Exactly, you can think of them basically as collaboration software targeted at the developer vertical. And, in fact, when you look at things like JIRA, and so on, they’ve expanded into project management beyond just that. But those are very different dynamics. And, in fact, Atlassian does a really good job of making that model work because they, in many ways, emulate the way that Salesforce works. In that, you start thinking about, okay, what’s my cost of acquiring a user, how do I go and basically colonize a company, in terms of number of seats. Once I have that, how do I then roll out incremental applications per user? Because you gotta have to have a growth model. And this is one of those things that we always forget, and I, no matter how many times I’ve lived through this, I always, it always sort of strikes me as like, this new discovery, that it’s like wow, you know what, selling more stuff to your existing customers is a hell of a lot easier, that’s where your growth happens, rather than having to go and talk to new people.

    Tim: But a lot of people don’t get that. I mean I was just talking to a company, a very famous company, last week, that, was rolling out a bunch of features, and I’d like, you know, specific plans to like, give them for free to new users and enterprise users, and not charge for them. It seems like that’s a mistake, right, that a lot of people still seem to be making.

    Ed: Well, you’ve got to be growing on all axes. On the converse of things, you’ve got the classic dilemma of an enterprise company that has its hundred big customers that it bends over backwards for and keeps happy, and never actually reaches the larger market. So, don’t interpret what I’m saying to be the idea that you should not be focusing on getting new people on to your stuff, but, you know, I do think that when you get to that point of, you also have to be thinking about how do you actually sort of increase the number of services, different things, that your existing customers are

    Tim: Right, that thought is interesting ‘cause, I guess, you know, in my mind, and maybe that’s wrong, too, it’s like a lot of enterprise software wasn’t about volume, it was just about big ticket items that you could sell to a few, to the Fortune 500, right, and kind of sell them as a big deployment, a big site-wide kind of deployment, but now it seems like a lot of it does go through volume, right, of how many employees can you sell this to, how many features can you get each employee to buy into, and it’s not just selling people databases or ERPs or whatever, it’s also a lot of almost consumer-like type of volume play in terms of the number of seats you’re selling and engagement throughout you’re driving for your application.

    Ed: Well, you’ve thrown a couple of ideas out there, and so, I think that one of the things you have to figure out, when you say it’s, and this is part of why enterprise, when so many start-ups go and throw the enterprise term out there, as kind of this magic umbrella, like, oh no, we’re not doing consumer, we’re… it’s a lot more nuanced than that. And then you start to get into, okay, are you applications or infrastructure? And that’s like the first cut. And then, if you’re applications, is it per user based, you know, or is there some other metric. And, then if you’re infrastructure, you’ve got a whole other can of worms.

    Tim: Right, and we talked a little bit about that, right, we talked, we mentioned quickly like open source plays, those are your straight up like closed-source big license plays, what else do you see in that kind of, on that side of the business? ‘Cause I think a lot of people our feeling was kind of, at least, some of the current and emerging pricing models for applications, but infrastructure is,

    Infrastructure Software gets Real Interesting

    Ed: Well, I mean, again, when you’re in an infrastructure it gets real interesting, I mean look, this is the whole… So, a lot of people go and talk about big data. You and I have had this conversation quite a bit about big data, so. A lot of the data, a lot of the data that people deal with in start ups, for example, or even, a lot of these large start ups, large tech companies, that we go and fawn over in terms of the data sets that they work on. It wasn’t like you couldn’t do those in Oracle. Wasn’t like Oracle just wouldn’t scale. You know, not to sound like an old timer, but a lot of people, who have never actually used Oracle, would go and say no, Oracle could’ve never, Yeah of course they could’ve, actually. Oracle would have found a way to do it. But the thing is, that the unit economics of big data were fundamentally different than the unit economics of the way Oracle did business. So Oracle, of course, would run your credit card transactions, which are obviously high volume, large datasets, and so on. They would charge you an arm and a leg for that, because if you think about it, what’s the business value of a single record, in that situation? Like, think about what would happen if I lost a record that was for a credit card transaction? Or a deposit, I mean that’s actually a pretty significant thing. The big data databases, they had a different set of trade offs. They were like, if I lose a single record, well I’m using that database to log page hits to my website. I don’t care if I lose a single record. So I’m willing to go and basically make entirely different economic trade offs for what database technology I want to use. Whereas, Oracle is about like no single record ever lost. And I know this is oversimplification, there’s some people who are going to be like DBAs and stuff are going to be scoffing at these assertions. But no, that’s really what it came down to. And what Oracle found was that, there was like no way that they could price for the situations where people had billions of records of which they were happy to lose 2% on, versus the situations where people were like, where every record was worth this much. You basically have to sort of try to figure out what the value is, and base your pricing around that. And, it became, it becomes very hard. And that’s where the disruption opportunity is. Where if you are doing Cassandra, or you’re doing MongolDB, or you’re doing any of the, or CouchDB, or any of these new class of databases, where the customers have different economic trade offs, and it’s almost impossible for the incumbent, in a different use case, to go after that.

    Tim: Right, and I guess the gap, right, for people that are just getting started in this, or want to go to an enterprise, is lack of awareness of those business objectives. We think about consumer applications, oh, like, I’m a consumer, and I would want an app that does this, or I want a game to catch Pokemon, or, you can kind of relate to it, but really, even if you run your own company, on some small scale, relating to the business objective and trade off and economics of something like a Fortune 500 company is pretty damn hard, right?

    Scaling to an IPO on Dozens of Customers

    Ed: Well, so the issue, the challenge you get into, is to try to figure out, it’s like the old adage says: there’s only 500 companies in the Fortune 500. So, then you have to go and figure out, I’ve seen plenty of start ups over the years, go and have sold to some large percentage, or decent percentage, like x percentage of the Fortune 500, 30 percent or something like that. And you’re like, wow that’s fantastic. But, what you find, is that some of those are doing fantastically well, and others are challenged, and you’re like, well, with all these great customers, what accounted for it? Well, the question was what was it that they got? Were they in some particular little pilot project? Were they in a strategic project? Were they, did they have a foothold that allowed them to actually, within each one of those companies, go and expand to dozens of projects? Did they, could they expand to thousands of seats, if it was per user? Like, those are the questions you have to ask.

    Tim: Yeah, and no, you’re right, it’s a multiple layers of difficulty, right, you kind of have to understand, whether you’re more application or infrastructure,

    It’s like the old adage says: there’s only 500 companies in the Fortune 500

    Ed: What’re you going to do with your technology? You have to think about an area near and dear to your heart, right. And, the question now becomes, and you’re selling to the enterprise. Is this for their consumer-facing app? Is it for their employee-facing apps? If it’s a consumer-facing app, how many consumer apps are they building? The reality is that for any of the major brands that are part of the Fortune 500, they have at least one, probably. Some of them have, without naming names, I know some of the major consumer brands have, maybe between their, for different languages, and different markets, they might have actually thirty or forty. For consumer markets, I’m talking about putting it in an app store, right. Now you’re thinking about, okay, how do I price for that? You’ve got to figure out, you’ve got to basically backwards intuate it, because the reality is that you need to be figuring out, how the life, if you are going after the enterprise,

    Tim: You need to make it to the IPO in a way, quote unquote, you know, with 500 Fortune 500 companies, that are going to, some of which are going to buy your product, and they all have between, one, two, 30 apps, so the math has to work that way. If you think your value is tied to the number of consumer applications…

    Ed: Again, going back to that question of, doing, being able to do Excel. You’ve gotta figure out a way that you’re going to get, ideally, you’re thinking, assuming that’s your end state, but regardless, I mean, you’re talking about some areas that you’re getting to a hundred million a year in revenue. And ideally more than that, but just to be thinking about that, just to be at the point where you’re thinking about doing an IPO. So, again, how are you getting to the point where each one of these customers is spending at least a couple of million with you per year. How does your $50 a month pricing model, or call us for 75k a year, that we see on all these developer websites. How does that get that, for enterprise? How does that add up? It doesn’t, for a lot of them, and that’s where they dig themselves into the problem. So, again, that’s where I go back to, so then you back out of it, and you’re going to say, okay, well, that means I need x amount of projects, right, okay. And how are you going to get to those multiple projects?

    “How does your $50 a month pricing model, or call us for 75k a year, that we see on all these developer websites — how does that add up? It doesn’t, for a lot of them, and that’s where they dig themselves into the problem.”

    Tim: And we talked about it, it’s either more seats if you’ve seen less employees, or more features, or you have someone to cross sell or up sell, or just have the higher. And one of the things we did at Apigee that was awesome was just driving up the average contract value. Just having this constant push to try to get more value out of the product that we have, and get more dollar out of that.

    Ed: Well, I think you will see, for any company that’s successfully executing on the enterprise, and I think this is the thing that over the years you and I have had this conversation, we’ve seen it looking at plenty of companies. The ones that make it work, have figured out how to do it. And, the ones that have gone in and embraced this completely friendly, self-serve pricing that translates into 20, 50k a year, but for companies of that size, the math doesn’t end up working.

    Tim: Yeah, it’s a really tough situation. Cool, that makes a lot of sense. Thanks.

    Back To Top