Editor’s Note: The New Stack Makers is an ongoing series with the people who are leaders in app development and back-end infrastructures.
Just before the launch of The New Stack, I had a chance to talk with Adrian Cockcroft of Battery Ventures. The following is the first part of the interview with Adrian that we did at the Monitorama conference in Portland. We cover the research Adrian and the Battery Ventures group are doing on the SaaS market, We also discussed the new requirements for getting products and services to market and the role of developers and operations in that mix. Adrian is speaking today at the Structure Conference in San Francisco.
Adrian Cockcroft: What I am interested in is year-on-year on the cloud. This is the baseline. We will see in a year’s time which bits of the cloud, which buckets within SaaS, are growing faster. What are the new buckets? The hard part is getting any real data on what’s happening in public and private cloud spaces – sort of what is going on really in Amazon versus Google versus Microsoft. So everyone wants to know “How big are they?” I don’t really have any public information that anyone else doesn’t have, so I’m trying to gather together what the collective wisdom is from everyone else. But the interesting stuff that I haven’t seen done before is a real breakdown of SaaS.
Alex: How are you doing your data research?
Adrian: If you are a VC company you see everybody. We’re doing research all the time in order to make the right funding decisions. You look at thousands of companies, you invest in tens, that’s just the numbers. But you’re still looking at thousands. You’re doing detailed research for hundreds. And that research just sits there. I’m leveraging a bit of that research into an aggregation and part of it is to get Battery better known, as I’m doing some interesting stuff. We have something we can market as an interesting thing to talk about. That’s part of it. But part of it was that I just keep getting questioned: “What’s big in cloud and what’s really happening?” You see people saying the cloud market is a hundred bazillion whatever, most of it unsubstantiated. When you drill in, you find they were making numbers up, top down. Enterprise markets are trillions and trillions, so it’s got to be some percentage right? So we’re trying to go from the bottom up to see if it makes more sense.
We’re looking at the money that’s being invested in cloud companies rather than looking at the money that’s being spent by enterprises. I think that’s more of a leading indicator of where people think future value will be.
Alex: Because they’re investing their money actually in the technology itself.
Adrian: Somebody somewhere believed enough in the company to put a ten million investment series A round into it.
Alex: And they had to base that on something?
Adrian: Yeah. And so they add the wisdom of the crowds as measured by the money that VC people are putting in that’s publicly recorded, just it’s not everything. But it’s a substantial amount. It’s a theory. We’re still collecting the numbers. We’ve got until the end of June. We’re burning through the days and weeks a bit too quickly on this. And then there are a few research tools that you can use for digging up all of the raw information. Once you get the list of companies, you figure out how to bucket them into a segment. Collecting, finding out who’s got funding when it’s made public, there are tools that will dig that out for you.
We’re trying to figure out who the breakouts are, which sections, and what is the rapidly growing bit of the cloud market that no one has noticed yet, or whatever.
Alex: Any hunches?
Adrian: The late adopters are coming online – CFOs, people like that. People that historically were conservative, now they’re becoming data driven. So the CFOs are getting into big data. Now they need their own Hadoop cluster, or whatever, their own thing. But traditionally they were just running SAP, an ERP thing or something and now they’re running their own big data analytics and trying to be data driven. And there’s a whole bunch of SaaS companies starting to pop up in that space. So that’s one that is fairly nascent.
Alex: Why now the CFOs, and why the people who are running the businesses?
Adrian: I think the CEOs flipped in the last year or so at most companies.
Alex: They flipped?
Adrian: Yeah. The CEOs are saying, “Have you done cloud yet?” It’s become a “me too” at the executive level. If you’re not doing it, then you’re behind the curve, and you may miss out. So if you’re an enterprise CEO, you don’t even know what it is necessarily, but you know that it’s a thing that your people should be dealing with and that’s bleeding into the CFO land.
Alex: Is this interest also originating from just the changes that they need to adapt to as more of their divisions and groups of employees start using cloud services?
Adrian: I think so. The sort of slide deck I’ve been pitching is really about speed of development. So most companies are at some level trying to innovate on their products and get them out faster. And if you say: “Would you like your product to take longer to get to market?” they answer: “No, no, no. We want it to get to market quicker, not slower.” So hardly anybody wants their stuff to get to market slower. But it always takes too long.
Adrian: So anything you can do that speeds up time to market between having an idea, or seeing a market need, or a customer pain point, and getting it to market and discovering whether you’ve got it right and iterating, that’s a universal. So the time it takes to do stuff is a universal need. It doesn’t matter what segment you’re in. So I’m kind of focused on that thing and a lot of the speed there comes from taking away friction and getting stuff done. And a lot of these SaaS companies take friction away because they just start using you. It’s right there. You don’t have to have a huge procurement cycle and you can try three or four of them at once and see which one works and then drop the ones you didn’t like or whatever. There’s a lot of things you can do for getting much lower overheads for getting into something. So anything that cuts that time. So it’s more about time than money. It might cost you more in the long term to scale out a SaaS app rather than run it yourself. But the time you saved is worth more than the money that you might save.
Alex: That begins to affect the whole organization.
Adrian: Yeah. And what people are seeing is that their competitors are starting to do it. And so if you read the Phoenix Project, it’s like the company is Parts Unlimited. Have you read the book? You could do a whole section just on the Phoenix Project, a whole blog post or something.
Well, Gene Kim’s here [at Monitorama]. He’s one of the primary authors of the Phoenix Project. Just interview him about that book and what it’s trying to tell and read the book. Because the book is about a company that’s about to go out of business because it didn’t think IT was important, and it was getting in the way. All it wanted to do was make car parts and sell them to people in a car parts store. That’s what it’s about. It’s the story of this company and it’s on the verge of going out of business, and they eventually figure out that they’re being put out of business by competitors who have better IT. They have better software. They have better programs, and their market’s integrated, and they have customization of things in stores. All of the stuff that’s software driven is being done by the competitors and they’re being driven out of the market. Their response to that is to start this project or the Phoenix Project that’s going to fix this. But the project is dying because they can’t get the IT done. The story is about basically how they figure out how to do effective devops-like stuff for getting it together to be able to be competitive.
Alex: There are all different kinds of issues that begin to surface when you start adopting SaaS because you start to look at things much more differently.
Adrian: And a lot of enterprises are trying out the public shared multi-tenant thing. And once they decide they like it, they buy an enterprise license and they want an in-house one so they can control it. And they want to just pay once and not have to think about these things. So you tend to find the small-to-medium companies using the pay-by-the-month, pay-for-use thing. But once the CIO comes in, he wants a predictable, “I want to lay out some money and know I’ve got this thing for three years or a year,” or whatever it is. So they tend to buy extended licenses and want it installed personally. So they get their own build. It may be in their own non-tenant AWS account or it may be in some VMs running a vapor version of whatever it is that runs as a bunch of VMs and you stick it in a VMware environment somewhere on premise. So there’s quite a bit of that going on, driven by customers that aren’t ready to put stuff outside or they’re too big to need it.
Alex: Are there corollary shifts that take place with this SaaS adoption that are becoming evident now? There are a lot of behavioral changes that occur with shifts like this.
Adrian: The corporate change that happens is known as “Zero Trust” networking. It’s where you start treating everything you have in your enterprise as being effectively a SaaS app, whether it’s in-house or not. So you authenticate your users and their laptops and you authenticate against every service individually, internally. And everything is behind a firewall individually. So the analogy that’s been used is you go from being a hard crust with a crunchy center, just like a warm M&M or something, to a candy bar with nuts in it. So it’s soft on the outside but there are hard nuts in it. So every individual application is hardened and the environment around them is soft.
Once you get inside a firewall, everything is accessible to you and we trust you. So once you penetrate the firewall, if you’re an attacker you’ve got access to everything. Whereas if you look at a typical cloud infrastructure, every individual machine is so locked down that if you break into a single machine you actually have gotten nowhere. The machine will probably disappear in a few hours anyway and the fact that you’ve compromised it isn’t very helpful to you. So it’s actually enormously hard work. A cloud-native architecture is inherently harder to compromise and the data is normally encrypted at REST, and there’s a whole bunch of key management going on. It’s very hard to figure out what’s happening. And you end up breaking into one microservice — maybe. And it is not very interesting. And you can’t get into others.
Alex: Now there are corollary shifts that occur. I went to the ALM Forum up in Seattle—
Adrian: What’s ALM?
Alex: Application Lifecycle Management. And there was someone talking about the shifts that they had to go through. And he wasn’t talking so much about SaaS. He was mostly talking about how they had to adapt.
Adrian: There’s a team I met that built something on an AWS account in a few weeks. The rest of the company didn’t even really know what they were doing until they finished doing it. And then they (the company management) said: “Oh this isn’t right.” And then they said, “How much did you spend on this? You did this in this amount of time and the cost of running it each month is only that much?” And then internally people said, “Let’s not shut that down. That’s a tiny fraction of what we would normally spend trying to get that done.” And these guys are now kind of the trend. It’s becoming a little more viral internally that you can do these things. But it was done on a credit card on an AWS account. So that’s an example of how to get something done.
So you get these pathfinder projects basically whether they’re sanctioned or not or what level they’re sanctioned at. But this is like the generic pattern. You get the people who think they can do it and you get out of their way. They do a pathfinder project and they learn a lot and hopefully they have a good result. And then you kind of broaden slowly from that and you bring a few more people into the pathfinder project and it starts to become a thing. And that’s kind of how Netflix did it. Netflix had two or three pathfinder projects and I led one of those projects. It was the platform for personalization that I was leading. And then they said, “Okay that looks like it works,” and other teams adopted it and it became the overall architecture for it.
Part two of the interview with Adrian will be posted next week as part of our series: The New Stack Makers.