In the latest episode of To Be Continuous, Edith and Paul examine how monetization approaches vary depending on the developer market you’re in.

They discuss how developer markets can be divided into four spaces and examine the companies operating in each space. They also consider the pitfalls of startups that push their product in too many directions and consider how to go about evaluating the threat of Amazon cloning your service. This is episode #34 in the To Be Continuous podcast series all about continuous delivery and software development.

This episode of To Be Continuous, brought to you by Heavybit. To learn more about Heavybit, visit While you’re there, check out their library, home to great educational talks from other developer company founders and industry leaders.


Paul Biggar: So, you want to talk about developer market. What is the developer market?

Edith Harbaugh: So I think the very word developer market is a misnomer.

Paul: Okay.

Edith: I think if you look at an engineering organization, there are many different kinds of developers who have different roles or responsibilities who care about many different things.

Paul: So you’re saying that there is no one developer market.

Edith: I think there’s not only one developer market. I think there’s many different dimensions of a developer’s market. I think it’s on a flatland but even an n-dimensional problem.

Paul: So when we say a developer market, do we mean the dev tools market or do we mean broader than that?

Edith: I think they’re kind of lumped together. I think when people think of developer tools they initially think of, something like a GitHub or Atlassian.

Paul: Okay.

Edith: But however, there’s also database developers.

Paul: Right, there Cassandra as a service is a developer tool.

Edith: Yeah, the wants and desires of a database engineer are completely different than a JavaScript developer.

Paul: So, Scott Raney, when he was on the show a long time ago, divided the developer market into four different spaces, right? Do you remember this? There’s the as a service.

Edith: You’re stressing my memory.

Paul: I know, mine too. There were four. What the fuck were they? So, there’s like editors and like language tools. There’s SaaS tools like Stripe that are sort of developer first but are providing a non-code service. And then there’s SaaS tools like yours and mine that are sort of developer productivity. I’ve forgotten the fourth.

Edith: Shaun made a big impression on us.

Paul: Scott. Sorry, Scott.

Edith: Scott, yeah. This kind of came up recently cause RethinkDB just did a really good post-mortem. I’ll sum it up by. Well they summed it up by, they didn’t get, they didn’t monetize and they had to shut down.

Paul: Okay, yes. Well, if you don’t make money you have to shut down. I think that’s a usual, a very standard approach to business.

Edith: I don’t know, some people continue to lose money for a long, long time.

Paul: Well, I mean, if you’re Facebook, you can lose money for as long as you like because you have, you have numbers that go with it.

Edith: Yeah or I suppose like Medium who very recently, though they had, I think, I don’t know their numbers but I think they had pretty good numbers, just realized that ad support wasn’t going to cut it.

Paul: Right, I think they also have a lot of money in the bank.

Edith: Medium.

Paul: They’re cutting long, long before they hit the runway. You know, clearly it’s thinking several years in advance and projecting that far and realizing this path is not the right one.

Edith: Yeah, ad supported businesses are really hard.

Paul: Yeah,

There are almost no developer tools in the ads to developers space. 

I mean that could be a viable monetization unit, ads to a specific niche audience but no one sells ads to developers. So, as I understand it, I hope this isn’t private information Joel, apologies if it is, they make way more money off careers than they do off ads.

Edith: That makes sense, I mean at TripIt, so we had ads but it was basically breakeven, even with 10 million users. How we really monetized was through our premium products, through TripIt Pro and the product I started, TripIt for teams, for business.

Paul: Yup, I’ve been thinking of sponsored builds and that sort of thing, at Circle, paying to advertise to developers and we didn’t go too deep into the numbers, but it wasn’t really a thing that you could monetize too well.

Edith: It really starts to decay very quickly. When I was at Concur, and I think this was long ago that I can talk about it, they tried this experiment where they would sell ad spaces within an expense-reporting tool and this hugely pissed off their customers. And also, by the way, the ad revenue was pretty minuscule. When you start to decompose it, like say TripIt has 10 million users and now I’m just going to make up numbers because I don’t remember the real ones.

Say two million of those are active in a month and you’re getting 10 impressions per user. That sounds great. That’s 20 million impressions, except for if you’re getting 0.1 cent per thousand impressions, obviously this is not much money at all. And that was with a huge base of 10 million users.

Paul: Yeah. So you gotta make a bunch of money, ads is a shitty way. What was Rethinks?

Edith: Their monetization, they said they never really succeeded in it, that they couldn’t really convince people to pay for anything.

Paul: And so I find that this is a common thing in the space that Rethink is in. So they’ve got an open-source product and they find it hard to monetize licenses and so they created a hosted product and I think in Rethink’s case they were right up against Meteor.

So Meteor created not a database but a realtime service and Rethink was trying to catch up with that and Firebase had that too. So Meteor has all the traction behind it. Firebase has all the customers behind it.

Edith: And not to mention the deep pockets of Google.

Paul: Deep pockets of Google, right. So certainly now and when was that acquisition, last year?

Edith: It was a while.

Paul: 2014 maybe.

Edith: To Google they could be bleeding money but the scale of bleeding money at Google’s scale is very different than you or mine.

Paul: So Rethink has to deal with those two big competitors and it seems like they were kind of going in all three directions at once. They’re building this essentially a better MongoDB and they’re building the realtime enabler portion of it and the cloud. And they, I guess, never hit monetization in any of them.

Edith: Yeah, I mean that’s what they put on their own post-mortem which was that they were just, they were just spread too thin.

Paul: Yep, those are I think three really interesting ways to look at how you sell to developers and developer market. So there’s this open-source plus plus plus.

Edith: Which I’d say like GitLab is trying to do.

Paul: GitLab is trying to do it. I guess mySQL was the original one of these right, cause I feel Red Hat was a different sort of model, but yeah, Mongo, MySQL, both database companies.

Edith: JBoss?

Paul: That was an application server on top of a free something.

Edith: Yeah, this was back in like 2002, 2003. They were an applications server and Citrix acquired them as I recall.

Paul: Gotcha. Were they profitable, made lots of money?

Edith: I don’t know. To be honest, they either could have released their numbers but it was long enough ago that I’ve forgotten or they could have never released them.

Paul: So MySQL was licensing plus around the open-source product. Was there an enterprise version of MySQL?

Edith: I don’t remember.

Paul: What’s Mongo doing?

Edith: Services.

Paul: Services, okay. That’s a shitty business.

Edith: Or, well no, so Mongo actually had this huge thing where they were trying to key some features out and keep them at the enterprise level and their base rebelled.

Paul: That’s going to be difficult.

Edith: Their base just rebelled and it was actually a lot of very ugly hacker news posts. Well that’s always the dichotomy if you’re open-source.

Paul: Well James Lindenbaum has a very good view on this and he talked about Heroku in the day. What goes into the enterprise product and what goes into the developer product and what he said is, “Anything that developers “feel is developer-y, has to be available “in the developer product.”

And in their case that was a freemium product. So you want to do a git push, you want to support HTTPS, those are developer-y things. If you want to have SLA’s, access control lists, you want to have data retention policies or data destruction policies, all that is enterprise-y and you’re happy to charge for that.

Edith: Well, I think that’s a very good way to look at it because it differentiates what an end-user cares about, and I’m saying the developer is the end-user, versus what the company cares about.

Paul: Right.

Edith: So an individual developer, to be really blunt, doesn’t care at all about data retention. And I don’t mean that in a harsh way. It’s just like, they are individual developers, like job to be done.

Paul: I don’t think it’s just a job to be done but I think developers like to think about technologies and not necessarily about policy.

Edith: Yep.

Paul: So if it’s policy it goes in the other one but I think

Developers buy tools like they’re building their side projects. 

Edith: Yeah, they’re all about individual productivity for it.

Paul: Right. So actually, so John Sheehan from Runscope has this excellent tweet where he describes what the developer market is and he says, “It’s a consumer go to market “with enterprise dollars behind it.”

Edith: Totally agree and I think where you get lost is where you’re building something and I say lost in that eventually gravity hits, and you could have something that many people love but if you’re not monetizing it.

Paul: Right, so I guess the two ways to fail then at being a developer company is to build something which doesn’t or can’t have consumer like adoption or that can’t have enterprise dollars behind it.

Edith: Yeah and those are a little bit, to go all greek mythology, like the Scylla and Charybdis.

Paul: I don’t know that myth.

Edith: You kind of have to thread this needle of you need to get enough developers to love you that the enterprise will pay money for it.

Paul: Right, okay.

Edith: Because if the enterprise loves it but developers hate you.

Paul: Yep. So, developers loved RethinkDB but they didn’t use it, they used Mongo, and people hated Mongo. I guess anyone who really looked at it hated Mongo, it was like, it’s this marketing machine. And the product didn’t work that well and it had all this random crap in it that, I don’t know if you saw people dissecting? It logged randomly, so there was like a call to random in the logging.

Edith: Non-deterministic?

Paul: Non-deterministic, it was like, if rand is less than 0.1. So the idea was that it would log 10% of the time.

Edith: Well, I know why they did that to be honest.

Paul: Well, yeah, there’s a certain amount of sanity to it.

Edith: Well, it’s funny because logging features like that are extremely expensive from the supplier side.

Paul: Yeah. I keep hearing you talk about how analytics are really expensive and the only people who make money on analytics are AWS.

Edith: Well, we haven’t actually talked about this on the air.

Paul: Oh, okay.

Edith: I mean like, I think,

Analytics are extremely expensive. You have to log every single event that’s happening and then store and process it somewhere. 

Paul: Yeah.

Edith: And then the other flip-side of it is that Google Analytics has completely commoditized this. Like, so the price of Google Analytics is free.

Paul: Right, logging in a database is important. I feel you should log, I feel a database should log when something goes wrong but deterministically. Well like, getting a non-deterministic log just sounds like the most irritating thing in the world when you’re trying to debug something.

Edith: I can see why they made that decision though because it’s really expensive, cause you’re just generating so much content. I’m not defending them, I’m just saying I’m sure the logic of it. So my background is content management and if left unchecked it could generate a lot of content.

Paul: So, I would suggest that rethink did not fail because they logged every message but they were generally regarded as, you know, what Mongo should have been.

Edith: Yeah so, a side-story about why I feel this way is, so I used to work at a big content management system and we had an analytics product. The analytics logged such that, it would log basically a day worth of content and it would take roughly 26 hours to process this data. So it was like one of those laws of physics, they were constantly falling behind and their customers were getting more and more pissed off at them.

Paul: It’s getting worse.

Edith: It’s getting worse.

Paul: By two hours every day.

Edith: Yeah, by two hours every day, so at some point you’re like literally months and then people were just basically declaring logging bankruptcy and wipe everything out. So, it’s a tough problem about what to retain and what not to retain.

Paul: So, in terms of business model, it sounds to me that that and I don’t remember everything that Slava said in the rethink blog post but they clearly didn’t hit the licensing around an open-source tool.

Edith: Which is hard.

Paul: Which is hard. Has anyone done a really good job there?

Edith: The one example people hold up is Red Hat.

Paul: But they sell consulting around the tool, yeah?

Edith: Yeah, besides that it’s just really difficult. And so we go back to MongoDB, they tried to retain some stuff in their enterprise layer and the people in the open-source community who’d worked on that part were up in arms.

Paul: Oh right, yeah, if you got open-source people to write it then obviously.

Edith: And they were like, well, not even that they’d written that part. Basically the thing was, “Well, we’ve contributed so much “to your free side.”

Paul: Oh okay.

Edith: You know and then gets, you know, we’ve helped you get all this free stuff going, we kind of looked at that as barter was the contributors feeling

Paul: Wow, so they’re not selling additional services on top of the free product? So, the next step you go to is hosted and I know Mongo didn’t go in this direction so there’s MongoHQ which is Now Compose and there was MongoLab, which I think also has a new name, that actually provided hosting and, I think, does Amazon have a Mongo hosting now or someone has a Mongo hosting.

Edith: Amazon’s really all about Dynamo these days.

Paul: Ah right, that’s their NoSQL thing?

Edith: Yeah.

Paul: So rethink could have gone hosting and didn’t.

Edith: What they said, and this is all a kind of second-hand reading the same post, is that by the time they tried to go there, they were just spread too thin.

Paul: Right.

Edith: They tried to do that but you can only do so much. I think one of the most important things that a start up can do is say no.

Paul: Right, so if you push yourself in three directions, you’re kind of fucked. And so they were trying to do the real-time stuff as well as the, yep.

Edith: So they pushed in three directions and they end up executing maybe 40% of all three. Instead of just picking one.

Paul: Course, that means you have to pick one.

Edith: And we’ve talked about this before, the lack of a decision is a decision.

They said all three of these look like viable paths, and all three of them could have been market winners, so they went down all three paths and the net result was by not making a decision about one they could not execute well on anything.

Paul: Right, they needed to make the decision far earlier.

Edith: Yeah, but that’s tough cause you probably within the company they had people strongly advocating for each of these.

Paul: So, I actually interviewed at Rethink.

Edith: Paul, I love how every story collapses down to that.

Paul: Yeah! I’ve interviewed at a lot of people. So yeah, just when we were starting Circle I interviewed at RethinkDB and back then, so this is before they had their Mongo replacement, the premise of Rethink DB then was a database that ran on SSD’s. And every other database ran, back then, on spinning discs and there was no optimization for how SSDs worked.

And I think they got out of that but when I went in to talk to them and I eventually got an offer was roughly around the time where I needed to make the decision of was I going to start CircleCI and I think I was still in the, “Okay, I’m going to work “somewhere for a little bit to see if Circle gets legs.” and I didn’t want to commute to Mountain View.

Edith: Such a killer commute.

Paul: Yeah, it really was.

Edith: And so you went to Looker instead?

Paul: Lookout, yeah.

Edith: Yeah, I made many similar career decisions based on commute. It’s easier now but I remember it used to be harder to find a job in the city.

Paul: Yeah and so I think people started moving to the city around 2009, 2010?

Edith: Yeah.

Paul: People really hadn’t started moving up here by, certainly when I was doing YC, 2010.

Edith: Yeah my litmus test as I remember, at the time I didn’t own a car and I actually had to borrow my friend’s car to drive down for an interview in Mountain View.

Paul: Right, you lived in the city?

Edith: I lived in the city and I drove down there and I interviewed and they actually made me an offer, it was just a hassle just for the interview.

Paul: Yep.

Edith: And same, I interviewed in South Salito.

Paul: Oh jesus, yep.

Edith: South Salito I could kind of justify because it’s beautiful.

Paul: Yeah, you could have a nice brunch overlooking the bay.

Edith: Yeah and I could ride my bike.

Paul: Yeah.

Edith: Except for when it’s raining and hailing.

Paul: And when you want to go back uphill.

Edith: Yeah, but I think people underestimate how big of an issue commute is.

Paul: Yep, I know we’re getting tangential but I drove down to Palo Alto today and I was driving against traffic both ways so it was totally fine but.

Edith: You drive?

Paul: I learned to drive to fundraise.

Edith: When?

Paul: When I was raising the Series A.

Edith: That’s right.

Paul: So I learned to drive to be able to drive down to Sand Hill Road.

Edith: Is it confusing?

Paul: Being on the wrong side? No, I didn’t learn to drive when I was in Ireland so I’ve only really driven in America.

Edith: That’s amazing, so did you go to driving school or what did you do?

Paul: Yeah, I got someone and I drove dozens of hours and then I did the test. It was like four weeks from when I started I did the test and I passed.

Edith: Did you make a conscious decision, like I need to go fundraise?

Paul: Exactly and I needed to be able to drive down to Palo Alto to fundraise or else it’s too difficult because when I did the seed, I was taking the Caltrain and then I was cycling to Sand Hill Road. I remember walking into VC’s offices dripping with sweat because Sand Hill Road is a hill, it’s a real fucking hill.

Edith: There’s no sand but there’s definitely the hill.

Paul: This is where you see the VC’s mountain biking on their $10,000 bikes and then all their spandex around there because it’s really difficult to cycle around there.

Edith: So that’s a great story so you wouldn’t pause in the parking lot and kind of towel off?

Paul: Sometimes in the bathroom, yeah, it was a disaster. I wasn’t in my spandex, I was wearing my shirts and my jeans and converse. I was commuting, there was just a hill in the way.

Edith: So the cool thing about when we, when we fundraised, I ran my bike everywhere and what I really liked, even between the seed and the A, is how many more people were up in the city.

Paul: Oh yeah, that makes it a lot easier.

Edith: So South Park is now like a mini Sand Hill Road, down to like, they’re all cheek and jowl.

Paul: Cheek and jowl.

Edith: Redpoint is right next to Shasta, it’s right next to Sear, you know that.

Paul: Oh right. They’re on the same, like 101 South Park or whatever?

Edith: Yeah but it really saved so much time.

Paul: So I learned to drive to fundraise and then I hadn’t, I learned San Francisco driving, which is not difficult but it’s not freeway driving and so I wasn’t comfortable on the freeway and so whenever I arrived down in their offices, again, drenched with sweat, because of the fear of dying on the freeway. So, in the end, I ended up Uber-ing up and down.

Edith: I thought you were going to say like you took El Camino Real all the way down there.

Paul: Oh, that would have been a good idea but what I discovered is that there’s an infosec leak with Uber drivers. I think it was a Lyft, I Lyfted a lot at the time. A friend of mine was coming to my house and he said, “Were you down fundraising “in Palo Alto today?” and I was like, “Yeah, how did you know?” and his Lyft driver had told him.

Edith: What?!

Paul: That there is an Irish guy that he dropped off here earlier. Yeah it was crazy, a $80 fare down to Palo Alto and then I waited and then I drove him back.

Edith: Well, it could have been the intercom dudes, how did he know it was you?

Paul: Because it picked up from that building.

Edith: Yeah but I mean that it was an Irish.

Paul: I don’t think I lived in the same place as the intercom dudes.

Edith: You’re not the only Irish guy in the city.

Paul: But I was the only Irish guy who lived, I wasn’t actually the only Irish guy who lived in that building but, fundraising, it was a reasonable guess.

Edith: Oh yeah, I remember we were going down to pitch and we were pitching different firms. I realized that I was going to pitch the two firms, like they’re all right next to each other on Sand Hill Road, and they were literally in the same building.

Paul: I’m surprised that they don’t have something set up in their wireless routers to know when an entrepreneur connects to their wireless thing, that they’ve maybe connected to it before. Wouldn’t that be smart?

Edith: No but they all have their own individual Wi-Fi.

Paul: Yeah but if you’re in the same building, you can see all the next door Wi-Fi’s.

Edith: Oh, but I don’t think you can see that somebody else has connected.

Paul: No but maybe you connected to what you’ve, your computer’s been connected to before?

Edith: Can they sniff that out? I don’t so.

Paul: So, if you’re in like Redpoint and Trinity are in the same place and Sequoia’s in there as well. So, you go to Trinity and you’ve previously been to Sequoia and you connect to Sequoia’s Wi-Fi because your phone automatically connects to it.

Edith: Ohhh, well the nice thing about SoftTech was our C investor and then I kind of ran on South Park so I always had kind of plausible deniability if somebody saw me around there, I was just like, “Oh, I’m just visiting.”

Paul: Yeah, okay, nice.

Edith: I’d just hope somebody wouldn’t see which offices I was actually going to.

Paul: Right.

Edith: My deep fear was that I wouldn’t be paying attention and I would open the wrong door.

Paul: Oh and just walk into the wrong office?

Edith: Because they’re all like. So I was like what if I walk into the wrong office, because a couple of times I would walk out and I would be unlocking the bike and someone maybe would walk by.

Paul: Yeah.

Edith: Because I would ride my bike.

Paul: I feel like people know when you’re fundraising anyway, like oh you’re having a meeting in The Rosewood? I think you’re fundraising.

Edith: Well that’s a pretty. I mean, like, maybe you’re just I don’t know, hanging out?So you don’t believe in innocent meetings with VC’s?

Paul: No, I mean, you need to have those relationship building meetings. In fact, I’m having innocent meetings with VC’s right now. So fundraising for my new thing, except I’m not fundraising. I’m just starting it off, getting feedback, all that sort of thing but like, inevitably it’s the start of a relationship that is going to turn into a fundraising cycle at some point.

Edith: Yeah, that was the same strategy I followed. To me, an investor is a big deal. This is somebody who is basically as important as a founder to your company.

Paul: Yeah, except at seed when they’re disposable. I’m joking.

Edith: No, you’re not.

Paul: No, I mostly am, but you see a lot of people with the party rounds where they’re like, where it doesn’t matter, you know, they’ve got 40 or 80 people in there thing and I doubt they would know an investor if they ran into them in the street.

Edith: Oh yeah, like I heard horror stories. I heard this actually at a conference we went to together, I sat next to the guy and there was this trend, about five years ago.

Paul: For the big party rounds?

Edith: Yeah and people were just dumping money and so he met one of the partners at some event and he’s like, “Hey, your firm invested in me.” and the guy just ran away. Because they weren’t going to do the follow on and he just didn’t want to deal with this guy he’s like, “Remember me? I’m here.”

Paul: Give me more money, please.

Edith: Yeah, he said like literally the guy bolted. So, to get back on topic, I think you have to eventually monetize.

Paul: Yeah, I mean that’s how you make a business.

Edith: Yeah, I think VC’s will give you money at some point for just your promise but eventually that check comes due in terms of what have you done for me lately.


I think the most obvious way to monetize is to host it and I’m surprised at people who delay this. 

So like Mongo could have been the Mongo host. RethinkDB could have been the RethinkDB host. Meteor is obviously the Meteor host, Meteor actually is the Meteor host.

Edith: I think it makes sense for some things and not for others. Then there’s extreme outliers, so our advisor Sean Byrnes’ company, like I just totally said advertising never worked. Advertising is what worked for them.

Paul: Right, this is Flurry?

Edith: Flurry, because they had sufficient domination in the mobile market that the economies of scale worked out for them to do mobile advertising.

Paul: Gotcha.

Edith: So I said, TripIt was 10 million, I don’t know their exact numbers because they weren’t public but I think they literally had some huge numbers.

Paul: Ten times that?

Edith: Well, no because they had an SDK the app developers installed so think how many average apps there are on somebody’s phone and let’s just back that out, say they had.

Paul: So 0.4, the average number of apps that I think most people have is less than one. I’m being facetious. People that actually install apps.

Edith: Yeah, so just say they had 50% of the overall mobile analytics market and multiply that worldwide and suddenly that’s a lot.

Paul: Yeah, that’s a lot of users.

Edith: That’s a lot of users, that’s a scale when you can actually be successful in advertising.

Paul: Who did they exit too?

Edith: Yahoo! bought them.

Paul: Yahoo! bought them? I’m sure they made good use of them.

Edith: You know sometimes I’m not sure if you’re sarcastic.

Paul: So this to their Verizon now?

Edith: Yeah, which is funny because Sean actually started off at Verizon but he’s not been there for a while. Sean is great, his thing is that everybody eventually tries advertising.

Paul: Right, I guess Rethink didn’t try advertising. Maybe they should’ve tried it. I’m joking.

Edith: I think advertising works if you have immense, like, Google makes a ton of money.

Paul: So hosting, I think, is this strong model. Firebase is pure hosting, right? There is no way to use Firebase without it, is there an open-source Firebase?

Edith: I don’t think so.

Paul: Okay, I wouldn’t be surprised if someone tried to make an open-source Firebase.

Edith: I wouldn’t be surprised if somebody is, right now, trying to make an open-source LaunchDarkly or CircleCI.

Paul: Right, yeah, there’s like 12 open-source CircleCI’s. So the problem you end up with an open-source product which is hosted is you spend a lot of time building the hosting and you aren’t open-sourcing the hosting and there’s a challenge for your customers. If they’re going to run it themselves because they don’t have the migration part of it. So Heroku Postgres was a hosted version of an open-source product and they ran stock Postgres but the tool that they actually built is their infrastructure tool around Heroku Postgres, which was not open-source.

And you kind of expect the same is true with Meteor and with all the other database hosts and that sort of thing. But Firebase went and just said, ” the service isn’t open-source at all, the whole service is the database and, you know, you can download the libraries because the libraries are open-source but the thing that makes Firebase Firebase is it’s hosted”.

Edith: That’s kind of very similar to LaunchDarkly to be honest. Our libraries, our SDK’s, are open-source. We love people who want to modify our SDK’s but the core of LaunchDarkly is all closed. Our advisor today was like, “Why don’t you have a “swift SDK?” and we’re like, “Why don’t you build one?”

Paul: Also I guess my point is that if the core of LaunchDarkly was open-source, what would that mean? So, you’ve got some data store which someone can run on their own machine but then when they need to scale it up, they hit an inflection point or a multiple of 10. That would be extremely difficult for you to build a product that allowed customers to do that migration themselves. In fact, it’s probably challenging enough for you to do the migrations yourselves and controlling the whole architecture.

Edith: Yeah, that’s something I’d said today was I’m not smart enough to build a successful open-source dev and monetize it, I’m just not.

Paul: I can’t think of anyone today who’s doing that successfully. So a big big unicorn at the moment is Docker, in the dev space.

Edith: They’re not monetizing.

Paul: I don’t think anyone knows how they’re going to monetize and I talked to a whole bunch of investors and the investors like, “It might be zero.” investors are, typically, incredibly shrewd about this sort of thing.

Edith: Well everybody thought they were going to be the next VMware so they’re definitely, I definitely think containerization will continue to accelerate. I don’t know if it’s monetizable? And now just Docker, and I know many people who work at Docker and I like them but they have this issue that they have such a high valuation. It’s very hard for them to get more money.

Paul: Where do they go?

Edith: So I think they’re a good example of something that can be extremely successful with developers but…

Paul: But not have anywhere to go?

Edith: Yeah.

Paul: It might be that the people who make money, I mean, hosting, so right, the people who are going to make money on Docker are AWS and Google, in particular Google.

Edith: You don’t think Microsoft?

Paul: Microsoft, but Google owns Cooper Netties and so you can kind of think of Google’s Google Container Engine, GKE, as the de facto Cooper Netties–

Edith: Google compute engine or container engine?

Paul: Container engine.

Edith: Okay.

Paul: But that might be where the money goes and that it’s spent on infrastructure.

Edith: Yeah, I don’t know. So I went to both AWS Reinvent and Microsoft Connect and for them it was just all about the core, core–

Paul: Services?

Edith: Yeah.

Paul: The computer you mean or just everything?

Edith: Azure is where Microsoft is bidding their future and that’s another factor that’s played into the dev tools market, is like, Microsoft is basically giving away now Visual Studio. Because they