Martin Kleppmann: To Silicon Valley and back again, with two exits along the way

Podcast transcription - 24th January 2018

Peter Cowley:                  So, welcome again to The Invested Investor. Today we have Martin Kleppmann, who I've known for probably over 10 years now, and was on his first journey, first successful journey. He then had an even more successful journey in California, which I wasn't part of. That's part of the story, perhaps. So first of all, can we just ... Martin, can you just describe your background and why you became entrepreneurial to start with?

Martin Kleppmann:        Hi, Peter. Yeah. Thanks for having me on here. I studied computer science as an undergraduate, and then after graduating was not really sure what I wanted to do with my life. I kind of looked around and thought, "Well, I could get a programming job, but it seems kind of boring just working for an existing company." For a while I went off to study music actually, thought, "I don't want to just be pure computer science geek," but then that didn't really work out that well either.

                                             I ended up back in Cambridge and thought, "Well, okay. I'll just try and start a company." Had no idea what the company was going to do at that point. I just had this feeling that, "Well, I'll give it a try because it seems like an interesting thing to try." This is actually where I met you early on as well because I knew that I was clueless, and so I went to our department's alumni mentoring scheme where people with more career experience offer to help out younger ones like myself. That's actually where I met you, Peter, and you remember that as well.

Peter Cowley:                  I do remember meeting in a, was it a pizza restaurant in Cambridge to start with?

Martin Kleppmann:        Yeah, something like that.

Peter Cowley:                  Yes. Then we worked together. This really was an angel investment. In fact, I say that I learnt what an angel was almost from you and the journey. Let's just talk through what happened. I invested a little bit of money. You put some money into it. Talk through that journey.

Martin Kleppmann:        I was bouncing around with all sorts of crazy ideas of stuff I wanted to do, mostly driven by technologies I found interesting really. It was super helpful at that time having you there just continually reminding me that actually what I was trying to do was actually start a business, not do a PhD or something that was pure technology. Actually, I had to think of stuff that I might actually be able to sell to customers or that customers might actually be interested in, which was, I think, a pretty novel concept to me at the time. I think I'd sort of theoretically internalized it, but not really got in the habit of thinking that way. So having you continuously prompt me as to like, "Okay, who's going to actually pay for this?" was actually tremendously helpful there.

Peter Cowley:                  You had this idea. I remember it right at the beginning, which I insisted you went out to do some market research on, which was a rollback for an operating system in a school environment I think. You went out. I seem to remember you went out to several hundred I think and got quite a lot of good feedback from that.

Martin Kleppmann:        Yeah. I had this idea that, "Oh, it's terrible if computers crash, so wouldn't it be great if you had a kind of reset button that if your computer crashes or something goes horribly wrong, you can just go back to one minute previously and get the entire computer back into a state as it was then." I still think it's kind of an interesting idea, although that would have been technically very hard to actually pull off. It seemed like a nice idea on paper, but I think we just weren't the right people to do it.

Peter Cowley:                  But also the market research, I think, did some pushback in terms of pricing and whether it they would want it. Many people said they'd like it, but was there a market for it?

Martin Kleppmann:        Yes, yes. Whether anyone would actually pay for this was super unclear.

Peter Cowley:                  So you then formed a company, and I think the first thing you did was to go out and contract, didn't you, and worked for another company?

Martin Kleppmann:        Well, I realized at some point that having no income was not really an ideal state of circumstances, so I realized that yes, just doing some web development, consulting for random companies was an easy way to generate some income. I was effectively freelancing, but doing that through this company that we had started. In the process of doing that, just tried to keep some money on the side, which gradually built up in the hope of using that to fund some actual product development later on. Which did work out.

                                             Through the process of doing this consulting, I then realized that there was actually a need for a kind of product in the realm of testing websites, because I was developing all these websites and they had to work in various different web browsers. This was a time when various versions of Internet Explorer were still quite big, and so it was actually quite hard work making JavaScript work across these different browsers. I thought, "Well, it would be good to be able to automatically test websites around different browsers." I spoke to a couple of web development companies, consultancies, that did this kind of thing, and I got the impression that they would also perhaps use such a product, so ended up building that then. It was a boosted service called Go Test It where you could record the steps of users interacting with a website like clicking on some buttons, typing some text in a box, pressing another button, pressing some links, and then checking that the right behavior happens across different web browsers. We made the thing work as a product.

                                             Then I found actually getting people to incorporate it into their workflow was still really hard. It's the kind of thing where loads of companies would say, "Yeah, that sounds like a really great idea. We'd love to use that," but then actually sign up for it and never actually use it because just actually getting into the habit of using a new tool like that, which does have a bit of a learning curve and a bit of effort in setting up the tests, it was still very difficult to get anyone to actually habitually use it.

Peter Cowley:                  I remember, great, you had the red and the green light, and if it ran correctly you got a green, and you used to sort of woo! If it didn't, you got a red, and it had to work out what had gone wrong in which browser at that point.

Martin Kleppmann:        Yeah. Well, the red and green lights were for our own internal systems, not the users' tests. But still, it did have the problem that often tests would fail spuriously for not particularly good reasons. Maybe you just changed some unrelated part of the website and then the test would break and you'd have to spend time fixing it again. If people experienced that a few times, then they get put off very quickly.

Peter Cowley:                  How big was the team then, at this point?

Martin Kleppmann:        It was primarily me driving it, but I hired contractors to do various parts of development. So probably five or six people who contributed to it part-time in total.

Peter Cowley:                  And this was based in office space in Cambridge, wasn't it at this point?

Martin Kleppmann:        Yeah, it was.

Peter Cowley:                  Now, the acquiring company, can we say how they got involved?

Martin Kleppmann:        Yeah. It was a funny story actually. Redgate, a software company in Cambridge, had just moved into a new office. They had a new larger office that was way bigger than they needed for the current size of their team, and so several people who were doing start-ups in Cambridge somehow got wind of this and started talking to the founders of Redgate whether they could just borrow some of those spare desks that they have lying around. They said, "Oh yeah, sure. Why not?" It turned out that even Redgate was setting up a really nice canteen there, and the start-ups who borrowed desks there even managed to wrangle free food in the Redgate canteen, ostensibly on the basis that we would help bring some buzz and entrepreneurial excitement to the internals of Redgate, the company.

                                             I heard of this, that some people had this arrangement, and asked myself, and got in on the arrangement as well. Probably the third or so of the-

Peter Cowley:                  With the team as well?

Martin Kleppmann:        Yeah. I and a few other people who were working on Go Test It at the time then borrowed some desk space there. It was really a nice environment because it ended up being sort of like the best parts of an incubator, several like-minded people working together all the time. We would have lunch together all the time. I would get to know the people from the other companies there quite well. We ended up helping each other out actually, and they contracted on some of my stuff and I helped them out with some of their stuff. It was a really great community there.

Peter Cowley:                  And at some point Redgate decided it would be worth talking about buying the company, wasn't it?

Martin Kleppmann:        Yeah. I had got to the point with Go Test It, as I realized just actually making it work commercially was still going to be quite hard even though we had a product that generally worked. Making it work commercially, I was getting kind of sick of it by that point, to be honest. So I was thinking of just giving it up, and just going back to university, doing a PhD or something like that. I thought that would be a nice exit route.

                                             But speaking to the CEOs at Redgate at the time, they ... Well, their whole business is developer tools, and what we had built was a developer tool. We thought, actually, this could fit quite well within their product lineup. They have a whole lot of existing companies who already use Redgate tools for various tasks including database administration and debugging tools for .NET. We thought this would actually fit in quite well potentially. So we talked for a while and decided that it would be a good idea to sell Go Test It to Redgate.

Peter Cowley:                  And they acquired it on the basis of an amount upfront and then an earnout. During the earnout when you're working there, you'd got to know both the cofounders of your next business. So let's move forward now to the next thing. I remember having a curry with you in Cambridge on Castle Hill about this thing called Rapportive. I really couldn't get my head around what it was going to be. Let's start talking about Rapportive now, shall we?

Martin Kleppmann:        Yeah. Rapportive was a little browser add-on that would manipulate Gmail. The idea was that if you are someone like us, a start-up founder who does a lot of emailing with people they don't know very well, maybe like user signups from our website, then what we would do is look up the email address of whoever has just emailed you, and we would try to find that person on various social networks. Especially on LinkedIn, but also on Twitter and Facebook if the profiles are public. Rapportive would take that social media information and put it right inside Gmail, inside the web interface next to the email. You would get a little summary profile of the person you're talking to right there next to the email.

                                             This really came out of something that we'd wanted ourselves at our previous start-ups. Me while I was doing Go Test It and the others, they were working on various other companies at the time. What we wanted there was to establish rapport, to build a personal relationship with our user signups. We wanted to know is this person we're talking to, say, a technical person or is it a more managerial person, and thus reply to them appropriately given what kind of role they had. So we would spend a lot of time looking up people on LinkedIn and seeing if they have personal websites and so on. Rapportive was essentially an attempt at automating that.

Peter Cowley:                  Where did Y Combinator come in? Did they approach you, or you approach them?

Martin Kleppmann:        Well, we thought given that we're doing this social media stuff, it would be a good idea to be in San Francisco because that is, after all, where most of the social media companies are based and where a lot of that stuff happens. Especially the consumer internet type of businesses. We were based in Cambridge in the UK. Didn't really have any contacts in the valley at all, and so we thought actually if we join Y Combinator, that would be a good way of becoming established there.

                                             What we did was we just went through the regular Y Combinator application process. We put up a website like Y Combinator expects you to have some kind of demo online generally that you can use. So we just put up a website. It wasn't password protected or anything, but we didn't promote it either. Just explaining what Rapportive is and allowing you to download, install it, and use it in your web browser.

                                             Then a friend of ours decided it would be a good idea to send a link to this website to a blogger, who then tried out the thing and found it exciting and wrote a blog post about it. Then other blogs picked this up, and several others then, fairly well known blogs at the time, picked up this Rapportive thing that apparently had just launched. We didn't really have any intention of launching at all, but suddenly we were all over these blogs and people started downloading it and installing it. Within a day we had 10,000 users based on something we hadn't even planned to launch. At that point then we thought, "Well, okay, we're really on to something here. This is something that people find genuinely exciting."

                                             It was, unfortunately, right during the period when I was still working on the earnout for Go Test It at Redgate. I then went to the CEO and had the very awkward conversation saying that we've just started this thing accidentally and it's going really well and we'd kind of like to continue working on that. They were really understanding.

Peter Cowley:                  Yes, and one of them invested, I think, in the next business.

Martin Kleppmann:        Yes. Yes, they did, and it was really fantastic to have that support and understanding there. So even though Go Test It didn't end up really going anywhere, but somehow while this new thing Rapportive got started, wasn't really intending to conflict. I was intending to continue working on it until the end of our agreed period, but, well, you know, sometimes unexpected things happen and it's great when people are that understanding.

Peter Cowley:                  We'll explain what Y Combinator is in the show notes. We'll put that in the show notes. You've mentioned we several times. Can you just describe who the team was at this point?

Martin Kleppmann:        Yeah. There were three of us. There was me, Rahul, and Sam. All three of us knew each other from this shared office space where we'd all borrowed a desk from Redgate essentially, and we got to know each other quite well simply through having lunch together every day and sometimes sitting there till late at night hacking on stuff. Rahul and Sam had previously worked together on a different start-up that was-

Peter Cowley:                  Mojo, wasn't it?

Martin Kleppmann:        Yeah. Mojo.

Peter Cowley:                  Yeah, yeah.

Martin Kleppmann:        You can probably explain it better than me.

Peter Cowley:                  Yeah. I've just seen Stew actually, the co-founder of that.

Martin Kleppmann:        Right.

Peter Cowley:                  But let's move on with your story.

Martin Kleppmann:        We knew each other from that, and we'd been discussing this idea of Rapportive for quite some time. We described it amongst ourselves as we wanted this head-up display for the web, but you get this additional information that is relative, contextual, right next to what you're doing anyway. It's especially to do with when you're dealing with people you don't know.

                                             Rahul put together the first version of Rapportive, but we'd all been discussing it together. The three of us then formally founded the company together-

Peter Cowley:                  Which was an American company, wasn't it at that time?

Martin Kleppmann:        Yes. We actually then incorporated it through Y Combinator. That was a funny story in itself actually because after our rapid 10,000 user in a day launch, accidental launch, we then talked to Y Combinator and said, "Well, actually, given this traction we have, we might to want the seed round quite soon. If Y Combinator wants to get in at their usual terms, then they should do that quickly so that there's a bit of time between the Y Combinator investment and the seed round."

                                             And so, we did with them what I don't know if any other Y Combinator company has done before and after us, actually, which was to do the interview with them over Skype. Normally you always have to fly there for an in-person interview, but they were happy to just do it with us with Skype and invest over the phone, essentially. We then incorporated the American company with their help. We moved over for the summer for the three months of Y Combinator's program there. Then we had to move back to the UK to get our visas sorted out, but once we had our US work visas sorted out, we then moved back to San Francisco, and we were based there from then on.

Peter Cowley:                  And you raised how much in that seed round? Not the Y Seed, the seed round.

Martin Kleppmann:        The seed round was in the months following. Then we raised about a million dollars there.

Peter Cowley:                  From whom?

Martin Kleppmann:        Mostly angels. There were some small investments from VCs who were interested in essentially just keeping their foot in the door there.

Peter Cowley:                  What did you spend the money on, the $1 million?

Martin Kleppmann:        It was almost all salaries. There's an inconvenience with the American visa system that requires that for the type of work visa that we were going for, which is called H1B, it requires that all of the people who are employed under this work visa are paid market rate salaries, and so we had to pay ourselves market rate salaries for a software engineer in the Bay area, which at the time it was just based on some list that we were given by our immigration lawyers, was like $75k per year, which is way more than we would have wanted to pay ourselves, really, but that's simply what we had to do in order to meet our visa requirements.

Peter Cowley:                  And I think you lived together as a team anyway, so you were sharing combination costs and food.

Martin Kleppmann:        Yeah. We just shared a flat together. Even when we hired two more people, one of those extra people then continued living with us. There were four people in one little start-up flat.

Peter Cowley:                  So you're doing tech development, but presumably also you were doing market development at this point, i.e. increasing the user base?

Martin Kleppmann:        Yes. We were trying to grow the user base and at the same time just build the features that we wanted really. The features we ended up working on were basically things that we wanted ourselves, and so because we were our own customers, that actually worked reasonably well. Generally whenever we launched a feature, people were quite happy, though it was still a free product, and so any sort of revenue seemed reasonably far away. Then we did start working on premier features, which we wanted to sell under a fremium model.

Peter Cowley:                  Had you worked out pricing for that?

Martin Kleppmann:        We had some ideas for it, and had the conversations with people, but nobody had actually put any money down yet. We were expecting it to be a small software service subscription.

Peter Cowley:                  A dollar, $2, $3 a month, or more?

Martin Kleppmann:        Maybe like $10 per user per month or something like that. That would then be for people using it for business purposes. The premium features would be stuff like integration with your CRM systems where actually this kind of stuff is really valuable because often your customer interactions actually happen in your CRM if you're a business rather than in your email.

Peter Cowley:                  How many regular users, or downloads, or whatever metric used, had you got to before ... We get on to the exit soon, but where did you get to? How big was it?

Martin Kleppmann:        I don't remember the exact numbers. On the order of 200,000 active users, which was reasonable, but was not really enough to get the VCs excited to want to invest in series A. So we went around trying to raise a series A because of our high salaries. We were burning cash fairly quickly, so the seed round gave us about two years of runway-

Peter Cowley:                  Two years. Okay, yeah.

Martin Kleppmann:        ... with all of the salaries and then the server costs and other stuff. So we ended up actually only having fairly limited time, and when we started then talking to the VCs about series A, then they thought, "Well, like 200,000 users. It's okay, but it's not really that exciting." It was difficult. If we had had a million users, then it would've been a different story, but at the time it was already the case that a few hundred thousand was not really considered enough-

Peter Cowley:                  To actually raise $3 or $4 million or $5 million.

Martin Kleppmann:        Yeah.

Peter Cowley:                  Okay. Before we go on to the exit, who else was helping you at this point? Were the angels helping you, or did you have other advisors? Did you have any gray hair around you?

Martin Kleppmann:        We had quite a large set of angel investors for this one. I think it was like 20 or so, and with most of them we only had very little contact, but there were a few people amongst those who were super helpful, often with introductions. The Bay area is very introduction-based, so if you want to get to talk to anyone, you need warm introductions. That was tremendously helpful there.

Peter Cowley:                  What about government? It's unlike the UK. There isn't really a board as such, was there?

Martin Kleppmann:        No. In that regard we were very informal really. We didn't really have formal board meetings. In practice it was just the three of us founders running the show, and that was okay.

Peter Cowley:                  What about information flow to shareholders? Did you do that regularly, or ...?

Martin Kleppmann:        Nothing very formal. We would occasionally send out bulk emails to the investors if there was something to say, but we were not actually very systematic about it.

Peter Cowley:                  Was it essentially you went to an A round rather than just ... I mean, in the UK it's common that one would raise a second round from the existing investors angels. Would you have been able to do that with the existing investors?

Martin Kleppmann:        We probably would have been able to because we had quite good relationship with our investors.

                                             In the end though, when we were facing the options ... Well, raising more money from the investors would have been a bridge essentially, so we would have had to have some pretty clear idea of where that got us to at least some decent revenue base or series A. We were just about to launch our premium product, but hadn't actually launched it yet and as such we didn't really have any data as to whether it was going to work commercially. We could kind of sell some kind of great vision about this, and of course the valley is all about selling great visions without having any concrete evidence for it. That's how, surprisingly, many things work there, but it wasn't really ideal.

                                             When we then got talking to LinkedIn, who ended up acquiring us, that was partly on the basis that, well, if that acquisition discussion was not to work out, then we would be able to go back to our investors and raise a bridge round probably.

Peter Cowley:                  It was only LinkedIn you were talking to at this point?

Martin Kleppmann:        We did talk to the other usual suspects, but only with LinkedIn did we really get to term sheet and have a proper conversation.

Peter Cowley:                  Okay. Who did that negotiation?

Martin Kleppmann:        We had some incredibly helpful help there from some gray hair. There was a broker who had M&A experience, mergers and acquisitions experience, previously, both on the buy and on the sell side from several previous acquisitions. He helped us there.

Peter Cowley:                  And took a fee?

Martin Kleppmann:        He took a fee. He took a percentage of the overall deal, but it was perfectly fair. It was totally worth it. He helped us navigate that process, which it's very psychological at times.

Peter Cowley:                  Was the negotiation on price as well?

Martin Kleppmann:        Yes.

Peter Cowley:                  I mean, what other terms were there? I mean, earnout obviously as we'll come to in a minute, but ...

Martin Kleppmann:        Yes. The price was really the main factor, but other questions are what sort of ratio of cash versus stock? What other compensation is there for the employees as well? We wanted a good deal for employees who were only small shareholders, but who had contributed massively to making Rapportive actually successful. We wanted them to have good terms as well.

Peter Cowley:                  How big is the team at this point?

Martin Kleppmann:        We were five people.

Peter Cowley:                  I see, still only five.

Martin Kleppmann:        The three founders plus two.

Peter Cowley:                  Two others. Yeah.

Martin Kleppmann:        And, yes, there's the vesting schedule, which is the period over which the actual acquisition price gets paid because it's not usually the case that you just get all of the cash up front because generally there's some milestones it's tied to. Ideally those milestones are only time-based, so not based on launching some product features or whatever or some commercial milestones, because with those you never know whether you'll actually make them. In our case, it was just pure time-based milestones.

Peter Cowley:                  The headline figure of things in the public domain? How much was that?

Martin Kleppmann:        Not officially.

Peter Cowley:                  Oh, it's not-

Martin Kleppmann:        The leaked figures of around $10 to $15 million were about correct.

Peter Cowley:                  In dollars. Yes, okay.

Martin Kleppmann:        Yes.

Peter Cowley:                  Clearly not the earnout, but the earnout was two or three years, wasn't it, I think, in this situation?

Martin Kleppmann:        Yes.

Peter Cowley:                  Yeah, okay. You then joined LinkedIn. The team joined LinkedIn and formed Rapportive. What did you do? I know you had a very close relationship with a lady over here, and I went to your wedding even. You came back over here and worked remotely. But how's your journey been? And we want to talk about what you're doing now, of course, as well.

Martin Kleppmann:        Yeah, sure. The team stayed together at LinkedIn. This was really important to us. It was not just hire a team of techies and get them to work on random LinkedIn stuff. We actually continued working together as the Rapportive team working on our own product vision. That was really fantastic, and LinkedIn gave us a lot of support there in actually being able to build the next version of the product if we wanted.

                                             The product we then launched after a year and a half or so was called LinkedIn Intro, where we were trying to do the same thing on mobile. Essentially what we'd done previously on the desktop browser inside Gmail we wanted to do on mobile phones as well. Unfortunately, that product ended up failing because of a whole bunch of combination of reasons that were really outside of our control. But LinkedIn was very friendly and reasonable about that. Even though the product ended up getting shut down, we were all still really treated very well.

Peter Cowley:                  And Rapportive itself, is that still available as a plug-in for LinkedIn?

Martin Kleppmann:        I think the last time I checked, Rapportive had been subsumed into LinkedIn's sales solutions product business. So it's not under that name anymore, but something of that sort still exists I believe. It's not a paid for product, I believe.

Peter Cowley:                  And you left LinkedIn. We'll come to you. Rahul also left LinkedIn, and did Sam leave them as well?

Martin Kleppmann:        Yes.

Peter Cowley:                  Yeah.

Martin Kleppmann:        Yes. So Lee is still there I think, but four of the five Rapportive gang eventually ended up leaving LinkedIn.

Peter Cowley:                  And you went on to write a book, which you will talk about, and become an academic again.

Martin Kleppmann:        I stayed with LinkedIn for about two and half years after the acquisition, and during that time worked on our own product and then after that failed, worked on LinkedIn's internal data systems, which was actually really interesting because I had got to work on data systems which were much bigger than anything I had done previously. There, you can imagine LinkedIn as an internet company has really large amounts of data, and so simply the infrastructure of processing that actually becomes an interesting technical challenge. I got to work on those technical systems there, and learn a whole lot from the good work people were doing there.

                                             I then decided I wanted to write a book, which would explain these kind of data systems. The infrastructure that you need for processing and storing data at large scale, there wasn't really any book that described this stuff pretty well. There were a whole bunch of different books explaining you how to use a particular tool like say Cassandra or MongoDB or whatever you want to use, but nothing really that helped you figure out what kind of tool you might use in which circumstances. That was mostly spread out in knowledge just in people's heads or somewhere in conference talks or in random blog posts, but not really systematically put together anywhere.

                                             Then I decided, "Okay. I've done the start-up thing. It's been interesting, but I actually want to concentrate more on these technical things because I feel I can add more value there." I ended up taking out a year after leaving LinkedIn and, just on my own dime, writing this book for O'Reilly.

Peter Cowley:                  Which is called ...

Martin Kleppmann:        It's called Designing Data Intensive Applications. It's got a wild boar on the front in case that's important.

Peter Cowley:                  That has been quite successful. You've sold more than you thought you were going to.

Martin Kleppmann:        It has been surprisingly successful, yeah. At the time when I was writing it, I thought, "Well, if I sell 10,000 copies over the entire lifetime of the book, I'll be happy. I'll consider that a bestseller." Now it's been less than a year since publishing it and we've already sold over 20,000 copies, which for a niche technical book is doing pretty well. I believe the number one bestselling databases book on Amazon at the moment.

Peter Cowley:                  Well done. Is there another book in you? We’ll come to that story.

Martin Kleppmann:        Not all that soon anyway. It's hard work. It's hard enough writing one.

Peter Cowley:                  Because your day job now is an academic role, isn't it?

Martin Kleppmann:        Yeah. Maybe six months or so after I left LinkedIn and I was working on this book full-time, I got talking to some people back in the computer lab, the computer science department in Cambridge, about some topics that I found interesting. It just so happened that they were trying to get a research grant to work on a topic that was exactly something that I wanted to do anyway.

                                             So I wanted to stay in touch with them about this, and then some months later that grant did indeed come through, and so they had a job opening for someone to work on some stuff which was precisely stuff I wanted to do anyway. I talked to them and, yes, ended up being there in this post doc position, which was a bit weird because I never actually did end up doing that PhD that I had talked about previously. But somehow nevertheless ended up in a post doc. I think just the book got treated as sufficiently academic work. It's kind of equivalent to a PhD.

Peter Cowley:                  Can you call yourself Dr. Kleppmann yet?

Martin Kleppmann:        No, I can't because I don't formally have it. I might at some point try and actually get that formal piece of paper just to seal that off, but at the moment I'm a doctorless academic.

Peter Cowley:                  Can you talk about the project, or is that ...?

Martin Kleppmann:        Yeah, of course. The vision there is really we want to make apps that have the same convenience as web apps. So say Google Docs, which is super convenient. You can collaborate with several people on a document shared through your web browser, but the problem there is you're having to trust Google with all of your data because all of the documents are just stored in plain text on Google's servers. What we're trying to do in this research project is figure out how can we have that similar kind of apps, but without having to trust anybody's servers.

                                             Also, while we're at it, enabling a few other convenient things like they should work offline so you should be able to just edit offline and then resynchronize online when you happen to have an internet connection available. We want to be able to use local networks so if you have Bluetooth between your phone and your laptop, you should be able to synchronize between those two even if you don't have an internet connection available.

                                             Really it's silly that everything has to go via a data center in Virginia. We're trying to build the foundations of applications that, in the future, will allow those kinds of data synchronization. This is really something that nobody knows how to do exactly yet. It's still very much a research project. There's no way we could just go and build a product on it right now simply because there are too many unsolved technical questions. I do hope that maybe with another two, three years of work, we'll solve the hard technical questions, and then it will actually be possible to build products on top of it.

Peter Cowley:                  That'll probably be open source, won't it, I suspect?

Martin Kleppmann:        At the moment everything we're doing is open source, absolutely, because it's academic, it's for the good of the world. We don't have any commercial tie in there at the moment at all, and that's really the way I think it should be because this is foundational technology that other people should use to build their businesses and their products as well.

Peter Cowley:                  Well, Martin, that's been very, very interesting. I've learnt more on this, you know, keeping very close touch. I've also got to thank you tremendously for you putting me where I am. It was only during that first two years where I effectively I cofounded rather than angel invested, although albeit I got a good multiple on a small number. It was during that process I even learnt what an angel was, and this is only about 10 years ago, nine years ago.

Martin Kleppmann:        I think we must have learnt very much from each other then.

Peter Cowley:                  That's right.

Martin Kleppmann:        I found it tremendously valuable.

Peter Cowley:                  It's been a very successful open relationship where we both obviously clearly learnt from each other, so thank you very much, Martin.

Martin Kleppmann:        Thank you too, Peter.