Cracking open science

Three academic types decided they’d had enough of the academic publishing industry stifling knowledge. Ivo Spigel sits down with Victor Henning, co-founder and chief executive of Mendeley.

Ivo: So tell me how this whole story began.

Victor: Jan and I have known each other for twelve years because we went to business school together. The actual story of how I met Paul is a bit more interesting. I was speaking at a conference in San Francisco. Afterwards, I decided to do a backpacking tour across the US. I travelled to Berkeley, Las Vegas, Chicago…

In Las Vegas, I went to a youth hostel and got there at about 4 am in the morning. Some familiar faces left the youth hostel and I thought: “Where do I know these guys from? Maybe I’ve seen them at the youth hostel in San Francisco?” You know, backpackers travel along the same routes. One of them said “Victor?” So it turns out that those guys that I randomly ran into in Las Vegas were computer science students from the Bauhaus university and their lab was one floor above my office on the other side of the world.

The idea for Mendeley was developed back in 2005 when Jan and I started working on our PhDs. We were in different fields – mine was social psychology and decision-making and Jan’s was information management. Both of us had tons of PDF stuff that we had to manage and we thought: “Why isn’t there a good tool that tries to automatically turn our collections of PDFs into structured databases?”

Kind of like when you have a collection of mp3s, and you drag them into iTunes and then iTunes extracts all of the relevant metadata like the band, track title, the album and so forth. We thought there should really be a similar thing for scientific papers that extracts the author, title, volume, journal, issue and so forth from your collection of PDFs, so you don’t have to enter them manually.

Jan and I had been discussing this idea on and off for two years and in 2007 we got serious and got Paul to join the team. We took our savings and went to an outsourcing company in Belarus to develop a first prototype of the software on the web site which we could use to pitch to Stefan [Stefan Glänzer, founder of Last.fm, White Bear Yard and Passion Capital] and got in touch with him.

Stefan had been a guest lecturer at Jan’s and my business school. We wrote him an e-mail saying: “Stefan, it’s Jan and Victor from business school class four years ago. We have this idea which is like a Last.fm for research.” He agreed to meet in London. After one further meeting, when we showed him the demo, we essentially came to a handshake. So our first seed funding came together without doing any contract or signing an agreement, we just exchanged a couple of e-mails with the key terms and then a handshake deal which gave us our first investor!

Once Stefan decided to invest – that was the first seed investment of about €250,000 – we decided to move to London. We started hiring our own development team and kept fundraising. At that time we were speaking with pretty much every strategic investor. All of the big publishing companies like Elsevier, Thomson Reuters, Nature… they all wanted to invest because they felt this was really something that could disturb their business down the road. But they were more interested in getting a foot in the door of something that could possibly threaten them rather than being genuinely excited about building something great.

Stephan then introduced us to the founding engineers of Skype. They are Estonian engineers who did Kazaa and then Skype together with Janus and Niklas. We met their investment manager, who also had been the director of product for Skype and the former director of strategy. Both are actually now quite active here in the European venture and startup scene as well: Eileen Burbidge is now Stefan’s partner in Passion Capital. The director of product who came in for that pitch was Taavet Hinrikus, who is the founder of Transferwise. That startup is making waves transferring money between European countries.

It was Taavet and Eileen who came into our office in Covent Garden and we instantly decided that they were a much better fit than the academic publishers. First of all, they were much more straightforward and it was a great brand association for us: Skype! They had experience with something that we were attempting to build, which was desktop software coupled with a client server model. They have great user interface chops and we knew that we wanted to build something that was extremely easy to use.

The third major investor, who got on board a couple of months later, and who today actually is our biggest investor was Alex Zubillaga, together with Len Blavatnik. Alex is an active angel investor who was also head of digital at Warner Music for a while and he was also involved in the Warner Music buyout. Len Blavatnik is a Russian oil billionaire who’s based in London and who actually bought Warner Music recently for £3.3 billion.

He is also an active philanthropist in science: he donated £100 million to Oxford recently to establish a new School of Government. So, a guy with a deep interest in science and higher education but also incredibly deep pockets.

Let me give you the key numbers so far. In the three and a half years since our launch, we now have now grown to 1.9 million users. The biggest group of our users is in the US, 20 per cent. From the start, we had this ambition to be global. We specifically targeted the US, and, when we launched in 2009, I went on a tour of Ivy League universities on the east coast. I visited eight universities in ten days: Yale, MIT, Brown University, NYU, Penn University, Cornell… lots of high-profile universities.

What has been equally as astonishing is the growth of our database. We’ve always had this vision that by crowdsourcing, we could build not just another metadata database or not just another article database but one with all this social information that makes a database truly special and useful. So, for example, for every document in our database we know how many readers that document has, where are they from, what their academic disciplines are, what else they are reading, and the tags and keywords they are applying to the document.

Is the interest in this paper increasing over time or is it decreasing? There’s lots of social information, and we call that the social layer to our database. And that’s been exploding. So we now have 285 million documents uploaded to our database and on average we are currently getting about five to eight hundred thousand documents a day added.

 Ivo: Why did you target the US initially?

Victor: There was a very specific reason for that. The founders and the first investor were all German, so our network was naturally very German. If we had started by getting our network into the platform, our worry was that it would seem like a platform for German academics and then you just wouldn’t get a Harvard or MIT guy to sign up to the platform. From the other way around, if you have the top research universities like Cambridge, Stanford and MIT on the platform, then you also get the Germans to sign up. And so we specifically decided that the best way to really position us was to target the US universities first.

Ivo: Thousands of start-ups are trying to do location-based gaming, or social networking, or doing photography, connecting your mobile phone with whoever is in the room or something like that. But the whole academic area is very non-typical for a start-up.

Victor: From the start, we met lots of people who basically discouraged us and said: “It’s a niche, why would you want to do that?” Jan and I went on vacation with a friend who was at Boston Consulting Group. His whole job was crunching numbers. We were presenting to him our big vision about changing science and how we wanted to build this and he said: “I’ve looked into this market and I can really only see a market of 20,000 people for what you are doing, so you should really scrap that idea and do something else.” We were really pissed off by that.

That was in a way typical of feedback that we got from a lot of the VCs. When we first met them they all went, as a reflex, “Oh, academia. There is no money in academia and it’s a niche market, how many people can there be, there are maybe 2 million scientists. This is not another Facebook and I’m not looking into this further.” One thing that showed us that we are on to something was the universality of the problems that we recognised for ourselves – we saw that it’s not just for full time academics, it’s also students – if you ever had to write a term paper you have faced the same problems.

And it’s not just academics in universities but also engineers, doctors, nurses… These days, even though the majority of our users are from academia, we have users ranging from the US Marine Corps in Afghanistan to Government agencies across the globe to a film production special effects company in London. They won the Oscar for Inception and they use Mendeley to manage papers on machine vision and artificial intelligence. We felt that the appeal was much wider than this niche that VCs were thinking of.

Ivo: Victor, let’s step away from the business model issue. I’d like to ask you to comment on this huge issue of intellectual property management and intellectual property rights. The most highly profiled furores have been SOPA, PIPA and ACTA, but I noticed on your blog that Mendeley is also very vocal about the same area, concerning, of course, scientific and research information and whether it is closed or open?

 Victor: I think that actually ties back to the academic publishing industry. As I said, academic publishing is a $25 billion business but what it is built on is… the only term that I can find for some of what’s going on is “insanity”. When I was an academic and I published my first paper I wanted to get a PDF copy of my own paper and the publisher wouldn’t give it to me. My university hadn’t subscribed to that journal so I had to buy a copy of my own paper.

 Ivo: That’s amazing!

 Victor: The way it works is like this. As an academic, you are paid by the university and by public money, by research grants. You are writing a paper and you give it to a publisher for free: you are not getting paid anything for that content. Then you provide peer review services. The publishers are always saying, “We add value by providing peer review,” but, in fact, they usually only appoint the editor for the journal. The editor accepts the (usually unpaid) honour of being the editor of said prestigious journal.

The editor then uses his own network to bring in the sub-editors and reviewers and organises the review process. The reviewers put in the time to read the paper, give their critique and redefine the content. So, all of that is also provided and paid for by universities – hence by taxpayer money and government grants. Some publishers even charge you for layout and graphics, so if you have a colour figure to go into the paper then the publisher will charge you thousands of dollars to put that in.

Now, they take the content, they charge you and then they literally turn around and sell it back to the university at a profit! You can look at the publishers’ annual reports. Profit margins are between 30 to 40 percent.

 Ivo: I’ve noticed.

Victor: That’s in the range of the Apples and Googles to the tune of $25 billion a year. Even Harvard, the most well-endowed university in the world, says that this model is simply not sustainable. Lately, publishers are attempting to extract more money from researchers and universities by introducing additional copyright barriers. It’s not just about distributing content but it’s also about text mining which I think it’s the next big area of growth and research in start-ups as well, like big data.

Elsevier tried to sneak in legislation through lobbying – the so-called Research Works Act – that was to undo the National Institute of Health’s Open Access mandate and therefore would have prevented public access to millions of bio-medical research papers. They had to pull back after outrage and boycotts from the academic community.

Our vision is to crack this open by, first of all, encouraging academics to be aware of their rights, be aware of the lobbying that is going on, and also by enabling them to post their own publications over which they retain copyright, from their Mendeley profiles. A ton of papers have been made available by Mendeley in this way, by academics uploading their own papers. So, that’s our attempt at making content more accessible and science more open.

Ivo: When you say “other platforms in this space” on your web site, you directly compare Mendeley to typical competitors. Are they going in the same direction as you are? Is your model or the way that you are functioning unique in terms of generating the interest and people using it for their papers?

Victor: There is nobody who does exactly what we do but we do have competitors in different areas of our activity. What you see on our web site are the competitors that we list for the document management and reference management part. That is still where traditionally most of our users get the most value and how they find us.

People tell them: “Check out Mendeley, it will help manage your papers more effectively,” and that’s the way they compare it to traditional other tools in that space. I would say, other competitors, obviously from the social side, are some of those “Facebook for scientists” tools that have being popping up.

The biggest academic databases, commercially, so far, have been Thomson Reuters’ Web of Science and Elsevier’s Scopus and both have about 48-49 million records. We surpassed that after a little more than two years and, like I said, we are now up to 285 million uploaded documents.

If you de-duplicate that – we have to de-duplicate content because popular papers obviously get uploaded by multiple different people – our database currently has about 65 million unique records, so we are already 33 per cent larger than the commercial ones by Thomson and Elsevier and we are still growing exponentially.

Ivo: So in a few years you will be probably buying either Elsevier or Thomson Reuters or both?

Victor: I guess we would need a hedge fund to blow a couple of billion into our balance sheet so we can leverage the hell out of it!

Ivo: There you go!

Victor: Our vision is that we really want to build a sustainable, long-term business. A lot of our investment has been for the long term. The biggest part of our engineering is the platform and the data mining team: together that would be 8 people and we are currently 36. None of that would’ve been necessary if we had been out for the short term, just building, let’s say, a freemium or subscriptions model to the productivity tools.

But for us it’s really about the long game of building an ecosystem that may replace traditional curating and publishing – something that might become a content distribution platform that enables third parties to built apps that we haven’t thought of yet against that gigantic interlinked database – that we could then, as Mendeley, profit from, by either charging for access to the data or by getting a revenue share.

Ivo: You wanted to talk earlier about the business model and that we kind of put that aside. Obviously you started as a freemium app and the only thing that is visible on your web site is “you can get this much for free and if you want more space and few more features then you need to pay this and this amount of money”. Is there more to your revenue model than that?

Victor: There is, yes. Apart from individual subscriptions which you can see on the site, we also have team packages and institutional packages. The latter, which we launched two months ago, have seen amazing adoption: We’ve already signed up some of the leading research institutions in North America, Europe, and Asia as customers, including Stanford, the Harvard-Smithsonian Centre for Astrophysics, the Korea Advanced Institute of Science and Technology and the Japanese Agriculture, Forestry and Fisheries Research Council.

There is a really huge pipeline of inbound requests, and we don’t have a sales force at the moment. In order to help us fulfil that need, especially with respect to libraries, we partnered with a Dutch company called Swets. Unless you are a librarian, you probably haven’t heard of them.

Ivo: No, I’m not. So I haven’t.

Victor: I hadn’t either until they approached me. They make $1 billion in revenue helping libraries manage subscriptions. They approached us a little more than a year ago wanting to tie Mendeley’s usage statistics, which I mentioned earlier, into their library subscriptions dashboard, so that librarians could make better decisions about which journals they should subscribe to and how the content that they have subscribed to is being used by their faculty.

So, we partnered up with them to develop and market the Mendeley Institutional Edition, which is sold at between $10,000 and 50,000 per year per institution.

Ivo: And obviously you are going to be revenue sharing with them on this?

Victor: Yes, there is going to be revenue sharing with them. But the nice thing for us is that we don’t have to build up a sales staff and Swets has established sales relationships with, I think, more than 11,000 libraries in 160 countries. So, they are going to sell Mendeley through those channels giving us a ton of exposure to all of these libraries.

Obviously there are network effects. For a library that has purchased the dashboard, the data on the dashboard gets better the more people at the institution use Mendeley, so they also have an incentive to push Mendeley. We feel that that’s also going to be another huge boost to our growth. That is the productivity and collaborations part of our monetisation, both for individuals and for institutions and enterprises.

In the future, we also want to monetise the content and the data. For example, because our web catalogue, the database of articles that we have, draws so much traffic, we have signed agreements with several academic publishers to provide us with their content that we can integrate in the catalogue and thereby drive traffic their way. We have now signed the first affiliate agreements which will give Mendeley a share of the revenue of the articles that been sold in this way.

The opportunity for Mendeley is to become the iTunes or Spotify for research: to help publishers to reach end users directly through our installed base and through our recommendations engine. The third part would be the applications ecosystem that I have described. There are almost 250 third-party apps using our data, and they’re generating more than 100 million API calls to our database per month.

We feel that this is not something we should monetise right now, but the first businesses are already being built through our API and maybe by 2013-2014 we’ll have substantial businesses, big data businesses, dealing with research data, that we can participate in.

Ivo: Have you been thinking about establishing some sort of a more formal, structured presence in India or Asia? Research and academic development is growing quite fast in China and India. Have you been looking at that?

Victor: We haven’t. We built the New York office mainly to do business development because lots of academic publishers are in that area and also there are a lot of universities. We have partnered with Columbia University, the New York Academy of Sciences, the Sloan Foundation. Obviously the US is our most important market.

We are quite wary of opening additional offices because of the overhead and the transaction cost increases so much between the teams of management. Also, it’s more fun if people are working together in the same office and see each other every day and can build relationships this way.

Ivo: What is the long term vision? what is going to happen with Mendeley 2, 3 or 5 years down the road?

 Victor: It’s really going to become focused on the platform and the applications. I could see the current productivity features that we offer becoming a prominent part of the platform, but with the platform being the main focus. We are currently investing a lot of effort into rewriting our APIs to make our internal APIs addressable by external parties so that they can use functionality which is currently only available to our internal developers.

From a business point of view, our main revenue streams for the next two years will still be on the productivity and collaboration side. In the long term, I really see us becoming a platform for applications built by third parties. The Mendeley collaboration tools, our desktop tools and our web tools will be the part of that platform.

To return to your previous question, I think that platform could be a number of things and it’s quite hard to foresee what it would be. It could be an extension of today’s networking and collaborations suite, so it could be very much focused on the documents, but it could also become an intelligent network of raw data processing.

For example, we have a research grant that we recently got from the EU of €2 million together with three partners which is about building semantic annotation interfaces. People are creating links between content and saying this paper supports that paper, this paper refutes the other paper over there, these two papers use the same method… Plugging this type of information into the app platform could enable incredible applications.

For example, the iPhone has Siri, we could hook into that and ask: “Siri, what is the relationship between, say, green tea and Alzheimer’s?” You could have semantic engines ploughing through all the content that we have, doing the text mining for currencies of the words “green tea” and Alzheimer’s in papers, how they are related to each other, what are the experiments that have been performed and then kind of answering the question for you.

I really believe that those applications will become possible in the next few years and Mendeley would like to be the core part of that.

Editor’s Note: This interview is the latest in a series by Ivo Spigel, taken from his forthcoming book of interviews with European high-tech founders, provisionally titled EuroGeeks