My thoughts on Mendeley/Elsevier & why I left to start PeerJ

I joined Mendeley as head of R&D at the start in early 2009 and left end of 2011. Here are my thoughts about my time there and why I eventually left to start PeerJ

First, in terms of business success, regardless of your stance on Elsevier acquiring Mendeley, this was a win for the Mendeley team, so congratulations. I hear many saying that Elsevier overpaid (FT states £45M), but Elsevier is smart and they are getting a huge team in Mendeley that knows how to operate like a startup. That’s very near to impossible to grow organically with an established organization the size of Elsevier. This also gives Elsevier a major operational unit in London, which they did not have before.  That and other factors say to me that the amount paid is about right give or take a few percentage points. 

In terms of mission success, however, I am uncertain if this was a win. Mendeley had become known as the darling of openness, which in my view was already closing off when I left. Selling to Elsevier sets up a new challenge to maintain that open ethos, and unfortunately we can’t immediately gauge what the outcome will look like. 

I joined Mendeley after abandoning my own fledgling online article manager and recommendation tool that I had started while still finishing up my doctorate. I started talking with the Mendeley founders in the Summer of 2008 just before they opened for a beta program. I was really impressed with their plans to “disrupt” the academic market then held by EndNote and also to allow wider discovery of the literature. At Mendeley I was responsible for building out new discovery tools with the data that we were collecting. 

The R&D group did a lot in a few years (deduplicating 200M documents, real-time stats, search, Mendeley Suggest, etc), so I’ll mention just three of my projects below and what type of indication it creates about how Elsevier approaches things that are open.

One of my early projects was the “Open API.” It took some convincing at first to the founders that by releasing the extracted article data we would actually grow, but to their credit it didn’t take long for them to agree. I wrote the spec in late 2009 and worked with Ben Dowling (now an engineer at Facebook) to develop the first beta version of the API that was released in April 2010. After Ben left, Rosario jumped in to improve and expand the API and still heads it up in engineering. The API is now used internally at Mendeley for its own tools. From the start, the API was very controversial as we were now giving away metadata for free (titles, abstracts, etc) that publishers had always held onto tightly unless you were willing to pay for it. 

Next up were “PDF previews” that display anywhere from 1-3 pages of a PDF on each of the article landing pages on Mendeley web. The inspiration here was the 30 second previews that iTunes gives (and now every other digital music provider). The thinking was that this would benefit publishers by sending them traffic that was ready to engage with their articles, something that only seeing the abstract can limit. And we were right, the data we saw showed that articles with previews sent more traffic to publishers than without previews. To many publishers however, this was even more controversial than the Open API, as we were now showing even more content, although you couldn’t get access to it via the API. 

The third project was a JISC funded grant, DURA, that we co-wrote with Symplectic (now owned by Digital Science) and the University of Cambridge. The idea was to utilize the Mendeley Desktop client to connect to institutional repositories to allow drag & drop depositing. Put your authored papers into Mendeley and they are automatically deposited with your institution’s Green Open Access archive. 

All three of these projects were impacted by Elsevier. With the PDF Previews out, Elsevier came out hard to limit what we could do. The Mendeley founders, citing little choice, cave to removing Elsevier abstracts from the API and taking down PDF previews of Elsevier articles. Meanwhile, another publisher, Springer, surprisingly took the opposite approach and wanted us to do more with the PDF previews.

With the JISC funded DURA project, I got word that behind closed doors Elsevier was allegedly trying to stop the project altogether at JISC. That project has since stalled since I left Mendeley. 

These events were entirely against my personal open ethos and why I had originally agreed to join Mendeley. It was clear that the founders and I no longer agreed on the future (a blog post I had written calling out Elsevier’s actions was censored during the intense Elsevier talks) and tensions led me to reconsider my time there. It eventually came to a head and we agreed to part amicably. 

If one is honest, from a business perspective the Mendeley founders did the right thing to comply with Elsevier’s demands. My personal passions about Open Access hindered that, so no surprise it didn’t work out for more than a few years. What I learned was that my next project had to have open at its core, rather than just tacked onto the side. For that reason then I co-founded PeerJ, an Open Access journal, with one aim of never being in the position to take shit ever again from a closed publisher. 

I think that Mendeley as it stands today will continue to be useful even at Elsevier. That said, I think it will be challenging for Mendeley to become a truly transformative tool in science, which is what had originally convinced me to move from San Francisco to London four years ago. I cannot take all of the credit for the “open street-cred” Mendeley has gained over the years, but the projects I worked on had a big contribution to that and I can’t help to think that the day I left was the last day of further open innovation. It will be interesting to see what happens. I sincerely wish Jan, Paul, Victor and the team at Mendeley well.

[Full disclosure: As an early employee I own shares in Mendeley]

Updated (10 April 2013): Let’s keep it classy Internet. Since I’ve witnessed some personal attacks against a few current Mendeley staff I thought I’d step in and state that there are still voices for change at Mendeley. I don’t want to leave anyone out, but since I saw attacks against them; three such people are William Gunn, Steve Dennis, and Ricardo Vidal. I know their character having worked together for years and know that they will do their best to make changes from within and wish them the best.