July 3, 2007

Interview with Timo Hannay, Head of Web Publishing, Nature Publishing Group

Welcome to the most recent installment in my occasional series of interviews with people in the scitech world. This time around the subject is Timo Hannay, Head of Web Publishing at Nature Publishing Group, publishers of Nature and other associated journals as well as web products such as Connotea, Nature Reports, Nature Network, Scintilla, PostGenomic, Nature Precedings and others. Way back in May I was contacted by Natasha Ighodaro of Nature to see if I would be interested in interviewing someone to talk about some of their new web products. Eventually, she put me in touch with Timo. As it happened, Nature was in the middle of rolling out a bunch of web products, so it took a while to actually get the interview down on pixels. In any case, I'm very happy with the results and very grateful to Timo for submitting to such a long interview and for responding with such Candor. Enjoy!

Q0. Timo, please tell us a little about yourself, your background and how you ended up as Head of Web Publishing at Nature.

It’s quite a long story, so here’s a slightly abridged version: I’m a scientist with an undergraduate degree in biochemistry from Imperial College, London and an doctorate in neurophysiology from the University of Oxford. (My specialty was synaptic plasticity.) I finished my doctorate in 1994, followed by a year of postdoctoral research at Waseda University in Tokyo in 1994-5. Back in those days I was also a freelancer for The Economist, and through a colleague at their Tokyo office I got to know the people at Nature Japan too.

I’d always been a big fan of Nature — my dad bought me a subscription when I was about 18 (which I’ll admit is pretty geeky) and when I was at Oxford my first paper was published in the journal. When I met the people at Nature they were just launching Nature Medicine, and in my spare time I started covering medical research stories for them from Japan. I then lost touch for a bit when I went back to London to join McKinsey & Co.

I worked as a consultant in the UK and Japan for about three years, which was an intense and brilliant introduction to the world of business. But too many of the companies we were serving were in sectors that didn’t especially interest me. So, through a series of happy accidents, I ended up joining Nature’s Tokyo office, working full-time on business development. I had been into computers since I was a kid, and by then I was especially interested in the web. It also happened to be the case that doing more stuff online, and doing it better, was the biggest business develop opportunity for Nature in the Asia-Pacific region at that time. So that’s what I focused on: developing their Japanese website, and adding Chinese and Korean sites. In late 2000 I moved to Nature’s London office to work on the main site, Nature.com. As part of that move, Howard Ratner, Nature’s CTO, put me in charge of a new team (of about 3 people) called New Technology.

We experimented internally with things like RSS, SVG, RDF and other three-letter acronyms, but as a technical team the scope for us to turn these into new services or businesses was somewhat limited. Sometime around late 2004 or early 2005 Annette Thomas, Nature’s managing director and now my boss, decided to create a Web Publishing department with a remit to experiment with the web in a much more user-facing way. Since then the team has grown to something over 20 people. I love what I do because it’s at the intersection of my main interests: science, technology and business. I’m only sad that I don’t have much opportunity use my Japanese any more. ;-)

Q1. Some of Nature's recent journal publishing decisions have been quite controversial among librarians. Nature Physics is a good example. Do we really need another Physics journal?

I have very little to do directly with our journals because my focus is explicitly on non-traditional online products and services. So I can only give my personal option, which is that if there wasn’t a need for any of our new journals then people wouldn’t submit their papers or subscribe to them. I honestly believe that we do a much better job than most other scientific publishers, and that we create better products. That’s why they’re successful. The Nature Reviews series is a great example. Until they came out, the typical editorial and production standards for review journals were, in my opinion, very low. Nature Reviews set a new standard. Our other titles do the same in their respective fields, and considering how heavily read and impactful they are, they’re also extremely good value. Cynics may think that I’m only saying this because I work at Nature and they pay my salary, but in truth it’s the other way round: I choose to work here because I believe that Nature does great things (and I certainly didn’t move from management consulting to scholarly publishing in order to improve my bank balance ;-).

Q2. First Connotea, Nature Network, the Nature Blog, Second Nature (Nature in Second Life): you seem to be getting into Web 2.0/social software in a big way. What's Nature's strategy for these types of initiatives in the longer term?

I think it’s important to realise that we don’t just work on participative "Web 2.0"-type services. We also do a lot in the area of scientific databases (see http://www.nature.com/databases) and podcasts (http://www.nature.com/podcast). But to concentrate on the Web 2.0 stuff: we’re basically trying to identify ways in which scientists can use the web as a collaborative environment. The web isn’t just a broadcast channel or a convenient way to ship PDFs around , it’s a completely new kind of medium in which our "readers" can connect with each other.

Since our job is facilitating scientific communication, if we can’t help scientists to make the most of the web — the most powerful communication medium that humans have ever known — then we’re not merely missing opportunities, we’re simply not doing our job. So at Nature we’re trying a bunch of different things, often inspired by interesting ideas we see outside science (Connotea is clearly based on del.icio.us, and Nature Network on things like LinkedIn and Facebook), but always tailored to what we think will be of most use to professional researchers and clinicians. We’re trying to test the boundaries of what we can do, and we’re not afraid to fail, though of course we always do our best to succeed.

In line with many web-based companies, but in contrast to the scholarly journals business, our services typically launch in a fairly basic form, then we develop them in response to usage patterns and feedback. Now that we have quite a few different services, you can also expect to see them start connecting together more.

Second Nature is a bit different. We’ve been following Second Life for two or three years now, and I think it has the same kind of disruptive potential that the web had in the mid-90s. (Whether it realizes that potential depends on a lot of things, so it’s far from certain.) It could become a profoundly important medium for scientific communication and education, and we want to be there early, understanding its strengths and limitations, and working with early adopters among researchers and educators to find out how we can add value. So far it’s been a positive and eye-opening experience; I’m optimistic about the long-term prospects.

Q3. How has the uptake been for Connotea, Nature Network and Second Nature? Is there a critical mass yet to make these social environments compelling to scientists and others? How will these social networks tie into the core journal publishing business?

Connotea has a user base somewhere in the tens of thousands (the exact number depends on how active a person has to be to qualify as a "user"). Nature Network is much newer so is still in the thousands. I don’t know the visitor numbers for Second Nature, but in terms of active contributors I guess we have a couple of dozen people engaged in creating things on our virtual land, which now extends over three islands. Connotea has enough usage to create interesting second-order effects. For example, it can do quite a good job at recommending things to you based in what you’ve bookmarked. Nature Network activity has grown extremely rapidly in the 4 or 5 months since launch and is approaching a level at which we would expect see that virtuous circle in which usage (e.g., in the form of forum posts) drives yet more usage (e.g., other people coming in to read the posts). There are numerous ways in which these could tie into our journals — Connotea-generated lists of recommended reading, links to articles authored by people in your personal network, etc. — but we’re much more focused in making these services useful in their own right.

If we achieve that then we should also be able to turn them into successful standalone businesses, even though that usually won’t be through the traditional route of selling subscriptions. We’re also trying to open up these services for others to use. For example, Connotea has an API (application programming interface) that we’ve used to create tagging and "related article" functionality for the institutional repository software, EPrints. Other people have used it to do similar things with their own web and desktop applications. The Connotea code is also open-source, so there are quite a few private instances, for example behind institutional firewalls. Some of those people have also contributed code back to the open-source code, which is great because we can’t possibly develop all the requested features on our own.

Q4. And speaking of Web 2.0, peer review is a core value in science. There's a lot of experimentation going on out there with alternatives to peer review, even Nature has stuck it's toe into the water. Where do you think this is headed -- no big deal or long-awaited revolution?

My personal view is that peer review is headed for a revolution at some point, but the timing is extremely difficult to predict because it depends mainly not on technology but on various interdependent and imponderable social factors. It could be in a year or in twenty years. Having said that, there are many people at Nature who are much more knowledgeable than me about these things and who think we’re going to keep more or less the current model of peer review for the foreseeable future.

The reason I think they may be wrong is that I basically buy the "wisdom of crowds" argument: there are plenty of examples of the web causing new, open and collaborative approaches to replace traditional, closed and proprietary ones -- from open-source software to Wikipedia. You don’t always get a better result to begin with, which is why skeptics find it easy to be dismissive (as they were with both open-source software and Wikipedia in the early days). But as they evolve, and particularly as more people join in, they get better until the results match or even exceed the traditional approaches, often at much lower cost. (Anyone who’s read Clay Christensen's work will recognize this as an important part of his "innovator’s dilemma" argument).

I also believe that the web is particularly well suited to a "publish then filter" approach rather than the traditional "filter then publish" approach that was required when publishing was necessarily a physical-world process. As you can tell, my belief is based on rather abstract reasoning, and by looking for analogies outside science, so even I’m not completely convinced by it. But I’m convinced enough to know that we ought to be pushing the boundaries, because peer review is completely central to what we do, and if there’s a better way to do it then we ought to be the ones to find it. But at least in science, no one has found it yet.

Q5. Tell us a little about your new product Nature Reports, how it was developed and what need you see it filling in the scientific information marketplace. Who do you see as its main audience?

I can’t take any personal credit for the Nature Reports series, but I can tell you a bit about it. It consists of three sites — on Avian Flu, Climate Change and Stem Cells — that aim to serve a couple of purposes. First, they provide sources of scientific information on topics that affect us all, and that are all too often the subjects of spin or misinformation. We want to provide a place for non-experts to go where they know that the information is scientifically up-to-date and unbiased. They go into more depth than the mainstream media, but not so much that any interested and intelligent person can’t follow.

Secondly, we want the Nature Reports sites to become places where scientists, policy-makers, business people, and other interested parties can come together to learn from each other. Particularly in the three areas currently covered by Nature Reports, science does not and cannot operate in a vacuum. It must be willing to give and receive information in a way that will help us all — together -- to make wise decisions on questions that could affect our world for generations to come. For example, there’s no way that scientists can decide on their own what we should and shouldn’t be doing with stem cells, because those decisions are ultimately social and moral ones, but science needs to inform, and be informed by, the debate.

Q6. What do you think the future of print journal publishing is in 5 years? 10 years?

The vast majority of journals are already accessed mainly online. Many forward-thinking organizations are morphing their libraries into places to work and meet, not primarily places where documents are stored. Some libraries are even becoming entirely virtual. And that’s even before you consider the rise of scientific databases, which are just as important as journals in many fields and are all online. So I think we’re already in a world where scientific information is primarily digital. Within 10 years, I think most journals won’t any longer exist in paper form because there won’t be any point. (Though people will continue to print individual items for reading.) Nature and one or two other journals will be exceptions because they are effectively also magazines that contain news and commentary as well as research. Many people (including me) still prefer to read the "front half" of these publications in print, but eventually they too will migrate to e-readers of some sort. However, predicting the timing of that development has already caught out a lot of people who are much cleverer than me, so I won’t try here. ;-)

Q7. How about journal publishing itself? In 5 or 10 years will we be able to recognize whatever it is that journals have evolved into? Is the very nature of scientific publishing headed for some sort of transformation?

I think the concept of the scientific "paper" will remain intact (even if that name will seem increasingly anachronistic). There’s real value in this unit of publication, which tells a story by explaining how something previously unknown has become know through a particular set of experiments. But beyond that, there’s a lot of potential for change. Smaller units of discovery will be published -- whether through blogs or databases or whatever -- because the barriers to publishing them are now so low. This, in turn, will create the need for new services to find and collate this information, preferably in a personalized way, and new measures of scientific impact that take into account such contributions, which will be much smaller and more numerous than published papers.

Journals will become better linked, easier to search, and more dynamic. Many databases will take more seriously the need for curation, peer review, citability and archiving. In this way, journals and databases will be harder and harder to tell apart, and I think the distinction between them will ultimately become meaningless. In cases where journals don’t add much editorial value -- whether through filtering or otherwise improving the content -- the concept of the journal itself may start to erode as readers become ever more concerned with the paper they are reading rather than where it came from.

Q8. Can you tell us a little about Science Foo? It looks like a lot of fun -- not something we normally associate with science publishing.

Science Foo Camp is certainly one of the most fun and cool things I’ve ever done at work. It’s based on a meeting format invented about 5 years ago by O’Reilly Media, the influential technical book publisher run by Tim O’Reilly. They run an annual event for techno-geeks called Foo Camp. (“Foo” is a word computer programmers use to denote some arbitrary value or name -- like “x” in algebra -- but in this case also stands for "Friends of O’Reilly".) Basically, Tim and his colleagues invite 200-300 interesting people to their HQ in Sebastopol, CA for a weekend of self-organised demos, presentations, brainstorming, contraption-building and musical jamming (basically whatever people find interesting).

The great thing about it is the quality and variety of people there: software billionaires, technically precocious teenagers, engineers, scientists, writers — you name it. The only criteria are that O’Reilly consider them to be doing interesting stuff, and they want to introduce them to others. So it’s a bit like a giant, manic, weekend-long dinner party for geeks. They’ve become something of a legend in techno-land. Anyway, I attend a lot of O’Reilly conferences (it’s where I steal most of my best ideas ;-) and have known Tim for several years. Last spring, at his Emerging Technology Conference in San Diego, following a conversation he had had with Linda Stone (a brilliant ex-Apple and –Microsoft person with a keen interest in science and medicine), Tim suggested to me that we organize a Science Foo Camp. I thought it was a great idea.

We then spent a few weeks looking for a suitable venue, during which time Tim asked Eric Schmidt at Google, who loved the idea too. That was in late May or early June last year, and we decided to hold the event in August, so we had only two months to get lots of interesting scientific people to the Googleplex. We were really worried that it would be too short notice to get the kind of people we were seeking. We were also worried that scientists from diverse fields might not have as much to discuss with each other as people from the technical realm. We needn’t have been concerned: it was a great success. Attendees raved about it and several went away with not just new ideas but new collaborations. One thing that worked really well — aside from the great venue and format — is that we included some non-scientists in the mix. These ranged from technology people with a strong interest in science to sci-fi writers and others with cultural links to science but from outside research. I think those people helped to foster a truly interdisciplinary mindset. We’re doing the same again this year, though this time we have had a bit more time to plan it. I really hope that we’ll be able to do this every year from now on.

Q9. Who do you think your biggest competitor is? Open Access journals, other society or commercial publishers or even just the notion that everything is available for free on the web?

None of the above. ;-) To be honest, I don’t spend much time thinking about any of those. Open access will come about mainly through funder-mandated self-archiving, not author- or sponsor-funded journals. Of course we compete with other established publishers too, but they are a relatively known quantity. Your point about everything being free is related to an issue that I think is critical for publishers of all stripes: how to create viable business models that don’t involve charging for content (whether readers or authors). That’s not because I believe it’s necessarily going to become impossible to do charge readers, but it won’t always be the optimal (or even a viable) business model, especially for collaborative online services, so we need other options. In short, we need to get much better at monetizing traffic.

But to answer your question, I think our biggest competitor is the unknown grad student in his (or her) dorm room hatching a plan to turn scientific communication upside down in the same way that Napster, Google and Wikipedia disrupted other industries. Such people are a threat precisely because of their obscurity and lack of any historical baggage. You no longer need a lot of money, or even necessarily a strong brand, to succeed online. Good ideas and implementation are much more important. That drives almost everything we do. I hope that we’ll come up with the best ideas and implementations first, not mainly because of a commercial desire to out-compete others, but because that’s how we can best support scientific discovery.

I see myself less as a scientific publisher and more as a scientist who happens to work in publishing, helping information about ideas and discoveries make their way as quickly and efficiently as possible from their originators to those who can put them to use. If I ever thought I wasn’t being effective in that role, I’d find some other way to spend my time, probably outside publishing, but almost certainly connected with science.

Q10. Nature Precedings almost seems like the boldest of Nature's recent web offerings, nudging the larger scientific community into the same direction as, say, the physicists. What was the rationale behind introducing the service, and what do you see as it's place in the Nature suite of web products?

The basic rationale is that it’s in the interests of science for researchers to share their findings with each other as early and openly as possible. As you say, this already happens in physics through arXiv.org (and Paul Ginsparg, who runs that service, has very kindly offered his advice as we’ve been setting up Nature Precedings).

There are all sorts of theories about why it doesn’t happen so much in biology and other fields, but we thought the time was right to try and kick-start it. For one thing, there seems to be an increasing acceptance and understanding of the power and value of the web in enabling open collaboration, whether through domain-specific scientific databases or much more general services like Wikipedia. We were also able to get public support from some outstanding partners: the British Library, the European Bioinformatics Institute, Science Commons, and the Wellcome Trust (with more to come, I anticipate). This is key because the barriers to adoption are much more social than technical, and no one organisation has the right mix of skills and influence to pull this off on its own.

For the same reason, we’re also reaching out to other publishers. I expect a few of them will be cautious at first, but many of them clearly appreciate what we’re doing, which is about complementing the journal system, not competing with it, and about building an open federated system, not a closed proprietary one. For our own part, Nature Precedings helps us to engage with scientists at an earlier stage of the research process, which supports our traditional journal activities.

Also, by moving early we hope to be among the first to work out how best to make this kind of service economically self-sustaining. We’ve already made clear that that won’t involve charging for access -- and we’re working with some of our partners to set up open mirror sites to guarantee that.

Q11. Scintilla, PostGenomic, Nature Reports, even Connotea, all seem closely related to me, all about organizing information and bringing it all together. Are all these services coming together eventually or are they going to get more differentiated?

To be completely honest, that’s not yet certain because it depends on how people use those services, and what they tell us about their needs. But my expectation, and our current intention, is to steadily integrate them in a way that will allow information from one application to be used within another, and for users to hop between them seamlessly. Ultimately the distinction between these different services should therefore become less and less pronounced.

I think that’s a good thing because people just want help with their scientific information needs, they don’t want to have to work out whether Connotea or Scintilla (or whatever) is the answer, and they certainly don’t want to have to visit several different sites to conduct a single task. That doesn’t mean they will all turn into one monolithic application, but it should become easier for (say) Connotea users to access Scintilla functionality, and vice versa. We’ve certainly put a lot of thought into making that kind of integration possible, but to what extent we pursue it ultimately depends not on us but our users.