Wednesday, July 4, 2007

pushing the boundaries of advertising in science

Soon after starting their first lab, most new professors are disappointed to find that they spend a disproportionate amount of time on marketing and fund raising. There are a lot of smart scientists doing interesting work, and in the end, the ideas that are shouted loudest and most frequently become the accepted doctrine at any given time. That's not to say that working hard towards solving important scientific problems isn't important. But if you work hard and solve an important problem, it is very likely that no one will know unless you go out and advertise it (Mendel, Newton, and Einstein were all lost to the world initially because they were closet geniuses). And since everyone is advertising the important problems that they've solved, science becomes something of a popularity contest.

To be popular, you need to be a constant member of the lecture circuit. Every field has their own set of important conferences. I do bioinformatics research. If you want people to know you in bioinformatics (unless you are old and have already established a reputation through many years on the lecture circuit), it would be very useful to give a lecture at RECOMB, PSB, and ISMB. Enough people hear your story, enough people hear your story again, and again, and again, and they start believing. They start telling their friends. Next thing you know, people with pocket protectors are coming up to you on the street asking for your autograph.

The problem with the lecture circuit is that it is pretty expensive. After you pay for the flight, hotel, etc... you've spent $1000. But you can't just bring yourself. You need to bring a lab member or two with a poster, so that your students can start learning how the lecture circuit works. So the conference costs you $1000-5000. Our tax dollars hard at work.

I know, I'm a little sarcastic, and sometimes useful things like long fruitful collaborations get started at conferences. But you get my point. What we're doing with all of this is a roundabout version of what coke, pepsi, nike, apple, and state farm do more directly. We're securing name recognition and our place in the marketplace.

I propose that we be more direct. Why not skip one conference a year. That saves about $3000. Take this money and invest it into Adwords, Google's text based advertising product. Let me giving an example. Recently I was involved in some network inference work that resulted in the PLoS Biology publication,
Large-Scale Mapping and Validation of Escherichia coli Transcriptional Regulation from a Compendium of Expression Profiles. Along with our analysis, we collected 445 E. coli Affymetrix microarrays, and we organized a set of software for benchmarking our algorithms and future network algorithms using the large amount of regulatory information that's already known for E. coli (yes, I'm marketing now, so go check out the paper if you're interested in network inference).

With Google Adwords, you bid on keywords; when those keywords are searched for and you are the winning bidder, your link and a little text goes up on the side of the Google search. For example if someone searches for "network inference", I might want a link that says Benchmark Your Network Inference Algorithm using our open source matlab scripts. Or if they type "E. coli affymetrix", I might want a sponsored link ad that says, Download custom datasets from a publicly available E. coli Affymetrix compendium at M3D. Since there's zero competition for those keywords, each click on my advertisement costs the Adwords minimum: 5 cents. And unlike my $3000 conference where the lecture hall is empty because I got the 8AM slot on the last day of the conference, these people are actually interested enough to click on my ad, which means they'll probably have a decent look at what I have to say. At 5 cents a click, my $3000 will give me 60,000 people that might find my data useful - more than the largest of conferences.

We want to create useful science that others can build on. And we want to build on other peoples' useful science. Why shouldn't we pay a dime to find each other? I know some people will think this is going too far, but we're doing it indirectly already. And why are we doing it? To a large extent, because no scientist has any free time. So if you want another scientist to look at your science, you've got to stick it right in their face. Once it's there, they can decide if it's the science they're looking for.

I think Adwords could free up a lot of time for professors, allowing them to spend less time marketing and more time on the original focus of their professional lives: solving important scientific problems.


Jean-Claude Bradley said...

I have used AdWords to recruit students to our department. But for science research, I don't think it is necessary. Just maintain an active blog or wiki and Google will take care of ranking. For example we have been working on the Ugi reaction in my lab. A search for "Ugi reaction" in Google pulls up 3 of the top 10 hits to our work.

Jean-Claude Bradley said...

It is true that conferences can be expensive and inconvenient. A very simple way to leverage the talks that you do give is to record them by screen capture and make them publicly available. That way people that missed your talk can still watch it. I've started to encourage our grad students to record their thesis defenses as well.