In the February 16 & 23 issue of The New Yorker, James Surowiecki devotes his The Financial Page column to the unproductive R&D departments in big pharma. My first thought was, “That sounds just like Microsoft Research.” Now, I don’t know about Microsoft Research in detail, but I’ve occasionally run across their work since 1998, when I saw a couple folks present at CHI on whether web sites perform better as broad and shallow, or narrow and deep.
As this list of projects suggests, Microsoft devotes tons of money to research and development. But to what end? I admit I can’t say much about many aspects of the computing world, but with respect to interaction design, all that I know that’s actually emerged from Microsoft research is the office assistant, AKA Clippy, and we know how well that’s done. There are numerous other projects that have produced academic papers and little else. Marc Smith has recently gotten a lot of attention for research he’s been conducting over the last 5 years on social software, though what people outside of the industry don’t seem to realize is that no one cares about methods of conversation on USENET.
I think about this as I sit in front of my 12″ PowerBook G4, which is riddled with interaction design innovation. Perhaps my favorite is LaunchBar, an app that lets me launch pretty much any program or file with just a few intuitable keystrokes. Or OmniGraffle, whose smart alignment and distance guides make positioning graphics a snap. Or SubEthaEdit, enabling easy-as-pie collaborative document editing.
And, on the web, we’ve seen things like Google (started by two guys from Stanford), blogging tools like Blogger (started by Pyra when it was three people), MovableType (Ben and Mena, so, two people), Slashdot (CmdrTaco and Hemos), Netomat (mostly Maciej),
Perhaps Steve Jobs was right to kill Apple’s vaunted Advanced Technology Group. It seems that product teams are responsible for their own innovation, and, what do you know, it’s working (Rendezvous, iTunes, Expose, etc. etc.)! Contrast this with Remail, from IBM’s research labs, which garnered some buzz when the site launched, but which I doubt we’ll ever see in any piece of shipping software. I mean, I love Babble as much as the next geek, but where is it getting us? And where is it getting IBM?
I am in complete agreement with this post, Peter, and I was surprised when Yahoo! announced it’s own advanced r&d department a few weeks ago. My first thought was, “oh boy, they’ll shut that thing down in less than 3 years, gauranteed.” Product teams are the only people motivated enough to innovate as quickly and practically as they need to, because their lives depend on it. R&D in a bubble gets management a bunch of really cool demos that end up having little real world value, or worse, some truly practical innovation that gets buried in the lab, where the fun is always in prototyping the next cool thing, not thinking about how to “productize” the current project.
Peter, I’m not sure writing off all of industrial research in HCI is necessarily the right idea. It’s worth considering that most research looks out several years in advance, attempting to identify concepts that may not make it to the market until several years later. All of the work at Xerox PARC didn’t make it into the Macintosh for over a decade. And work such as Lifestreams out of Yale that was done seven or eight years ago is just now making it into the plans for Longhorn. So while it may be true that industrial HCI research rarely becomes productized within 2 or 3 years, this doesn’t necessarily mean that the whole enterprise is a waste of time.
Part of the reason I think it takes so long for this work to make it into products is that the product development teams often like to design software that is familiar to customers; radically different interfaces have a much harder time getting adopted. Nonetheless, a lot of research never becomes adopted even after several years. So a better question might be: How can industrial research more effectively create concepts that are sufficiently useful in the shorter term to become adopted?
Why assume that the goal is to productize the R&D output?
Microsoft, to take one example, strives to build their brand on being The Best Company to Plan YOUR Future. Today they’ve got PR for SenseCam (http://www.globetechnology.com/servlet/story/RTGAM.20040304.gtmsideamar4/BNStory/Technology/ should be one link to the story), a product to document everything in your life – probably related to MyLifeBits, etc. This stuff makes great stories that we tell each other, even if it’s with skepticism. Bill Gates published a book of hand-waving techno visions a few years back, and everyone talked about it.
That brand – the intangible asset that makes Microsoft so incredibly valuable – is one of the very useful outcomes of that work. Do they have to ever productize any of that stuff? Doesn’t seem so – they can keep innovating in the future and let the Office 2112 team worry about the product stuff. I think it works.
What about patents? IBM produces a great deal of income from patents stemming from research that may never produce products. I don’t know how many patents Microsoft Research produces, but it is a good strategy.
Well, I have been reading your blog posts daily and the reason I come on your blog frequently is its compelling content… Regards…