Three things struck me as I reflect back on my trip to Microsoft Tuesday. One, it’s not wise to dismiss the company as being “at the first grade level” or “behind” the rest of the pack. Second, the fact that Microsoft came late to the search game just might be an advantage. And third, I need to get some face time with the Windows team, because I went into the day wondering if MSN isn’t becoming MSFT’s future interface/platform play, with Windows relegated to a supporting role (as DOS was to Windows) and, well, nothing I heard convinced me otherwise. And that certainly can’t be right. Or can it?
I met first with David Cole, who runs MSN and has a long history at the company, in particular with Windows and web technology. Much of what we discussed I need to save for my column, but suffice to say we covered a broad range of topics, including Longhorn integration (yes, MSN and search are being built with an eye toward that eventuality), the Web OS meme (while not dismissive of the idea, Cole thought it was just one of many approaches to solving computing and information service problems), and of course, competition with Google and the rest of the field. Cole began by outlining how MSN has shifted to its current strategy, based on building scaled software services that break into two major buckets: communications services (MSN Messenger, Hotmail, etc) and information services (search, content, etc.). Yusuf Mehdi, who I met with next, runs information services, and we had a pretty detailed chat about the present and future of search. That conversation was for the book alone, due to timing issues. I can report, however, that Yusuf was pretty charged up about what they are building.
So why did I leave thinking that MSFT’s late start in search could be an advantage? Well, think about it. The company has massive resources, and the folks in charge are pretty smart. So they get to tackle the search problem with no legacy issues and no presumptions with regard to approach. There are any number of hurdles in search – starting with how to scale your infrastructure and moving into how to integrate results with personalized data – and many of these might best be tackled by starting fresh. Plus, on the talent side, MSFT is really the only viable player that can offer engineers unlimited resources and the chance to start from scratch. I know, the Valley mill says MSFT is having a hell of a time hiring, but when I asked that question up in Redmond, I got quite the opposite answer.
I found both Cole and Mehdi personable and open, and willing to listen as much as hold forth. I did ask them any number of questions suggested by Searchblog readers, and I got a chance to talk to Scoble before I went in as well. I asked Scoble what I should ask the MSN execs, and he said he wanted to know when MSFT would launch blogging on MSN, what their RSS strategy was, how they plan to kick Google’s ass, and what they were going to buy. I asked those questions and more, and while some of the answers are under embargo for the book, here are a few of the highlights I can report:
– MSN Premium, a new subscription product, already has a rudimentary blogging tool, Cole told me, and blogging will be improved and incorporated in some way across the system.
– 3 degrees, a social networking application that is in testing in the Sandbox, has taught MSN a lot and we can expect to see some version of it attached to MSN Messenger. In particular, watch the music application.
– RSS: Well, Cole was not hip to RSS, but he said he would get hip pronto. (He also was not aware of Scoble, but that’s changed now!)
– Google: Very diplomatic on this front as you might expect. Respectful of how the company changed search and the internet for good. It’s not often that Gates admits defeat.
– What might they buy? We didn’t really get into this one – ran out of time. But I sense that MSFT is pretty into rolling its own right now. When the search platform is stable, I’d wager they might go looking for neat things to plug into it.
On the platform idea (point three), my general thesis is this: Over time, more and more of a typical user’s desktop real estate is devoted to web-enabled apps. I am an extreme example of this trend (and I’d wager the same is true of most of Searchblog’s readers): at any give moment, I’ve got ecto (a blogging tool), NetNewsWire (RSS reader), Firefox and/or Safari (browser), mail, and Office open. All these applications are web-enabled (Office is the lamest of the bunch, but not for long). Even OSX makes web calls – if only for software updates for now. So if you look at my screen, at least 80 percent of it is web applications. Compare that with five years ago, where it was just email and the browser, or ten, where it was just email.
Now, all these applications are migrating to the portals, and the portals are migrating to the model Cole described: software-based platforms replete with tools and applications – mail, calendar, blogging, rss readers, the works. At some point (and this certainly is not a new idea) the very idea of the “desktop” will become pretty old school. We’re building an entirely new architecture on top of our OSes. So…what does that mean for the traditional OS? In essence, it loses the glory role, in the eyes of the consumer. The OS does the hard stuff – files systems, security, connectivity, etc., but the interface, the stuff the user sees, is migrating to the web. And MSFT’s play there isn’t Windows. It’s MSN. At least, I think it is….
This, it seems to me, is why the company held its nose and lost untold billions through MSN’a rough patches (the service is now on a path to sustained annual profitability). And it also seems to me that this trend has something to do with Longhorn’s delays and missteps: how do you build a new operating system when its interface will live on the web, instead of on the desktop? I’m looking forward to learning more.
7 thoughts on “Microsoft: Built From Scratch, The Interface Migrates to the Web”
Thanks for the report. Hope you emjoyed you’re time up here in Seattle. 🙂
Microsoft emphatically does not see the future being about networks and servers. Sit through a Longhorn presentation and you’ll understand that they are in the midst of an all out assault to bring computing back to the PC client.
Also, Microsoft has a miserable track record of creating desirable products and services. Virtually all of the progress they make in new product categories owes itself to Windows and Office. Xbox, Expedia and SQL Server are fairly impressive. But Windows CE, Windows Media, Windows Media Center, MSN, all of Microsoft search, Mactopia, Microsoft Money, bCentral, etc. are lousy.
Want to make a good search engine? Start indexing more pages, start showing relevant results, stop charging (bcentral) $90 freakin lousy dollars for ONE page to be indexed by b crappy central every 48 hours, google indexes my page at least every 48 hours FOR FREE, this guy didn’t know what RSS is but he works for msn? hmmmmm maybe they should consider hiring some new staff…
just a thought
Well, the whole division knows about RSS now. (I did a presentation to them today on that topic). One thing is they are quick learners if they find something that they need to know about.
pb: agreed, but look at the Tablet PC, Media Center, OneNote. More coming.
I’m interested in the idea that applications are migrating to a Web browser platform. (If I’ve read correctly you forsee for the directions on the OS platform idea.) For the last 4-5 years, almost everything our company has built has had a Web interface. Our latest product is built on top of .NET, and the rich client UI that this affords.
Our experience has been that the ability to innovate and deliver new UI facilities rapidly through a Web browser (which Web browser for a start) is much more constrained than it is on a client desktop environment. Not that impressive UI cannot be delivered through a Web browser – the new Web front end to Exchange looks scarily like Outlook. But I suspect that most of the important lessons about how the UI should work were determined from years of building and improving Outlook first on the PC.
Of course, there is a middle ground, where services are available from a network through a Web services interface API, and the client interface can be built either using Web browser technologies or as a thin client. The customers can then choose which interface they prefer.
I agree with John’s remark on the increasing dependence on network connectivity for everyday applications. The other day, our internal VPN failed (for no apparent reason) meaning my Outlook couldn’t communicate with our Exchange server, located in our other office. Everything else worked just fine on the Web, even the Web interface to Exchange interestingly. Had the entire connection failed however, I would have needed offline versions of information to be available to do any work whatsoever.
In an old job of mine inside a computer science department, you could always tell who was really dependent on the network by the rapidity with which people would start clustering around the system admins when network connections went down for some reason. For some of us, it was usually within a couple of minutes 🙂
Here’s something REALLY REALLY easy for you to suggest Microsoft fix – quit stripping trailing “/’s” from URL pathnames that should be there in your search results!
I.e. trying doing a search for “hulkmobile” on http://search.msn.com/ and look at the entries for “Alek’s Hulk Car” … note that the target URL is http://www.komar.org/hulk … WITHOUT the trailing / … but if you surf there, the web server has to do the right thing FOR you and redirects with the trailing / …
The search engine should really show the correct URL (with the trailing /) and this is such a simple thing that Microsoft should get it right … on the other hand, Yahoo makes the same mistake! 😉
Google, no surprise, does do the right thing …
But why is the OS significant when it comes to web apps? I can already e-mail, shop, get maps, blog, chat, read RSS, search, and maintain a server-side archive all through one simple and secure program that looks, handles and works the same on Windows, Linux, or Mac OS – Firefox.