From Service to Application

Late last week – and it was certainly an odd week for all sorts of reasons – I had the honor of appearing before the SDForum's Search SIG in Mountain View, on the Microsoft Valley campus. First I was interviewed by Dan Farber about the book (here's Dan's write…

Late last week – and it was certainly an odd week for all sorts of reasons – I had the honor of appearing before the SDForum’s Search SIG in Mountain View, on the Microsoft Valley campus. First I was interviewed by Dan Farber about the book (here’s Dan’s write up), then I got to interview four entrepreneurs in the search business – folks from Trulia (real estate search), Truveo (video search), Healthline (medical) and Simply Hired (jobs). Om has more on that here (though I have a rant in me about the “exit” – more later).

As usual, Dan focused his write up on what proved to be, for me anyway, the most interesting comment of the night. It came from Simply Hired CEO, Gautam Godhwani, when I asked him if he feared Google. “Google does search very well, but we have yet to see Google do applications well,” was his reply.

Interesting. As I thought about that, it struck me that what we are seeing right now is indeed the evolution of search companies from their roots providing a single service – one thing, done well – to a application suite that does many things. What does that mean, exactly?

Well, in Google’s case anyway, let’s give credit where credit is due. Google does do a few applications pretty well. Gmail, for example, is still considered by most to be a very good mail application. Blogger, while not for pros, is also a pretty successful application (“AOL for blogs” is how one pundit put it at Web 2.0).

But neither of those are really executions of search as an application – even though Gmail has really good search. Huhm. What Godhwani was saying is that in the search field, applications are the next thing, and Google is just as new at this game as his company – if not more so in certain vertical fields.

Dan further quotes him: “Finding a job takes a few weeks or months, doing research and using the power of referrals. You can’t do it on a basic search engine, so we are complementary to search.” As Godhwani said this, I was thinking to myself – “Well, as soon as there is an economic reason to do it, Google will do it, and then what?”

By then, I sense, Simply Hired (and all the other vertical search engines on the panel) hope to be so far along that the only logical move would be an acquisition, or direct competition in which the upstart actually has a chance of winning. It’s how it’s been for ages in this industry.

But back to this larger idea of search becoming an application. It’s probably obvious to you, but for some reason this idea provides me with a way of grokking a much larger trend – why is it that Google is so focused on Toolbar, Desktop Search, Accelerator, Local, and Ajax-y things like Maps, etc.? It’s because to create a decent search application, you need to have a far more robust interface, and you need to know far more about the intent of the person that is using your application. A web-based service, on the other hand, does one thing well, and does it the same for everyone. Search is becoming an application, indeed, and that more than anything else explains very well Google’s recent moves.

8 thoughts on “From Service to Application”

  1. John –

    thanks for speaking at the Search SIG & along with Dan being our emcee for the evening… definitely the best SIG meeting we’ve had so far, imho!

    with regard to vertical search & vertical applications, i think you (and dan) hit the nail right on the head. while we certainly don’t underestimate Google (or Yahoo or Microsoft) and their potential in any vertical should they choose to enter, each vertical domain requires domain-specific expertise & focus to do a good job. in particular, to address the issues of a vertical community you need structured data & metadata relevant to that domain, and on top of that the vertical applications that solve problems for that domain.

    While the majors have advantages in broad-based search technology, they don’t necessarily have advantages in domain metadata & applications. And while they certainly could choose to focus on an area & catchup, they probably can’t do that in 10-20 different verticals very well. They probably *will* choose to enter a few strategically (local, shopping, audio/video), and then enter a number of others with more limited initiatives.

    However it should also be noted that Google & Yahoo (& Microsoft in the near future) are also PARTNERS with all of the vertical search engines — they sell traffic “wholesale” to vSearch portals, who in turn develop the user profile further, address user needs thru vertical search & apps, resulting in a higher-value lead which they can then sell “retail” to other companies & play the arbitrage.

    Why enter a vertical & compete with all of your channel customers, perhaps only to take 20-35% of the market, when you can be a wholesaler to all of the players in the market? Again, this is the Microsoft playbook all over, re-tooled for search — doesn’t make sense for platform players to dive into a market with their channel & create channel conflict unless they’re sure they can own 50% or more of it. Much easier to be a wholesaler to everyone in the vertical, and get their cut that way.

    In any case we think there’s a lot of market opportunity ahead in vertical search for both major players and startups. We’re not selfish — we’d be happy to split the market 60-40, and still leave a piece of the pie for the big guys 🙂

    – dave mcclure

  2. michael: actually, Microsoft *is* the one company of the big 3 GYM portals that does applications pretty well.

    but in general, even Microsoft in the 80’s & 90’s didn’t compete at the “vertical” level — they built a very broad class of HORIZONTAL desktop productivity apps (Office) & NT Server Back Office apps (SQL Server, Exchange), then left most of the vertical apps (healthcare, insurance, financials, manufacturing, etc) for their partners to build out on top of the Windows/Office, & NT/BackOffice platforms.

    in fact, during the 90’s Microsoft built a fairly big & friendly ecosystem for many downstream channel partners by making their OS platforms great APPLICATION PLATFORMS for other folks to leverage. the small consulting company i started in ’94 was a Microsoft Solution Provider in this mold, and we basically built our business by developing Internet, Intranet, & E-Commerce solutions on top of the Windows NT platform. many other larger ISVs created multi-million dollar businesses on the Microsoft desktop & server platforms.

    only recently has Microsoft finally started buying / competing in some of the more mature channels (such as financials & ERP systems), but their standard game plan was always to work with [multiple] vertical market partners to develop various apps on top of Microsoft platforms.

    what is interesting is that here in 2005, Google can effectively do this whole strategy all over again with Google Search (& perhaps GoogleBase) as a *hosted* platform for others to build on. this certainly appears to be the strategy for things like GoogleMaps and Gmail, although they have yet to announce any formal plan or strategy to address the various vertical markets. curious that they haven’t clarified this better, because they certainly could step on the gas more in that space. if they wait too long, Microsoft will figure out & could again use their apps expertise to do a better parter platform strategy.

    perhaps the best example of someone combining a Web 2.0 services offering with a great platform partner evangelism program these days is Yahoo, with their Yahoo! Developer program:

    folks like Jeremy Zawodny, Jeffrey McManus, Russell Beattie & others @ Yahoo are doing a great job taking Microsoft’s old traditional developer program evangelism playbook and re-energizing it for today’s Web 2.0 Internet services. (Jeff Barr @ Amazon and my former colleagues at Ebay / PayPal also do a decent job on evangelizing more specific developer platforms for product catalog & payment solutions).

    all 3 companies have opportunities, but i would say if they’re smart they’ll focus on building a hosted platform with a rich set of Web 2.0 services for other startups to complement, rather than competing with potential partners by trying to take on every vertical.

    – dave mcclure

    ps – for more on vertical search & vertical applications, check out “Top 10 Rules for Vertical Revolutionaries” from the June 2005 SDForum conference on Vertical Search:

  3. “Gmail has really good search.” Do you believe this for real or was this an assumption based on how they advertise it. I can’t find a bloody thing unless I spell it dead on and none of my email subscription lists contain the keyword I’m looking for (i.e. never). Spell something wrong – forget it. Got a few seconds to back that comment up relative to any other email search?

  4. Google is currently building what I personally call “diagonal search”: ie a combination of horizontal and vertical searches (Gnews, upcoming Gase, ravel icons already given here on this blog, etc…)

    It means that when you enter on Google via the regular horizontal search, it will then try to attract you on its vertical services where monetizazion capabilities are far more juicy.

    so, horizontal+vertical -> diagonal search.

    I think that it is the current strategy of Google

    For those reading French, all details on my blog at

  5. To Dave McClure’s extensive comment above:

    Great summary of the Microsoft strategy, now being executed by Google.

    I totally agree that Microsoft didn’t extinguish the various vertical markets. They only focused on the ones that they thought would have the broadest appeal (read market and money), and went after it.

    A lot of verticals were safe and ok, as long as
    1. their markets were huge enough to attract microsoft’s interest.
    2. or microsoft needed them as customers/evangelists/resellers for their broader product offerings.

    To paraphase Esther Dyson, vertical competition by the biggies doesn’t destroy the growth in those markets, it just erodes it.


  6. have you played around with the mobile map client? it is a very sweet java application that allows for the brunt of the work to be done locally, saving the puny bandwidth for the map data only.

    i think google does apps well.

Leave a Reply

Your email address will not be published. Required fields are marked *