Tuesday, May 31, 2005

Search: The Rise of Specialization

Adam Smith once said that specialization is a function of market size.

It is no surprise then that the search market, as it grows, is seeing the emergence of specialized, vertically-oriented search algorithms and companies. Searching for good answers wrt the types of search companies starting today and how an investor might want to participate in Adam Smith's prophesy of specialization, I attended today's Under the Radar conference in Mountain View. Under the Radar is a showcase for emerging search and mobile application companies and is targeted at the press and venture investors.

The common abstraction across all the presenting search companies is that while generalized, web-searching is useful, certain classes of queries demand a move from page-link analysis to algorithms more germane to the query at hand. Categories of specialized search include
  • Mobile search (optimized for handset form-factors, location-based services, and carriers): Medio Systems
  • Product search (optimized for product reviews and comparison shopping): Fatlens and Become
  • Rich media search (optimized for video and audio): Blinkx, GoFish, and Meevee
  • Travel: Kayak
Wrt product search, an example helps to make the point. If you search on Google for "Red Sox," the first listing is, appropriately, redsox.com. If you search on FatLens for Red Sox, you get a listing of available tickets to buy for Redsox games. Fatlens is an example of a company optimized to solve a problem ( help find tickets to events), rather than generic information about a category/topic (help me find the Red Sox home page).

The elephant in the room is obviously, what about GOOG, YHOO, and MSN? The major giants are investing heavily in search and extensions to web search. Will consumers seek results optimized for specific queries - travel, rich media, product shopping, event tickets, etc? Will the very models of analysis that make GOOG so powerful on the web (link analysis) leave room for specialists to enter with algorithms with limited general value but strong, vertically specific search results? Will the search start-ups be destination sites or will they syndicate their specialized search results to aggregators of queries (today's big three)?

A further observation is that the business models of various search companies are coalescing around a few key variables:
  • Advertising (Ex. Google Adsense as a partner, direct sales to advertisers)
  • Lead generation or referrals (revenue per click, revenue per customer acquisition, etc)
How does a search site generate revenue? A breakdown of revenue per search helps to understand.

At a high level, Revenue Per Search = Coverage x Click-through rate x Price Per Click


  • Coverage (#searches that show ads/total # searches)
  • Click-through rate (total # clicks on ads/# searches that show ads)
  • Price per Click (total amount received from advertisers/total # of clicks)
  • or total amount received from advertisers/total # searches
The specialized search vendors' business models presume that specialization will lead to higher value to consumers (greater relevancy) and hence more qualified leads to advertisers. Sounds reasonable.

WRT verticalization, a few thoughts:
1. Future revenue f(revenue per search or "RPS")

2. RPS f(domain specific problem resolution)

3. Domain specific problem resolution f(verticalization and specialization)

4. Verticalization f(investment in domain specific search algorithms)

5. However, venture returns are a f(scale and market size)

6. Focus limits scale

7. Therefore, future scale must be f(indirect sales and extensible product platform)

Accordingly, I buy that all queries are not created equal. A search for Red sox tickets, Red Sox jerseys, Red Sox highlight clips, and Red Sox pitchers may all require search engines optimized for the nature of each request. I do not buy, however, that consumers will want to visit N number of search companies to answer N number of queries. Nor do I believe that it is capital efficient to raise money to compete to become a destination site.

Rather, I believe the vertical search space will only see scale and success to the extent that major content or web properties see sufficient value in specialized search that they syndicate results from specialist search engines and play the role of aggregator of traffic and integrator of best of breed search function.

The key bet for a vertical search vendor today, in my mind, is that will engineering driven cultures, like GOOG and MSFT, overcome their not-invented-here bias and grant that specialized vendors produce better results than they are capable of producing through incremental hiring of specialists and tweaks to their search engines. If the answer is no, then I think the specialists lose. If the answer is yes, then MSFT and GOOG may morph from technology innovation companies (the best search innovators) to integrators of best of breed function and ad networks that drive queries across underlying and optimally suited third-party search engines. As a user/consumer, I clearly believe we will "win" if GOOG, MSFT, and YHOO aggregate specliazed search on our behalf rather than force us to alt-tab across N search engines to get us our much needed N answers!

GOOG's market cap and traction is clearly heralding in a new golden era of search innovation - will it pay off for venture investors?

Thoughts?

17 comments:

  1. Anonymous12:32 AM

    Sounds like you should attend Vertical LEAP where we'll most likely talk about that very topic.

    Re the economics, the ability to target more precisely classes of users certainly attracts much higher CPMs - provided that page views and uniques have reached a "meaningful" scale. And meaningful depends obviously of the vertical.

    As to the question of payoff for VCs, I think that it is going to be challenging. Large players will have to make a buy/build decision reasonably early on promising verticals. Early take-outs will be rewarding for the early stage guys or angels, but not VCs.
    The opportunity might be in local "copycats", like Kelkoo in Europe.

    ReplyDelete
  2. Anonymous7:27 AM

    I like the blog and too was in the audience at Under the Radar. You definitely capture the risk/reward.

    My concern with these is that:
    (a)distribution and customer acquisition is paramount, and none of these guys made it clear how they were going to pull that off. Building a destination is an expensive and time consuming effort, not likely to end up winning in this day and age
    (b) From the 5 minute pitch, it’s not clear to me how they improve the quality of search results so that the relevance/accuracy is that much better than the generalized enginge. With a few iterations of a search, you can probably get to the content that you are looking for on one of the ‘generalized’ search engines - in fact, some of these portals are doing just that - whether through Froogle or Yahoo Shopping;

    The allure for 1st gen shopping search engines was being able to index and retrieve semi-structured and structured product info and lock up distribution via partnerships, because what they did was actually much more difficult to do than text based web search, and then monetize from merchants wanting more focused leads - not sure how these guys are that different technologically at the end of the day.
    I think that these guys get to a place where they get some traffic, and also get to some revenues, probably $5-10M annualized, and then flatten out - they may even be profitable. I’m willing to bet that many VCs fund many of these, hoping that one gets the right formula of according to your blog.
    I’d be looking for signs of rapid organic growth. It is is many respects the true data point about the utility of a consumer offering, which will also tell you how high margin the business can become. The revenue side is well understood now, in my humble opinion.

    ReplyDelete
  3. btw, the Vertical Leap link should go to www.sdforum.org/vertical. the event is being held by SDForum on Tue 6/28 in Silicon Valley. the 1-day conference will cover a variety f topics in vertical search, including local search, shopping, travel, classifieds/jobs, news/blogs, etc.

    ReplyDelete
  4. Anonymous1:12 PM

    Paul Kedrosky wrote an excellent article on combining vertical search with Greasemonkey:

    http://paul.kedrosky.com/archives/001257.html

    ReplyDelete
  5. Anonymous11:03 PM

    Great post, great discussion.

    briefly,

    The general search engines already aggregate specialized search engines in their main search results, particularly for category-level keywords. The more useful to the tool, the more links they attract, the higher they rank, and so on. Search for "travel" on Yahoo or Google and you will find no shortage of vertical search tools. I get where you are coming from re: deeper integration, but I had to make this point, because it is not like we are starting from scratch to find these tools.. they bubble up.

    Back on the topic of integration, since you mentioned the Big 3, you might want to check Ask Jeeves. They show advanced search forms for many verticals. Example, a search on "travel" gives an advanced search form for Orbitz. But can they take the next step and execute a travel search? In this case, no. They could do more with simply-structured keyword searches, but the point is there will be many types of vertical search tools, and many types of aggregation, this will not be a case of one-size-fits-all.

    ReplyDelete
  6. Anonymous5:04 PM

    What If This Could All Happen Automatically,
    with a simple push of a button.....

    ReplyDelete
  7. Anonymous12:08 PM

    Among these magnificent benefits is the chance to be able to promote your business, any business,

    ReplyDelete
  8. Anonymous2:40 PM

    Will Price, interesting post - this post.

    Personally, I've always found offshore platform foundation to be very intriguing, and interesting, but not always very informative. So I did some research and started my own webpage to help people get started with offshore platform foundation.

    If offshore platform foundation also piques your interest, come by when you've got a few seconds - check out the different resources listed. I'm sure you'll find it very interesting.

    ReplyDelete
  9. Anonymous4:18 AM

    Will Price, interesting post - this post.

    Personally, I've always found offshore hedge funds to be very intriguing, and interesting, but not always very informative. So I did some research and started my own webpage to help people get started with offshore hedge funds.

    If offshore hedge funds also piques your interest, come by when you've got a few seconds - check out the different resources listed. I'm sure you'll find it very interesting.

    ReplyDelete
  10. Anonymous12:52 AM

    And A Link Back To Your Web Site Excite You?

    ReplyDelete
  11. Anonymous4:47 AM

    Will Price, interesting post - this post.

    Personally, I've always found w t offshore to be very intriguing, and interesting, but not always very informative. So I did some research and started my own webpage to help people get started with w t offshore.

    If w t offshore also piques your interest, come by when you've got a few seconds - check out the different resources listed. I'm sure you'll find it very interesting.

    ReplyDelete
  12. Anonymous7:03 PM

    Bueno!! I was doing some research about juice.net monavie and came across your blog. Was glad to see you writing about juice.net monavie because I have been working on a site to help people who sell juice.net monavie!! Gracias!

    ReplyDelete
  13. Anonymous6:06 AM

    Will Price, interesting post - this post.

    Personally, I've always found offshore platform foundation to be very intriguing, and interesting, but not always very informative. So I did some research and started my own webpage to help people get started with offshore platform foundation.

    If offshore platform foundation also piques your interest, come by when you've got a few seconds - check out the different resources listed. I'm sure you'll find it very interesting.

    ReplyDelete
  14. Anonymous3:47 AM

    Will Price, interesting post - this post.

    Personally, I've always found to be very intriguing, and interesting, but not always very informative. So I did some research and started my own webpage to help people get started with .

    If also piques your interest, come by when you've got a few seconds - check out the different resources listed. I'm sure you'll find it very interesting.

    ReplyDelete
  15. Anonymous5:20 AM

    Will Price, interesting post - this post.

    Personally, I've always found to be very intriguing, and interesting, but not always very informative. So I did some research and started my own webpage to help people get started with .

    If also piques your interest, come by when you've got a few seconds - check out the different resources listed. I'm sure you'll find it very interesting.

    ReplyDelete
  16. Anonymous11:17 AM

    This comment has been removed by a blog administrator.

    ReplyDelete
  17. Anonymous6:34 AM

    This comment has been removed by a blog administrator.

    ReplyDelete