This week’s interview is with Grad Conn of Microsoft. We spoke about Microsoft’s unique new health search product, and how they have integrated that into the core experience of Live Search. Read the interview for the details of how and why Microsoft put it together, and comment below if you want to discuss it.
This past Wednesday my latest “By The Numbers” column went up on Search Engine Watch. This article reviews the Custom Search Business Edition, for those webmasters that need to do a bit more with their Custom Search Engine than the free version allows. One of the nicest things about CSBEs is that you don’t need to have any ads presented together with the results.
A while back I wrote about how building a search engine using Yahoo! Search Builder can increase the number of pages you have in the Yahoo! index. This notion did not come from me. It came from an interview I did with Tim Mayer back in December 2006.
Recently, we began a little experiment with this notion to see if we could see any impact from doing this. It’s still too early to have data from this test, but we have already noticed one very interesting thing. Take a look at this image:
Notice the source column on the right? On our domain, www.stonetemple.com, the data source is stated as being Search Builder. This does suggest that something has been tweaked regarding the indexing method. While the Stone Temple site is not the one that we are running our Search Builder test on, the above screen shot should provide some additional data regarding the potential for the Search Builder product to influence indexing.
Recently we had the chance to speak with Tim Mayer about Yahoo’s Search Builder product. Search Builder is the vertical search engine offering from Yahoo, that competes with products such as Google Custom Search Engines, Live Search Macros, and Eurekster.
In the discussion, Tim reveals that sites that use Search Builder on their site may well increase the number of pages on their site that get indexed by Yahoo. It amazes me that there has not been more play regarding this point. It seems to me that I see people complaining all the time about the number of pages that they have in search engine indexes.
So here is one quick answer: implement a Search Builder search engine on your site. I would love for someone to try this on a site that has a limited number of pages in the Yahoo index and see if it helps.
Somewhere during the days reading I came across this report on the Emerging Opportunity in Vertical Search. The report pulls stats and data from a report called Vertical Search Delivers What Big Search Engines Miss by Outsell, Inc.
What is fascination is that the Outsell’s data suggests that the B2B vertical search market will reach $1.0 billion in revenue by 2009. So it’s no wonder so many companies have launched vertical search platforms. Some of the most significant platforms, listed alphabetically, are:
- Eurekster Swickis
- Google Custom Search Engines
- Microsoft Live Search Macros
- Yahoo! Search Builder
All of these are vertical search platforms, that permit you to build your own vertical search engine. All of these vendors handle the crawling of the web, and the building of the index for you. You simply need to enter in lists of sites for your search engine, specify rules for its operation, and other related data through a web GUI of some sort.
Included in the Slack Barshinger report is some data from a Jupiter Research Survey done on search in 2004. The data shows that 41.2% of users in indicate that “Results are often not directly relevant to my query”. Frankly, I suspect that given a different question, such as “Some of the results in the top 10 results are not directly relevant to my query”, that the percentage would be far higher.
And this lines up with one of the conclusions, that there is an overall Internet search failure rate of 31.9%. What’s at the core of this problem?
Context. The search engines do not understand your context. In the Slack Barshinger report they provide the example of a dentist searching on ceramics. The result page shows pottery sites as the primary results. And even if the dentist refines their search, to include the word dentist, the results are not good.
This is because there are two context problems:
- The search engine does not know the searcher’s context
- The search engine cannot easily determine the intended context of web pages (i.e. is this page targeted at dentists of pottery buffs?)
Net-net, this is a BIG issue for search engines. They know that search has flaws, and they want to fix them. These problems are driving lots of innovation. There are hundreds of vertical search engine start ups, focused on a variety of narrow niche markets.
These companies perform their own crawls, build their own indexes, and even make use of non-crawlable data, such as purchased databases of information to provide more focused results. This space is going to be really big, and represents a big new frontier.