The keynote at SES Berlin this morning had the two major search engines speaking and then discussing the state of the search market. With Mike Grehan moderating the Q & A and two short presentations by the search engines we got a nice insight in the way Google and Bing think.
The biggest thing which came back from both talks was user intent. Both search engines are trying to figure out, each in their own way, what users really want.
Lars Lehne (Google)
First up was Lars Lehne from Google. He started by taking his kids as a metaphor. They keep on asking questions, wondering why dad doesn't know it all, he works for Google.Stefan Weitz (Bing)
He shows insideyoursearch.com to explain what his kids must think he's doing all day.
Some key elements of his presentation were:
- Search has moved marketing from push to pull
- The increase in measurable
- Gross rating point = Google Rating Point / Search Rating Point
- The more people search the more important relevance
competition is only one click away
- the information you want must be by hand. It has to be mobile
search is missing link between push and pull
He closes with a slight change to a Henry Ford statement: "a man who stops SEM to save money is like a man who stops his watch to stop time."
Microsoft Stefan Weitz got into search because he looked for the biggest challenge: he found search.
The end isn't typing in a keyword and getting results.
It's about enabling people to ask more questions. It should not be about how many links you get.
Weitz adresses 3 problems of search:
25% of queries result in quick click backs. This actually means the back button is very much used :)
the results are not precise enough
42% of sessions need refinement. People continue to refine the searches they make. That's an indication.
Almost 50% of time is spent searching is on these queries.
The average is half an hour pre query. This means very much time is spent searching.
He then adresses how Bing is built Their building Bing in 3 parts:
Multimedia, Index Quality, Speed, Relevance
Trying to compete in core search is tough so that is basic. You won't win on that.
Reduce Overload, Sessions not queries
Could be different visualization. Path through or refine through data sets. Break up results.
Pre Process Web Information, reduce anxiety
Its about understanding user intend
After the short presentations they started a discussion moderated by Mike Grehan. Some points which came forward from the Q&A:
- Weitz explains user intend is hard to find out but that they can get a lot from history. Gives the example of UPS, if you search on UPS in the US you won't get much results, you just get specific UPS information.
- Weit also says People expect searches to be correct. That makes it difficult to be good.
- Lars Lehne says sitelinks have dramaticly improved the CTR.
- Bing answers about Twitter: what do you do with all the Twitter data? They use Twitter to be more relevant.
- Google: Twitter is about learning from others about brands
- Grehan at the end asks the search engines two questions: where should we be concentrating in search?
- Google: don't give up on classical marketing. Many think search is transactional driven. The big companies are ignoring search but will change their minds in the next coming years.
- Bing: it's critical to model your pages on the most common tasks people will want to do. Bing looks at what does this site enable people to do?