Personalization, User Data & Search

Aug 20, 2007 • 6:07 pm | comments (0) by twitter Google+ | Filed Under Search Engine Strategies 2007 San Jose
 

SES: Personalization, User Data & SearchChris Sherman is moderating this session. Chris explains that every year we talk about personalization but we never really see it in action. But that is now changing, Chris says.

Gordon Hotchkiss, President & CEO, Enquiro Search Solutions Inc. is up first.

What drives personalization? - User volunteered information - Search history - Web history - Community choices - Task context - Geo targeting - Platform information - Origin of search

Does Personalization Work? Organic listings 3, 4, and 5 were personalized. Study of 16 individuals tracked history from a previous sessions. Provided listings based both on past clickstream and assumed place in buying funnel. The eye studies reveals, the personalized resulted doubled the performance (click throughs). Personalization works.

Maybe personalization works too well? Danny Sullivan asked "the concern might be if the natural results are getting better and better why would someone want to click on the ads anyway?"

What he saw with eye studies that if there is strong relevancy queues in the middle of the page, people will focus there and pay more attention to middle of page versus top result. What happens if the scanning patterns change due to it?

When there is an image in the search results, you typically fixate on that image. Without it, there is a triangle pattern of looking. There are more results fighting for attention when images are on the page (universal results).

As personalization becomes more important, keywords become less important and the user becomes more important, says Gord.

Optimization is going to happen around themes. There will be more long tail optimization. Optimization will come across content buckets. Understanding user behavior will be vital. User centric development will take hold. Emerging "buzz sites" will be an SEO tactic.

Circle of Importance: - A handful of sites or specific content will emerge for each "theme" - These will be inundated with offers for RSS content, widgets and gadgets - Scarping fo content from Circles of Importance Sites

He then showed eye studies based on mock new UIs for search interfaces.

Richard Zwicky, Founder & CEO, Enquisite is up next.

Enquisite is a search analytics software, he was an SEO, so he built it out of a need. He explains what personalized results are.

He then shows statistics, proving that if you are not on page one, you are no where.

He explains that the search results are different based on where you search on, also the look of the results are different. So the results and items on the page are different. Different results are shown based on where geographically you search from. Plus now every engine has different results.

To make things more complex, users are different and they have language differences.

If you want to show up for these results, in the short term, go buy some search ads. In long term you can do more client management, more content needs to be created and more SEO is getting a lot more difficult. SEO campaigns need to be more sophisticated and you need better research tools and reports.

Dave Davies, CEO, Beanstalk Search Engine Positioning, Inc. is next up to discuss patents.

User Behavior: Engines can watch user behavior to assess the value of web pages including click through, time to return to engine and type of query. These will impact your future results as well as the results of similar searches.

Personal PageRank: Google will be assigning a value to you as a search user. This value will be used to grade your activities on web sites. This will affect the rankings of sites based on an individuals activities on that site. Users who are trusted by Google will affect the results to a greater degree than those who are not trusted.

Group Data: The engines are/will be grouping data. Membership in common communities, common bookmarks, common search behavior. This will affect your future results as well as the results of similar searchers.

Influencing Engines: The Problems: - Rankings can vary from region to region - Rankings can vary from person to person - Some factors will have little to do with typical SEO

Influencing Engines: The Benefits: - For the search, they will get better results - For the engines, they will provide better results

Limitations: - Currently personalized results only affect those signed into a service

Solutions to React to This: - Design for time on site and page views, monitor stats and tweak for phrases with low visitor experience levels - Localize - Link out - Attract high PageRank visitors, sheer numbers and think like a high PageRank searcher

Jonathan Mendez, Founder & Chief Strategy Officer, OTTO Digital is next up.

Importance: - Nothing performs better than targeting to the intentions of users - SEMs need to step up and get recognized - CMOs think BT is just display - Higher search budget, more brand dollars - Everyone wins this way

Get parameters passed to you, so you can use this data, such as: host language, safe search preference, client browser, results language, query, search interface and more see joostdevalk.nl/wp-content/uploads/2007/07/google-url-parameters.pdf

Ads can provide even more parameters such as ad groups, ad creative, match type, and local parameters.

Tools: - Offermatica - Omniture (touch clarity) - Kefta - WebTrends

You need to then do segmentation, and use very high impact segments because it provides a large data set. You can segment on behavior, and source and temporal and environmental.

Personalization and Targeting Roadmap - Use data to discover segments - Hypothesize - Create and develop - Test and validate - Monitor

He then showed a quick case study, showing how using some of this data helped increase a conversion rate of 47%.

Match Types: - Exact had a +2.63% conversion rate - Broad match has a 4.36% lift in conversion - Phrase match had a 15.16% life in conversion

Temporal patterns: Weekdays a page had a 1% lift and one page had a 8.5% lift on weekends, so they used different pages based on weekend vs. weekday.

He then showed a geo example. And then showed more examples of how this data can really help improve your metrics.

On QA include: Sepandar Kamvar, Google (Engineering Lead at Google) He said he is impressed to see there is a sophisticated knowledge about what they do. The biggest impact of personalization is to design and optimize for the user and not the keyword, from an SEMs perspective. There is an importance of building useful gadgets because personalization works best when your query is simplified (short queries), so extend that to zero word queries. They built iGoogle to look at user's data to see what they want to search for without them searching, as gadgets. Also thing about other ways to distribute your content. So optimize for user and build useful gadgets.

Tim Mayer, Director of Product Management, Yahoo! Inc. Yahoo is focused on using social data, people's preferences. Yahoo is attacking the "subjective space." What doctor or restaurant should I go to? What is cool furniture or clothing, etc? Those types of queries. Yahoo uses the discovery experience. Leverage the power of the people to find the coolest thing out there or what is new. Become the known expert and the authority on the web and within your community (delicious, answers, etc.). Then have something unique. Have a lot of fresh breaking content, promoted through Digg, delicious, etc. Promote yourself across a lot of different services (Digg, iGoogle, delicious, sphinn, etc.)

Previous story: Ad Exchanges are Changing Everything
 

Comments:

No comments.

blog comments powered by Disqus