I attended a session today about Search Behavior where Inquiro, and Peformics presented their latest studies on how people are using the search engines and their affect on site traffic and customer buying behavior. (Keynote also presented a study, but I found very little of it to be relevant). Here are some of the very interesting takeaways:

Inquiro’s Study

• Contrary to their prior beliefs about the importance of titles and descriptions, Inqurio’s study found that the only statistical relevant predictor of driving traffic volume from Google searches is your rank and page position.
  
• Inquiro’s study employed some very interesting eye tracking technology from eyetools which let them see exactly where people were looking on the page, for how long, and where there eyes went next. From this they have defined what they call “Searches Golden Triangle”. This is the top left portion of the screen where customers quickly scan from the top down for relevant information, if their eye catches something relevant they will quickly scan to the right to verify. From a paid search perspective on Google, it highlights that the most important spots would be Google’s two top sponsored link’s that appear above the searches as opposed to the ones that show up on the right hand side. I will try and get a copy of their slide that shows this but in ascii characters it would like something like this:

  XXXXXXXXX
  XXXXXX
  XXXX
  XX
  X

• Inquiro’s study further showed that if searches didn’t find what they wanted in their first “F Scan” at the top of the page, 60% would scroll down and repeat in the organic searches, while 40% would start a vertical scan from the top of the sponsored links on the right hand side. Eye tracking tools showed that only the top 1-3 ads were relevant, with 50% going to the top, 30% to number two, and then quickly dropping down to less than 10% seeing the bottom 3 ads on a page.
  

Performics’ Study

• Performics’ study took a look at 30 e-commerce sites and then looked at cScoreData for people who purchased from those sites and relevant keyword searches they had done leading up to the initial purchase. They found that 50% of all buyers made a relevant search before their online purchase.
  
• There was an average of 4-5 searches prior to purchases.
  
• Their research showed that most searches started off as searches on generic terms, and most buyers never searched on a brand term.
  
• When buyers who searched on branded terms tended to search on these terms later in the purchase cycle, just prior to making an online purchase, which may be why some ROI tools show branded terms converting at higher rates than generic terms. A special note, that many of these buyers were introduced and “pre-sold” on a specific brand from the prior generic searches they had performed.
  
• 55% of buyers did their last relevant search more than 2 weeks before the purchase.
  

This was a very interesting session, and I had a few takeaways from this that I think are extremely relevant to companies doing paid search and search engine affiliate marketers:

1. For companies paying for their own search campaigns, it is important to take a full view of your ROI. If you are only looking at short term click conversions that you may be quickly out spent by competitors who are able to see long term click trends and conversions and offline buying behavior driven from search.
   
2. Affiliate marketers should review the cookie periods that their merchants are providing. Anything less than 30 days means that you are probably directly driving sales that you should be getting credit for.
   
3. Affiliate marketers may also want to stay away from companies that are only willing to let you bid on generic terms, because these generic terms may be driving a significant number of “pre-sales” behavior that is then turning into branded searches just prior to the customer making the online purchase. If the merchant is reserving these high conversion ads only for themselves, and overwriting affiliate cookies when customers click on their own paid search terms, then they may be taking credit for the sales you are playing a large role in generating for them.
   
4. Finally, with the relationship between generic searches early and branded searches later in the process, are your ROI tracking tools only looking at the last click, or can they show the full cycle of clicks?
   

   It’s all rather thought provoking and is worth thinking about and spurring further discussions.

   Adam