My colleague Pete Blackshaw has been talking about a search optimization metric he calls Shelf-Space Analysis, which looks at where branded web content appears within search results for general terms related to a topic of interest. So I decided to run a little test of my own. I searched in Google for 10 different diseases — from more to less serious, from well-known and widely diagnosed to more rare. I looked only at natural search results, not at sponsored search links. I wanted to see whether the branded website of any treatment (or any pharmaceutical manufacturer) would show up in the top 10 search results for any of these disease categories.

The result? Not one branded site appeared in the top 10 search results for any of the disease categories I searched. (In case you’re wondering, the terms I searched on were: asthma, allergies, high cholesterol, high blood pressure, fibromyalgia, breast cancer, prostate cancer, insomnia, Crohn’s disease and alpha-1 antitrypsin.) The top search results invariably included links to Wikipedia, well-known healthcare information sites and government sources. The national disease organizations or foundations also have good representation, as does the Mayo Clinic.

Here are some numbers behind these findings:

I decided to take the search a step further and look through additional search results for two diseases to see how far in I would need to go to see a branded treatment/pharma site appear in the results. I chose two diseases that are very different: breast cancer because it is so widely recognized and so many companies have initiatives related to breast cancer research, treatment and support, and alpha-1 antitrypsin deficiency because it is much less commonly diagnosed and therefore should have fewer players vying for search placement.

• For breast cancer, the first branded search result I found was #47, a link to the Avon Breast Cancer Crusade. After that, #81 links to the Avon Walk for Breast Cancer. Aside from these two, there were no other branded links among the first 100 search results.
• For alpha-1 antitrypsin deficiency, results #22 and #40 go to the Mercksource.com resource library. #67 goes to the branded web page for Prolastin, one of the primary treatments marketed for this condition. There are no other pharma/medication-branded links in the top 100 search results - including sites for two other main prescription treatments for the condition.

All I can say is, wow. I anticipated that search optimization might not be maximized for pharma-branded websites, but I am honestly surprised by the extent of these results. Given the value of search for finding healthcare information, based on our recent survey and healthcare webinar — one-third of survey participants told us that an Internet search engine is their first source for finding healthcare information online — an updated search strategy may be just what the doctor ordered. (Pardon the pun.)

Tags: disease, pharma, search

I just spent a lovely two weeks in the northern wilds with my family.  No Internet and, aghast, no cell phone.  We did spend a bit of time in the airport however, which led me to some thinking about the direction of the Internet and the importance of the current incarnation of search.  But first a story:

My wife and I were sitting the middle of an AirTran flight and I was looking up at first class (yes, they do have 1^st class in AirTran flights - news to me). I saw a distinguished looking older gentleman whom I recognized, but didn’t know where from.  As soon as we got off the flight in Burlington, VT the gentleman was talking to a news crew.  Still, I recognized him, but had no idea from where.  My first thought was that when I got home I would Google him.  But alas, there is no way to Google “older guy I recognized on flight to Burlington” and have it be remotely meaningful.  Since I appear to have outsourced my ability to remember things to the Internet, I never got my answer.

This is because, as we all know, Google, and most search for that matter, looks for words.  We’ve found work-arounds with tag clouds and other forms of meta-data, but no real answer to the root problem.

Why should we care now? Well total time spent watching Internet video us up more than 30% over 6 months according to VideoCensus, versus, basically flat overall Internet time.  The image based (non-text) Internet is growing.  Given that we are still lacking (good/real/useful) meta-data standards for video we have a problem.

We can’t search it.

I think this might be a problem.

Tags: internet, search, video

Here’s some data that provides an interesting perspective on the Google/Yahoo deal announced last week.

While this deal does seem to be a pretty clear win for both parties, the audience overlap data below suggests the possibility that we should be wary of the most optimistic incremental revenue scenarios. Given the fact that 77% of Yahoo searchers also search on Google, it is possible that there may be a material number of instances in which a person searches on Yahoo and does not click on a sponsored link, and then searches on Google and clicks on a sponsored link.

With the Yahoo/Google deal in place, this person would not need to execute the second search on Google. This would simplify the experience of the searcher, but would not generate incremental revenue for Google. Yahoo, however, would get a cut of search revenue that it would have otherwise lost. The scope of this behavior in the future may ultimately be what determines whether Yahoo or Google got the better end of the bargain.

Tags: google, search, yahoo