October 20 2008
Build SEO plans on complete market research
The September 2008 Search Market Share shows that Microsoft continues to shave search visitors off of both Google and Yahoo! Of course, the search market share reports based on number of queries performed still hide what is going on because they fail to filter out the non-search queries that people run at Google. Rank-checking queries and informational queries constitute a significant percentage of Google’s traffic.
As I have noted in the past, even measuring search market share by estimated unique visitors has its drawbacks (a problem that has become so complex I have given up on measuring search market share by Compete’s data).
I feel a more precise measurement would follow the number of referrals that search engines actually send to other sites. But how should that kind of data be measured? You cannot trust your own server logs and analytics, especially if you have historically only tracked progress for Google. If you don’t optimize for Yahoo!, Ask, and Live, you have no reason to expect significant traffic from any of those search services.
Hitslink measures market share according to its aggregated data, from people embedding its analytics code on their sites. They claim to measure 160,000,000 visitors across their network. Now, while that is all well and good, what Hitslink cannot determine is who among their analytics partners is optimizing only for Google. If their analytics customer base favor Google then Hitslink’s data is biased and unreliable.
100,000,000 million people visited Live.com and/or search.msn.com in September 2008. Fewer than 136,000,000 people visited Google.com in the same period. To suggest that Microsoft serves only 4-5% of the search market is ridiculous. On the other hand, if Yahoo! processes more queries than Microsoft, then why does search.yahoo.com receive fewer than 60,000,000 monthly visitors?
The data we’re missing for a proper analysis of the search market can only be found in the search engines’ tracking data, as they all appear to be tracking click-through results now. External estimates of market referral data are all based on incomplete and non-representative samples.
It’s vitally important for SEO technicians to understand which search engines are being used and how much for each vertical. A great deal of wasted effort goes into optimizing for hypercompetitive expressions on Google. Comparable expressions may exist on Yahoo! and Microsoft, where few if any people may be optimizing correctly for those search engines.
Your typical SEO plan should include:
- Keyword research for Ask
- Competitive research for Ask
- Optimization for Ask
- Keyword research for Google
- Competitive research for Google
- Optimization for Google
- Keyword research for Microsoft
- Competitive research for Microsoft
- Optimization for Microsoft
- Keyword research for Yahoo!
- Competitive research for Yahoo!
- Optimization for Yahoo!
Each search engine serves its own audience and maintains its own database. Relying on any one search engine’s data to analyze and optimize for other search engines is uncompetitive and inefficient. Don’t assume that search patterns are the same across all search engines. Things just don’t work that way.
Written by Michael Martinez
