I’ve been involved with Local Search for the past 4 years. I spent the first 2 years at The Press of Atlantic City trying to figure out the best way to give our local users what they were looking for when they queried a term. We realized early on that the majority of our queries were for top news or archives. This had a lot to do with the lack of a marketing effort which I will touch on in future posts.

I left them to go work for Planet Discover, the company that provided our search technology. After looking at fifty or more sites with similar looking search results, I began to wonder if this was the right approach. I wasn’t questioning the technology because I believe what Planet Discover offers is superior to their competitors. I was questioning the strategy at the local level but I didn’t have a solid solution.

Last November, I had the opportunity to see a presentation by Jason Calacanis, the creator of Mahalo.com. His take on search was eye opening to me. His belief is that the majority of National search results are poor and that human input could fix the issue. Mahalo employees and users with expertise in key areas will massage the search results to provide the best possible result. He will admit that this is a huge undertaking and plans on focusing on the top 10% of searches. Mahalo isn’t shutting out Google, Yahoo and MSN. They are actually going to give you the opportunity to see the search results that Mahalo doesn’t have a good result for, on their sites.

Anyway, I digress…

The Mahalo concept works beautifully in the Local Search world where 95% of the search results are tied to local information. If the editorial staff and web staff worked closely together they could create the best local search result for queries on many terms like museum, little league baseball, taxes, and many more. Providing this level of search result to the user would make local SEM advertising more powerful and the newspapers website more relevant.

Again, the technology is good but the system needs to be fed better information to spit out the best possible result. Every search term shouldn’t pull up jobs, cars and homes. We have actually trained the user to go other places to get this information and the initial search result in the federated search is usually useless making the user click through anyway.

User testing may be a good first step to determining the value of your search results. This could be followed up by some solid search term analysis. Look at the top 10 terms at the end of each month and compare them to the search result. Is it the best possible result?

For the top 10 search terms you should determine what content buckets you would like to show in the federated result. Go for quality over quantity. Use the SEM ad products to promote other relevant local content such as a page on the local high school’s site or any content your users could find valuable.