After Microsoft's on-again-off-again-maybe-not-done-yet bid for Yahoo or some part thereof, its acquisition of Powerset shows that it want to be a serious player in search. Better search results might drive users to Live Search, or at least keep them from leaving when the results aren't so hot. Yet there's still a lot more Microsoft needs to do.Powerset's claim to fame is the use of "semantic search" techniques. Put simply, semantic search improves search quality by techniques such as inferring meaning from sentence structure, using word synonyms, and determining the specific meaning of words from their context. A user searching for "trees" in a gardening context, for example, does not want to find an article about binary trees in a programming context. But how much do these techniques really improve search results?

I mention trees because Microsoft's announcement says, "Search engines don't understand today that 'shrub' and 'tree' are similar concepts." When I do a Google search for "transplanting large shrub," the results seem to cover both trees and shrubs pretty well. That doesn't have to be because Google applies a complex algorithm to find synonyms for "shrub," but simply because most substantial articles about transplanting large shrubs will also apply to trees. The first page has what appear to be several good hits.

Then there are Microsoft's search-related problems that have yet to be solved, even with better search technology. Somehow, Microsoft needs to reach and switch users who are comfy with their Google and Yahoo habits. The Live Search Cashback plan of paying users directly works with product purchases, where money is already changing hands. I wouldn't be surprised to see Microsoft come up with some sort of similar program for organic search, although the incentives might be harder to come by.

Somehow Microsoft has to monetize this search traffic, and that was one of the attractions of a Yahoo deal. Could Microsoft build its own search ad system? Yes, but it may involve a long and unprofitable learning curve. Don't forget that Yahoo's capabilities are not homegrown; it bought Overture to make the leap into search ads. I have to think that Microsoft is still on the prowl for help with this angle, but the pickings are slim.