SearchMash is an interesting new search engine. It turns out that it’s actually just a shell on top of Google’s search engine. It’s a pretty cool shell though. One thing you can do is to hit space repeatedly to see more and more results, without reloading. Ajax goodness. Another thing you get is image search results alongside with your normal results.

This article is about one of the least conspicuous features, which none the less might be the most important one.

You can rearrange results.

Doesn’t sound like much but lets think about it. What might rearranging the results be good for? Right now the features list says the rearranging is just for fun. This is coupled with a hint that there are ‘plans’ for the future.

Here’s the future.

Rearrangement is an intuitive way for search users to ‘fix’ their search results. For example, say that Jane searches for ‘Google’ but the top hit is ‘Microsoft.’ Microsoft has a big marketing department. Marketing or not, ‘Microsoft’ is not the result Jane was looking for. She scans down and finds ‘Google’ in the second result. Traditionally, this would be the end of the story. But with this technology, what Jane would do is to simply drag ‘Google’ to the top. She does this not because of some spiritual search for perfection. She does it because the next time she performs the same search, she wants the most relevant result on top. The search engine memorizes her rearrangement action and the next time Jane searches for Google, she gets the result she wants on top.

That’s nice and all but it’s only the tip of the iceberg.

If a large number of users rearrange their search results in the same manner, the search engine can take note. If a lot of people are placing ‘Google’ on top of ‘Microsoft’ when searching for Google, the search engine might eventually say, “Hey, that’s kind of funny. Maybe Google is actually the most relevant entry. In fact, I’m going to go ahead and make Google the top entry for everyone.”

There it is. The spam result just got downgraded, the search engine’s index became better, and every person searching with the engine in the future thinks the search engine is just terrific, not realizing the betterment was caused by humans rather than technology.

There is in fact nothing surprising about this development. Sites like digg have been touting swarm functionality as their thing for a long time. The philosophy is that when many users get together and do a small bit of work each, a lot of work gets done. Ultimately even huge tasks can be undertaken in this way.

The only thing that is surprising about user sorting, and Google’s secret testing of it with Searchmash, is that it took so long. To sort the whole internet by search terms is the quintessential swarm task. It can be massively parallelized – lots of people can sort their own little bit of the internet. Each step of sorting isn’t a lot of work. In this case the sorting is even rewarding for the user – they get personalized search results with their own sort order in the future. They don’t need to be altruistic. They just have to do their own little thing and in the end everyone benefits.

Swarm ranking is the next step up from Google’s page rank. The first internet search engine to feature page rank and swarm rank simultaneously will win this round by a significant margin.

If only our startup had a couple of million dollars.