What Google Can't Copy (Easily)
by Gabriel Weinberg, October 2009
Google just launched some new search options including "More shopping sites" and "Fewer shopping sites." We recently launched a very similar feature, i.e. "Mostly shopping sites" and "Mostly info sites."
I'm sure Google had their version coded before we even thought of this idea. And I'm not trying to overly flatter ourselves by saying that in this particular case they launched the feature because of us. I'm just re-pointing out an identified pattern to get at my main question: what features can Google not easily copy?
There was Search Wikia (now defunct) and Google's Search Wiki, Wolfram Alpha and Google's Public Data, Cuil's big index and Google's big index, etc.
Matt Cutts (of Google) took the time to carefully explain that these features were all in the launch process before the competitive threats emerged. I subscribe to a modified version of Danny Sullivan's theory, which goes something like this...
Google has tons of ideas already coded up and ready to deploy. Blog posts about them may even be in the works or complete. For core algorithm improvements, they just go ahead and push them out continuously. But why add actual new features when there is no immediate pressure to do so and possible drawbacks, e.g. user push back or confusion? As long as their market share climbs upwards, there is no compelling reason to do anything that might potentially impede it.
So they wait until either a) through bucket testing they are sure it is a good idea or b) a competitor adds a feature that users seem to like or the press makes a big deal about. They don't want any new search engine to really get a foothold so they need to react to these competitive feature launches lest a new search engine actually starts getting used :). The features are all ready to go anyway, so it really is only a matter of making the final decision to push them out the door.
I hate conspiracy theories, and I hope this doesn't come across as one. To me, this is the simplest explanation given all the data.
Which brings me back to the title of this post: what Google can't copy (easily). I find it hard to believe that anything is off limits to Google due to technical complexity. I suppose you could make some case around patent protection, but I'm skeptical even then.
Yet there are still a host of things that for various business reasons Google would have a hard time copying. First, they can't easily make drastic changes to their results pages. Yes, they can add a button here or there, but they're not going to go where Kosmix and SearchMe (now defunct) have gone.
And nor should they! However, this leaves room for startups to innovate in this area. Doing so and actually getting traction is of course non-trivial, but I'm convinced there are alternative interfaces that will appeal to non-negligible portions of the world population.
Similarly, Google can't easily make drastic changes to their home page. Sure, they can make the text box bigger, but it isn't going to become two text boxes or an intensely-visual flash interface any time soon. The farthest I've seen them go is Google Japan.
Again, this leaves room for UI innovation. You see this already with vertical search engines who often use multiple boxes/drop downs. Our shopping/info site feature also falls in this area, as it is implemented mainly from our home page.
Third, Google is reluctant to hard code search rankings. They have a number of a good reasons including scalability, censorship and anti-trust. Startups don't have these issues, at least not to the same extent. Mahalo has focused on this "fat-tail," and so have we.
We've also been way more aggressive at removing what we perceive as useless sites form search results. Personally I think this goes far to make better search results. Yet it clearly isn't a short-term win because users don't immediately notice the difference.
Google is also reluctant to develop its own content and promote it above other content for pretty much the same reasons. There is a fine line here. They do answer some questions, cross-promote their services, and do universal search.
But when it came to Google Knol they very carefully pointed out that Knol content would have to compete with other sites on a level playing field with regards to rankings, whatever that means. There exists opportunities for startups here too, which we have certainly focused on via our Zero-click info boxes, category and meanings pages.
Fifth, there is ads. A lean startup can get away with much less advertising on search results pages. How much this matters to users or not is another question for which I don't have a good answer. As far as I can tell, it doesn't seem to matter that much. But Google is nevertheless and understandably sensitive to this area, which I think is why, to many Adwords advertisers' dismay, they don't show ads for a vast majority of searches. Yet when money is really on the line, the ads are all over the place.
Can you think of any other feature areas that Google would have a hard time copying for business reasons or otherwise?
A lot of good comments are also at this HN thread.