I wanted to give you an update on underscores versus dashes in URLs. This is something that a lot of people have asked me about. And I had talked about it a long time ago. And so I figured it was time for an update. So first, let me give a little bit of history about why, whenever we see an underscore, we join that in the URL rather than separate using that. So what I mean? Well, if you say red dash widget in a URL, we view that dash as a separator. So we index the word red, and we index the word widget. And those are separate. Whereas if you were to have War of 1812 with underscores, so, war of 1812, instead of separating on the underscores we actually glom all those together. So that’s one term that you could find by searching for war underscore of underscore 1812. Seems kind of weird. So why does Google do it that way?
Well, whenever we started, AltaVista was huge. We were just this little tiny company. And we were all very techie. Lots of computer programmers. And we wanted to find exactly what we wanted as far as terms. We really cared about precision. And so whenever you are a programmer, you often have things like, if you’re a C programmer, you might recognize TMP underscore MAX. And so, if you are a programmer, you want to be able to search for that term and find TMP underscore MAX, and that exact term. Not just TMP and MAX that happen to be on the page. So it was because the original engineers were programmers, and the programmers wanted to be able to search for programming terms, that we joined based on the underscore rather than having that act as a separator.
It’s kind of what we call a second order effect. It’s not a primary thing that really makes a huge difference. For example, Wikipedia has a lot of pages that say war underscore of underscore 1812. That doesn’t keep Wikipedia from ranking. Because there’s page rank, there’s proximity, there’s title. There’s all the other signals that we use, over 200 of them. But if you are going to make a site and you’re starting fresh, so you’ve got a blank slate to work with, I would probably go ahead and go with dashes. And I would continue to go with dashes at least for the foreseeable future.
We had thought about doing a little project to split on underscores a few years ago. But it turns out the amount of impact it has in our rankings is relatively low. And it turns out, to get engineers to do that versus some other projects, there were other higher impact projects that we could have them work on. So at least for the time being, we still join on the underscore and separate on the dash. So a few people had asked, you were thinking about splitting on the underscore, do you do that yet? The answer is no. I don’t know when we will. Nobody is slated to be working on that.
Now if you already have a website, if it already uses underscores, and if it already works the way you want, don’t go back and rewrite every single URL. I would only bother when it’s a brand new website, when you’re really working on something fresh. When you’re trying to say to yourself, OK, I can do this anyway I want, then that’s a pretty good time to go for dashes. If you’ve already made the choice and you happen to use underscores, I really wouldn’t worry about it that much. It’s not a huge factor. But I just wanted to explain a little bit about the kinds of reasons why we would do that in the first place and just give a little bit of context and a little bit of update. So I hope that explains things a little bit better. Thanks everybody.
by Matt Cutts - Google's Head of Search Quality Team