Ask Google to choose your next domain

Last May, I packed up my tent, sleeping bag, and bike into my Jeep and drove down to San Luis Obispo to spend the weekend racing my first half-Ironman. Although there were lots of people with tents like me, most of the veterans drove down in campervans and RVs. After managing about 4 hours of sleep in my soggy tent, I was feeling pretty jealous.

As soon as I got back to San Francisco, I went hunting for an RV rental search site (something like AirBnB or Kayak) but I couldn't find one.

So decided to make one.

After 6 months of sporadic weekend work I’m nearly done; 414 locations and 2,018 RV models and counting.

But what to name it?

At first I was set on naming it Recreationist, but the more I thought about it, the more irrational it seemed to choose a domain based on my own preferences. While I had a good feeling about Recreationist, I’m fundamentally skeptical of my own and everyone else’s intuition; at 42Floors we even test designs that we hate just to challenge our assumptions.

A split test was needed.

The test

I spent a few hours on Domainsbot and coming up with a variety of names — short, silly, boring, startup-y, long, exact match, etc. I whittled that initial list down to 17 and created identical AdWords ad for each domain.

The ads started on January 2 and took 6 days to hit 98% statistical confidence. The winner, by a landslide, was

It had a click through rate of 4.63%, which was 53% better than the runner-up, Those extra clicks also translated into a sizeable discount on AdWords: $0.20 per click versus $0.26 to $0.44 for the other domains.

To run the test I spent $170 for the domain names and $293.04 for the ads, which included rerunning a subset of the tests to double-check the final outcome. Thus, for the bargain price of $473.04, I got to confidently pick a domain that outperformed my personal favorite by 127%.


There are two common objections to this approach:

  1. If the domains you want to test are expensive (already registered), it would be absurdly expensive to run the experiment.
  2. The optimal domain from a clickthrough standpoint may not be the best brand.

For the first problem, the solution is simple: just buy knockoff versions and run the test on those before investing in the winning domain. For example, if you want to see how will compare to, just test versus The .co domains have a performance penalty versus their .com brethren but the differences should be proportional.

As for the second problem, brandability, this approach will at least give you some cold numbers so that you can make statements like, "I'm willing to accept a 22% lower clickthrough rate because I believe we will more than make up that gap from repeat visitors who would otherwise forget our name."

What's next

I still have several hundred RV rentals to add and I haven't even done the bare minimum of SEO on the site, half my avatar photos are of toilets, and the site's slow as hell when the cache misses, but people are finding and using it anyway. It's kind of exciting.

Actually, speaking of SEO... "Check out the next time you need an RV rental."