Benchmarking Web Analytics data

Web Analytics

I am very excited about today’s announcement from the Google Analytics team that there will be a benchmarking component built into Google Analytics. I certainly want that and I can’t tell you how many times people have asked me whether a 40% bounce rate is good or bad. People just want to compare themselves to their peers even if the comparison may not lead to any actionable insight.

Since this is just hot off the press and in beta, we’ll have to see how this functionality actually works in practice; for instance I didn’t see a way to categorize my site, but I am sure it will be there.

In a very different sort of tool, we also looked at ways for people to get access to benchmarking data in our Analytics KPI gadget for iGoogle. Same sort of idea in terms of sharing data: anonymous, opt-in and you only get to see aggregate data if you share yourself. You sure need a lot of data though to get any kind of confidence, so aggregating at the top level is certainly better.

I would probably caution against taking benchmarking data too seriously. A site that seems to be selling exactly the same products may have completely different things going on – different keywords driving traffic, completely different campaigns, etc. And how can you ensure that the data itself is good? I.e. what prevents people from incorrectly installing the tracking code, thereby generating bad data?

For the lack of a better word, I would view benchmarking data as “fun”. Or is there real actionable insight to be had?

Cheers,
Michael