How do you A/B test SEO related changes?

[chal00d]

In many enterprise-level “big site” organizations the culture leans heavily into A/B testing the product, which tends to concentrate on measuring conversion gains. As an in-house SEO it can be challenging to defend site changes in the context of this product conversion testing environment.

While there are plenty of articles that cover the SEO effects and pitfalls of A/B testing, it’s harder to uncover material that focuses on A/B testing SEO itself.

In light of this I’d be very interested to get some comments & feedback from some of the experienced SEO's/webmaster's here;

Are there any easy ways to A/B test SEO changes to large websites?

Which metrics are the best performance indicators for version A versus version B?

Can existing, conversion-focused, A/B testing methodologies be tweaked for SEO test scenarios?

chal00d

[Pjman]

Q. Which metrics are the best performance indicators for version A versus version B?

A. 1 and only 1 indicator to worry about long term revenue. How does this tweak affect my bottom line, long term.


Q. Can existing, conversion-focused, A/B testing methodologies be tweaked for SEO test scenarios?

A. Yes, but only test on a small scale. When I test site-wide changes my rankings take a hit for a few months. I focus my site wide testing during my slow months and make no A/B testing during my hot months.

[rish3]

Not easy to A/B test when

- there's only one "customer" that matters
- you don't control the important variables. sure, you can change signals, but they can change the algorithm

:)