OK - a title likely to provoke knee-jerk reactions from all and sundry, but hold on a moment before cranking up those old cloaking arguments.
I think this is a whole different question.
For those that don't know, Google Optimizer is multivariate testing tool that allows you to split run test different versions of key parts of your page in order to improve conversion rates.
Its in beta at the moment, but using it on a number of client sites has shown it to be as effective as some of the high ticket price alternatives.
I like it.
..in fact, I like it a lot.
So whats the cloaking issue then?
This is a nice way of getting round the problem from which most other multivariate testing software suffer, the problem of showing many different page versions to the SE spiders and incurring a penalty.
IMO, its a sensible solution to the problem and I have no issue with its implementation.
It does, however prompt a few interesting questions:
- This is obviously presenting one version of the content to spiders, and another version to the visitors. Clearly the intent is honourable, so is this specific evidence that Google are prepared to accept this technique as valid providing the intent is not to deceive?
- Are Google prepared to accept this technique as valid from any other company offering a similar product, or does Google Optimizer have a 'get out of jail free' card?
- Assuming the intent is similarly honourable, are PHP based solutions OK to use if they use IP based cloaking to mask their test content?
- How will Google be able to prevent abuse of the technique - for example, at its simplest level, by setting up a test using high converting content in the tests and
SEO targeted content in the base "noscript" content?
- Are the other Search Engines going to penalise this use of the "noscript" tag?
- Or, is all the above irrelevant because once it comes out of beta, Google will advise that we only use Google Optimizer on PPC or e-mail campaign landing pages?
Its a tricky one, no?