Google permits and encourages A/B testing and has stated that performing an A/B or multivariate test poses no inherent risk to your website’s search rank. However, it is possible to jeopardize your search rank by abusing an A/B testing tool for purposes such as cloaking. Google has articulated some best practices to ensure that this doesn’t happen:
- No cloaking. You should not use an A/B testing platform to change the spirit of the page. If Google determines that the variation of your page is substantially and materially different from the original, not just in design but also in scope and content, then they may construe this change as cloaking, and your site may be subject to a penalty. However, according to Google, “Google does not construe the ethical use of testing tools to constitute cloaking.”
- For experiments involving redirects, use the rel=canonical tag. If an A/B test has multiple URLs, place the rel=canonical link attribute on all of your alternate links, pointing to your original page. This will help point bots indexing your website to your original page. Google's search quality evaluator guidelines say "Redirects from one page on one website to another page on the same website are also not sneaky. However, unexpected redirects from one website to a completely unrelated website should be considered deceptive." Experiments involving redirects should be fine as long as they don't redirect to unexpected or unrelated content. See Google support's article on sneaky redirects for some more information on this.
Optimizely snippet's loading times' effect on your SEOSite speed is one of the factors that Google evaluates for its search rankings. Since Optimizely is executed front-end and is loaded synchronously, it should not have an impact on the SEO rankings of your site. You can also check Optimizely's best practices on how to Implement the Optimizely Snippet to optimize the loading speed of your snippet.
How long to run an experiment considering SEO
- Update the content of the original page with the content from the winning variation. This is the best option as it preserves all link equity. You may still see a loss in ranking if the variation is significantly different from the original or has less content.
For more information, we encourage readers to view the following resources:
- Google Webmaster blog post on testing
- Google Content Experiments article on experiments and cloaking
- Hacker News thread with Matt Cutts, engineer at Google (“I’ve been quite clear that there’s nothing wrong with A/B testing.”)
- Google Webmaster blog post from November 2015 with an update of their search quality rating guidelines