A/B Testing and SEO: impacts and best practices to know [Exclusive Guide]

1:49 pm April 18, 201711576

use A/B testing is the best way to optimize conversions on a website today. What are best practices to know that tests don’t impact not referencing of the site or of the tested pages? Here is a summary of a few best practices learned from the white book “Maximize conversions by securing its SEO” proposed by Primelis and A/B Tasty.

reminder: what is the A/B Testing?

the A/B Testing is a practice that is to propose a certain percentage of Internet users a variant of the original page to find one that will get the best results (conversions, clicks on elements, use of a search engine, subscriptions,…). There are many types of tests it is possible to set up a site:

  • A/B Test classic
  • Test A/B/C
  • multivariate Test
  • Test by redirect
  • Test MultiPage

all the tests performed on site do not have the same impact on SEO the white paper proposed by these two companies is up on best practices to know.

Click here to download and read this free white paper

SEO and A/B testing: best practices to know

if most changes tested via A/B tests will not or little impact on SEO and positioning SEO of a site, some precautions and good practices are know to avoid nasty surprises.

what General SEO recommendations?

  • find your SEO of your tests referent: if many tests will not impact, it is important that the SEO Manager is aware of current and future tests in order to anticipate possible technical optimisations upstream.
  • do not exclude user-agent Googlebot in your A/B tests : If Google is favourable to A/B testing, search engine ask to have access to the same versions and variations than the user to avoid any downgrades by mistake.
  • avoid too different variations test: testing a page with too many variations is generally not very recommended, Google might as well believe in an attempt to manipulate (Google penalizes the cloaking). If such tests are required, it is therefore recommended to indicate to Google that this variation is actually the same page as the previous using the canonical tag (in the case of a test by redirect).

best practices SEO to complete my A/B testing by limiting risk SEO

1 – best practices in testing by redirect:

If you want to set up a test by redirecting your site, these recommendations can avoid you a downgrade or loss of positions SEO in the natural results of Google :

  • do not block Google : access to the variant must not be altered by a Disallow command: in the robots.txt and a noindex should not be added on the variant to which traffic will be redirected (it must however apply the second recommendation in this case).
  • use the canonical variant page attribute: the variant page must have an attribute canonical tells Google that the page to take into account and to keep in its index is the original page. If you carry out a test by redirection on a set of page, each page variation must individually have a canonical attribute to the original page.
  • redirect users via a 302 or a Javascript redirect : Google must be able to understand that it is a temporary redirect, the redirect 302 (Moved temporarily) is ideal.
  • delete and redirect the variant once the completed test page: after the test and set up the winning variant in production, it is recommended to delete the variant page and redirect 301 verse the original page.

test par redirection checklist seo

2 – best practices in testing by overlay Javascript:

the application of overlays Javascript on the same page is the most common method to do some A/B testing. In practice, these changes are applied when loading of the page (the rest page so the same) through some overlays Javascript generated by your A/B testing software.

as long as you do not offer a completely different content on your variant, this type of test is usually not a problem for Google even if it interprets better Javascript. It should however be borne these some best practices in mind:

  • the test should not change deeply the content and structure of the page : the number of tested items must be the lowest possible
  • the overlays should, ideally, not remove or hide important elements for the positioning of the page : ex: content main page text, blocks of internal mesh, Hn tags, important images for the page,… (This recommendation is especially valid for pages that have a goal SEO and my crawlées / indexed by Google.) If in doubt, see your SEO Manager.)

ex of tests that pose no SEO problems:

  • color change of CTA
  • change of wording on the CTA
  • change of location of some blocks
  • variant which hides one or several items not containing the main content of the page or the elements important level SEO (mesh blocks (, H1,…)

ab testing surcouche javascript seo

3-other good practices related to A/B testing mobile and engagement solutions:

  • and A/B testing on mobile index mobile-first : Google announcing the next arrival a mobile-first index (2017-2018), it is strongly recommended to make more and more attention to mobile tests involving to hide important elements for SEO as the texts at the top of page internal mesh blocks or breadcrumbs.
  • engagements and SEO, limits don’t cross : many A/B testing A/b tasty solutions now offer solutions to boost the commitment of the visitors. If most of the latter will not impact SEO, it is recommended to not display pop-in commitment taking too large a share of the page on mobile under penalty of decommissioning by Google .

you’ll understand, to do no harm to the referencing of your site, it is strongly recommended to integrate and apply these simple SEO tips in your process.

Rezndez you to this address to download the full white paper: http://www.abtest-seo.com/


your opinion on this info?