I set up an experiment on Google Analytics and serve two versions of the same page with slightly different content.
Both pages are available online, so Google will index them both.
How to make sure that those two pages don't get "punished" by the search algorithm for having the same content? Or is it fine and Google Analytics "understands" those cases and gives more priority to the original?
When Google finds duplicate pages it will just choose one or the other to index but not index both. Google almost never "punishes" for duplicate content. Google only reduces rankings when the duplication is scraped or causes Googlebot to keep crawling duplicate pages forever. See What is duplicate content and how can I avoid being penalized for it on my site?
When you run a Google Analytics experiment, Google suggests that you choose one of the variants as the "canonical". The other one should have a rel="canonical" tag. That way Google knows which of the two is the preferred one to index.
Some people also worry that Google will penalize them for cloaking when running GA experiments. That isn't the case. As long as the "spirit of the pages" is preserved between variations, Google won't penalize for cloaking during GA experiments
Google wants to know what the page will look like when it sends its users from search to your page. As a result, you should have 100% consistency between the page that Googlebot sees and search visitors see.
It is ok to serve up pages with modifications and variations if the person visiting your page has already been on your website or came from a different source than Google. But Google needs to know what page the visitors it sends are going to be viewing, and so your organic search traffic page should be the same as the one that Googlebot sees.
Using duplicate content on your website can indeed hurt your rankings. Google will decide which pages to index and send traffic to, but if there are pages with duplicate content then this can cause dilution in your rankings. Google may display both of your pages as the 25th result on its search engine instead of one of your pages as the 10th result.
If you are going to be using duplicate content on your pages then it is imperative that you use the rel="canonical" tag. The canonical tag indicates to Google which of the two websites you want it to index for its search results and which of the two pages Google should avoid indexing. This will prevent dilution of your pages and will ensure that your organic search traffic visitors are visiting the page that was intended for them.
