I'm planning to build an application which will end up displaying duplicate content.

app.domain.com/content/ID # Completely rendered on the client, behind a wall of JS
www.domain.com/content/ID # Server-rendered, SEO-friendly

Under app.domain.com I'll be serving a React SPA that renders completely on the client.

Under www.domain.com I'll be serving the same React application but rendered on the server, to improve SEO.

The content on both applications will be the same, but presumably the server-rendered version will have better SEO. To avoid duplicate content problems, I'll be following Google's "Consolidate duplicate URLs" guide. I will:

  • Specify a preferred domain: www.domain.com.
  • Set a rel=canonical header in my response for pages under app.domain.com/content/ID pointing to the corresponding www.domain.com/content/ID URL.

Here is my question: which application's page speed counts towards SEO? Is it the page speed of the client-side rendered application (slow because of all the JS that needs to be downloaded in a blocking way) or is it the page speed of the server-render, SEO-friendly version?

  • Why have the client rendered version at all if nobody is going to be using it? Search engines will send all the traffic to your SEO friendly canonical. Commented Mar 13, 2019 at 11:48
  • I'm planning to have people use the app.domain.com version because it will render the same content but inside of an application that allows the user to interact with it more heavily.
    – fnune
    Commented Mar 13, 2019 at 12:29

1 Answer 1


I have not seen this exact situation, but in similar situations (where there are two URLs and one points to another) Google tends to use the canonical version.

Word of warning though: Splitting up your site between app.domain.com and www.domain.com will be "eventually" consistent in Google. While they do recognize canonical tags, they don't always do so very quickly.

The duplicate domain will slow you down from an overall SEO perspective.

What kind of speed difference are you talking about by the way?

If it's the difference between 1 second and 2 seconds then it's probably not worth having them separate.

Use a tool like gtmetrix to run speed tests on your competition.

If all your competitors are loading after 5-10 seconds, then your improvement from 2 seconds to 1 second won't really make a big difference in the grand scheme of things.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.