4

We are moving a website from HTTP to HTTPS and I have read Google's support article which is quite good:

https://support.google.com/webmasters/answer/6033049

As I understand, 301 redirects may affect search engine ranking (at least temporarily), so I am considering a plan that could potentially achieve the migration while avoiding redirecting Googlebot, however I have had to make some assumptions which I would like to validate here.

The plan I came up with is:

  1. Install the certificates and get the HTTPS version of the site working in parallel with the HTTP version
  2. Update the canonical links on every page to the HTTPS equivalent. Also update the sitemap.
  3. Sit back for a few days or weeks as I assume the next time Google crawls the site it is going to find the new canonical links and update its index to the HTTPS URL. <-- ASSUMPTION
  4. Once Google's index contains only HTTPS URLs from the site, install 301 redirects from the HTTP pages to the HTTPS equivalents. This should have no impact on Google as the next time it crawls it will be crawling HTTPS links. <-- ASSUMPTION

What do you guys think? Are my assumptions right, and is this a reasonable approach?

  • 1
    You are trying to avoid unavoidable pain. Changing protocols is disruptive regardless of what you do. Do not try to make things easier by skipping a critical step - do 301 redirect. – closetnoc Apr 2 '17 at 15:28
3
  1. install certificate for the main domain and all subdomains
  2. make site accessible with https, check all subdomain combinations
  3. change all internal links in the whole codebase to https
  4. check twice, whether all internal links are indeed https now, inclusive static assets like images, pdfs etc.
  5. establish redirects 301 from http to https, better based on port (RewriteCond %{SERVER_PORT} !^443$ RewriteRule (.*) https://%{HTTP_HOST}/$1 [R=301,L])
  6. check twice, whether all redirects are working
  7. establish new search console property for https
  8. update sitemap and upload it to the new search console
  • Whilst I agree with your process (+1), you've not really answered the question with the points raised/proposed by the OP. – MrWhite Apr 5 '17 at 21:29
2

You should not do this exactly as you state, for a very simple reason:

You should view http and https are two different sites and in this case they'll have identical content, which means duplicated content. The canonical will partially prevent this, but it's not optimal. It's OK to wait a few days and perform your final https-migration-test by visitor usage, but don't wait too long.

The 301 redirect is a very common technique to migrate to https, I suggest you do this. I wouldn't be surprised (and this is speculation) that the smart people at Google have made their bot understand that "a lot a 301 redirects to the exact same content but https" means that you've installed https.

  • 2
    Google is very good at handling duplicate content between HTTP and HTTPS these days. Even if you have both up, Google will simply choose one to index. It usually prefers HTTPS. No penalties for duplicate content. – Stephen Ostermiller Apr 5 '17 at 20:08
  • That's interesting to know. Still, it's IMO good practice to asume it's two domains, this'll make you think before you act :) – Martijn Apr 6 '17 at 7:02
  • @StephenOstermiller if Google finds redirects from http to https, but internal links weren't changed to https, Google assumes redirects are created by mistake and lets rank old http urls. I realized this behavior multiple times and i guess the cause of is: to establish redirect is a one code line in htaccess, but to change internal links to another protocol needs more effort. Google means the webmaster created a redirect by mistake, so Google prefers to let the rankings unchanged. – Evgeniy Apr 6 '17 at 7:29
  • @Evgeniy Interesting thoughts. But have you got any articles to base them on? – Kalle H. Väravas Oct 26 '17 at 17:39
  • @KalleH.Väravas no, but many own experience with same setup: if a redirect is established, but internal linking remains old, Google tends to act, as redirect wouldn't exist. Maybe there are som other factors, for Google decision, like if there are other technical issues on site? If yes, Google means, the webmaster isn't much experienced - and redirect counts as erroneous. – Evgeniy Oct 27 '17 at 10:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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