3

Running HTTP and HTTPS versions of the site side-by-side.

HTTP is canonical right now. Goal is to change that to HTTPS.

Created a new site in Search Console and submitted a new sitemap 4 days ago.

HTTPS sitemap is identical to HTTP sitemap (other than protocol). HTTPS sitemap has lastmods showing the date the pages were updated.

Search Console shows all 1M+ URLs in the new sitemap as "Submitted" but it shows "-" for Indexed.

Should we:

  • Wait until Google has indexed the new sitemap before 301 redirecting and changing canonicals?

  • Or 301 redirect and change canonicals now? Maybe Google hasn't indexed the new sitemap because they know it is not canonical?

  • lastmod: Leave as-is, remove, or set to current date?

Chrome shows HTTP warnings soon, so we're eager to move, but don't want to lose Google juice.

unor
  • 21,739
  • 3
  • 46
  • 117
freehat12
  • 31
  • 2
  • It will take a while. Be patient. This answer helps you to understand that particular metric. https://webmasters.stackexchange.com/questions/72610/number-of-indexed-pages-with-site-search-less-than-reported-in-google-webmast/87808#87808 – closetnoc Oct 15 '17 at 00:35
  • @closetnoc Should we wait for the HTTPS sitemap to be indexed before redirecting users to HTTPS? – freehat12 Oct 15 '17 at 00:42
  • Most people redirect right away. But you may be wise to wait a bit. Not sure. – closetnoc Oct 15 '17 at 01:08
  • Your assuming that Google cares a lot about your sitemap, truth is... it plays little part in indexing your site. Indexes are useless for easy to crawl websites, this includes sites with high volume of URLS. Sitemap plays little to no part in how much time is spent crawling your site. – Simon Hayter Oct 15 '17 at 12:56
  • "HTTPS sitemap has lastmods showing the date the pages were updated." - When the "content" was last updated? Or when you moved to HTTPS? (The later doesn't really make sense IMO, but the lastmod date is probably ignored by Google anyway.) – DocRoot Oct 15 '17 at 23:27

3 Answers3

8

Google will never index HTTPS while the canonicals point to HTTP. I switched my largest site over to HTTPS using the following protocol:

  • Enabled HTTPS for the site without switching canonicals for about two years. During this time period Google sent all traffic to HTTP.
  • Switched the canonical version to HTTPS. It has been running that way for about 8 months now. At this point Google is indexing about 90% of the URLS on HTTPS. The number that indexed on HTTPS appears to be moving over rather slowly. I thought Google would have it done within a month, however it has been a much longer process. I was invited to the beta of the new Google Search Console report for indexing. Interestingly it shows that Google is not honoring my canonical on some percentage of URLs. It has crawled the URLs, but has decided that the old HTTP URL is more appropriate! Every month Google appears to switch over more and more of the URLs, but it is unclear when this process will be complete.
  • After about 9 months, I redirected HTTP to HTTPS. I was hoping to wait until Google indexed everything on HTTPS, but it is took so long I made the switch while Google was still indexing a small percentage of pages with less traffic on HTTP. At that point, I didn't notice any effect in my overall traffic when implementing the redirect.

We've had enough reports here of ranking problems with HTTPS switches that I wanted to take this gradual approach with this site because it gets tens of millions of visits per month from Google. The good news is that I haven't seen any ranking problems or traffic dips from Google having switched the canonicals but not redirecting.

I believe that you could skip my initial step of enabling HTTPS but not switching the canonicals without having any problems.

Here are the graphs from the index report of the Google Search Console beta showing URLs not indexed on HTTPS. Google actually has two reasons for this:

  • "Alternate page with proper canonical tag" which seems to indicate that Googlebot hasn't re-crawled the corresponding page on HTTP yet. Alternate page with proper canonical tag
  • "Google chose different canonical than user" which seems to indicate that Google is choosing not to honor the canonical tag and is actually preferring HTTP URLs in some cases. Google chose different canonical than user
Stephen Ostermiller
  • 98,758
  • 18
  • 137
  • 361
  • Just wanted to add, this is the same method I used to successfully move a large site back in 2016. My first stage was only a couple of days while I tested everything worked fine. Then the second stage (canonicals) was around 6 weeks; Google managed to index the https site fairly quickly, and I didn’t lose any rankings. – DisgruntledGoat Jun 24 '18 at 09:21
2

Redirect straight away, the longer you leave http accessible to users/search engines the longer you are dragging out the process.

301 redirects, submit new sitemap, fetch as google bot via search console and then wait it out. Providing your redirects are setup properly your current standing should not be effected.

Randomer11
  • 598
  • 2
  • 9
0

When we update our website URL from http to https than we also have to update it in Google, Bing & Yahoo Webmaster tools with new xml sitemap that contains all HTTPS URLs so that search engines can index your new website URL.

Ken Calhoun
  • 21
  • 2