Does Twitter have an impact on indexing speed

by Peter Young on January 21, 2011 · 14 comments

Indexing test
In the week following the Christmas break, we were talking in the office about whether Twitter would have an impact on indexing speed. It seemed pretty obvious to us that although Twitter links are nofollow that Google must use the data in some way when indexing webpages – after all, if a link picks up a number of key retweets then it must be interesting to the users that retweet, so therefore must carry some weight?

The tests
After gathering data for the Twitter test, it became obvious that I should be adding some more variants to the test – to see which was the best way of getting content indexed by Google.

The tests I undertook involved creating a number of test pages that were:

  1. Linked to from Twitter only with a number of retweets from powerful accounts.
  2. Linked to only from the homepage of my site (PR3, DA 33)
  3. Linked to only from Sitemap.xml for my site, with a priority of 1 and towards the top of the file. Was also resubmitted to Webmaster Tools.
  4. Linked to only from one external link (followed, PR1, DA 35)
  5. Linked to only from one external no-followed link (PR3, DA 44)

Notes: all of the pages had a very small amount of unique content on them, and no links on the page.

Expected results
I never really knew what to expect with the Twitter test – but I did expect it to be indexed – just not very quickly.

The order I was expecting the pages to be indexed was:

  1. Homepage link
  2. External link
  3. Twitter
  4. XML Sitemap
  5. Nofollow link

The Results

Test Created Tweeted/Submitted Cached Diff Order
1 06/01/2011 12:08 06/01/2011 12:09 06/01/2011 18:16 6 hrs 2
2 06/01/2011 14:33 06/01/2011 23:41 11 hrs 3
3 06/01/2011 13:30 06/01/2011 13:31 11/01/2011 09:23 5 days 4
4 12/01/2011 10:47 12/01/2011 15:22 5 hrs 1
5 12/01/2011 15:35 n/a n/a 5

Whilst not completely random, the results are somewhat different from what I expected. Why? Twitter links are all nofollow, so tests 1 and 5 are essentially the same, yet these two appear at opposite ends of the table.

This, in my mind, proves that Twitter links (and maybe others, Wikipedia anyone?) are treated somewhat differently by Google in their indexing. With Google obviously applying some sort of social weighting based on number of retweets and the number of followers of the account retweeting.

The homepage link and the followed external link show very similar results – as expected as they are both very similar links in terms of power (DA 33 to DA 35).

Another surprising result (for our team anyway) was the fact that the XML sitemap link got indexed at all. The fact that no links point to this page, either internally or externally and the page can still be indexed does highlight the power of the XML sitemap for sites that have problems with indexing.

Note: very interesting is that GoogleBot visited the Twitter test post almost to the second that the tweet was retweeted.

So which method is best – and which one should you use to get your content indexed quickly?

Well the answer is obviously to give the search engines as many quality signals as possible. For example, for an e-commerce site the best advice would be a latest products section on the homepage and separate sitemaps for each category of products. And tweet about the products too – eBuyer do this very well.

Dan Taylor is a SEO Manager at Just Search. A regular blogger and prolific tweeter, he is also a regular on the Manchester SEO scene.

Google+ Comments

{ 12 comments… read them below or add one }

Dean Cruddace January 21, 2011 at 1:02 pm

Happy to be a guinea pig Dan, it was interesting to watch the outcome as it happened. Deployed a very similar method for an affiliate site, zero link build beyond twitter, but had it indexed within a few hours. Hard to isolate exactly but it was definitely down to the RT factor imo.

Glenn Ferrell January 21, 2011 at 2:13 pm

Great topic. Several questions. What do the terms PR1, DA 33, etc refer to ? Also, test 5 appears to have no result at all in the spreadsheet. How do you draw conclusions from this ?

Dan Taylor January 22, 2011 at 2:57 pm

Hi Glenn,

thanks for your comment, PR relates to PageRank (Google), DA relates to Domain Authority (SEOMoz).

The reason that test 5 does not have any data is due to the fact the page has not been indexed at the time of writing by Google – my conclusion is that Google is adhering to the nofollow rule, however as I mentioned, Twitter links are nofollowed too so Google must operate a ‘special case’ follow rule for Twitter and maybe other sites?


Dan Taylor January 22, 2011 at 2:59 pm

It is interesting to see the two tier set of rules that Google is applying to the nofollow rule on certain sites – hence the reason people still chase Wiki links!


Dean Cruddace January 24, 2011 at 11:56 am

You could also consider twitter feeds being drawn via RSS into sites where what was a nofollow suddenly becomes a follow. I could be wrong.

A vanity search + twitter might expose some twitter scrapers where the nofollow is ignored or not applied, which in turn may influence.

Thos003 January 24, 2011 at 8:17 pm

If you are publishing good content then why wouldn’t you post it on twitter?

Still, thanks for the test. It’s always fun to test things.

Josh January 24, 2011 at 8:17 pm

This is some good information to look at Dan, but I’ve got a question. Do you think the authority or even the age of the Twitter account used played a part into the speed of indexing? For instance, a brand new twitter account (less than 24 hours old) have the same indexing speed as an account that has been around for a few years and have a great number of followers? Your thoughts?

Dan Taylor January 25, 2011 at 5:34 pm

True – that could be that case, but many have believed it true of many powerful sites with nofollows for a while now, what are your thoughts on that?


Glenn Ferrell January 27, 2011 at 4:48 am

Dan — Thanks much for the clarification !

Dan Taylor February 8, 2011 at 9:35 am

I think that you have hit the nail right on the head – the post was retweeted by a number of high profile accounts, as well as my own which ia fairly established.

I would definitely look for a few powerful retweets rather than thousands of no-power tweets.

G.Brunner translation March 21, 2011 at 3:46 pm

Quite interesting. I have always thought that the nofollow attribute would be one of the most interesting aspects of SEO to investigate, and completely agree with you that Twitter links can impossible be ignored by Google.

Web Site March 16, 2013 at 7:28 pm

Thanks for sharing your thoughts on dan taylor.


Leave a Comment

{ 2 trackbacks }

Previous post:

Next post: