We’ve been tracking Twitter’s integration into the Google search results since 2014, and today I’m publishing the fifth of our studies on the topic. The TL;DR of today’s results is that Google’s indexation of Twitter is at the lowest level ever. Read on for the details!
For reference, here are links to the prior versions of this study:
- November 2016
- July 2015, after Google had integrated the Twitter API firehose into their algorithm
- Feb 2015, after Google and Twitter announced their new deal, but prior to completion of the integration
- July 2014
The Core Data
In this data set, we have 693 users who posted a total of 47,799 tweets. On an aggregated basis, only 5.2% of these tweets are in the Google Index. The following shows the breakout of the indexation through all five editions of the study that we’ve done:
[Tweet “Google’s indexation of tweets falls to its lowest level ever. See the complete study at”]
This is the lowest we’ve seen in all five of our studies. It’s also interesting to see how quickly tweets get indexed. We’ve shown through the history of the study that tweets are NOT indexed quickly. This surprises most people, as they assume that Google looks at Twitter to capture breaking news, but that proves to not be the case. Here is the history of the percentage of tweets that indexed within seven days:
Only 1.6% of the tweets were indexed within the first seven days! In our five data snapshots of this, we saw lower indexation levels within the first seven days only in February of 2015. Note that this was back before Google and Twitter announced a new API deal.
Wonder why Google doesn’t bother indexing all social posts? See How Social Media Affects SEO
Factors That Impact Twitter Indexation Levels
Follower count
We also took a look to see how the level of indexation correlated with the user’s follower count. Here is what the data shows:
[Tweet “Study finds that percentage of tweets indexed by Google increases in proportion to the tweeter’s follower count. More at”]
I’m in no way suggesting that Google looks specifically at follower count. However, some aspect of the person’s overall authority does matter. That could be the way that people engage with their account (people with more followers tend to get more engagement), rather than follower count, specifically.
Favorites
That brings up the next question. Do tweets that get more likes (favorites) tend to get indexed more? Let’s take a look:
[Tweet “Study finds tweet engagement affects likelihood the tweet will show in Google search.” ]
The answer is a clear yes. Tweets with more engagement do get indexed more. This is consistent across all editions of the studies that we’ve done. Could this be a direct consideration in whether or not Google indexes a tweet? It’s a possibility.
Retweets
Let’s also look at retweets:
As you can see, a higher level of retweets also correlates well with the potential for being indexed, but only up to a point. However, this occurs at a lower level than favorites. This suggests that the number of favorites is a stronger consideration than the number of retweets, possibly because there are many accounts that are just bots auto-retweeting other people’s accounts.
Twitter verification
We also pulled data on the potential impact of having a Twitter Verified account. Here is what we saw:
[Tweet “Having a verified Twitter account is highly correlated with having more tweets show in Google search. Learn more at]
So having a verified account does correlate with higher chances of indexation, but it’s not clear that this is a direct factor. It could just be because people with verified accounts tend to get more engagement with their tweets.
Length of tweet
Last but not least, we also pulled data to see if the length of the tweet was a factor. Note that due to the timing of our data collection, which was as Twitter was rolling out tweets over 140 characters, we excluded tweets over 140 characters from this data. Google did not appear to be indexing those longer tweets during the beginning of the data collection, but started to near the end of it, and that was going to skew the data. Here is what that data showed for tweets up to 140 characters:
As you can see there is basically, no difference today, for tweets under 140 characters.
Wrapping it Up
The main conclusions of this study are:
- Google’s indexation of Twitter is on the decline. At an overall indexation level of 5.2%, it’s at it’s lowest level ever. Note though that 5.2% of all tweets is still a pretty large number of pages!
- Google continues to index tweets in its own sweet time (not rapidly). This provides support to Google’s public statements that they treat social media sites just like any other site.
- The metric that most increases chances of indexation appears to be favorites, and then secondarily retweets.
- Our available data suggests that tweet length matters little to indexation.
Matt Cutts also mentions that while Google does index Twitter (and Facebook) web pages, it’s certainly at a smaller scale compared to non-social media websites. This is due to people constantly blocking or reporting each other, removing content, privacy settings and more. Not to mention social media websites have their own algorithms and index directives.
I’m also curious if the Twitter index decline is in part due to Google’s focus turning to semantic and voice search. It’ll be interesting to see your next year’s report.
Love the simple images too, thanks!
Good points, JL. I covered similar statements from Matt in my How Does Social Media Affect SEO? (which I should link from this Twitter study!). This is probably the last in this series of Twitter indexing studies, at least for now, as we think we have established the trend, and we want to put the large amount of resources it takes to do this to other studies.
Yes. The expired agreement just applied to the direct real time data feed from Twitter to Google, which powered the for-now defunct realtime section on some search pages.
It did indeed, but it also allows Google to “see” and therefore, if they wanted to, index all tweets. Without this access they would have to scrape Twitter to get tweets, something they would probably not want to commit resources to, and which Twitter might resist. The interesting thing we found is that even with this capability, Google is selective, and chooses not to index all tweets.
I suspect that Google wasn’t treating Twitter content any different than to any other web document.
In each of your measures above, more activity produces more internal links to the thing that you’re measuring and all of the internal links within Twitter are dofollow.
Users with more followers naturally have more links pointing to their profile, which flows more equity to their tweets, which provides them a higher likelihood of being discovered, crawled and indexed – the same is true for retweets and favourites.
Twitter verification is in a similar boat I suspect, not everyone can get verified and I suspect that verified Twitter accounts tend to have more followers, more favourites, more retweets and so forth since those accounts tend to be more popular which is why they were eligible for verification in the first place.
Hi Alistair – Twitter links are all NoFollow, so they don’t pass direct SEO value.
Hi Eric,
I realise outbound links on Twitter are nofollow and don’t pass value. My comment above said internal links, they are dofollow. If you re-read my comment with that lens, what are your thoughts?
Al.
Hi Al – thanks for pointing out that I’d missed the “internal links” part of what you said. I agree that Google is likely treating social media pages like any other web page (and that’s what they say they do). Hence tweets with more links are likely to get more indexation. Interestingly enough though, as I point out in the article, favorites appear to have a stronger impact on indexation than retweets do. My speculation is that the number of robotic accounts that simply retweet things need to get filtered out by Google.