The average Google ranking of a URL is highly correlated with the number of Tweets about that URL, says Branded3 [download page] in an April 2012 report, though close examination of the report reveals that other factors could have a major role in the report’s conclusions. The company looked at data gleaned from its Twitition website, in which users who sign a Twitition have a tweet automatically sent from their account. With a sample size of 8,528 Twititions collected on February 28, 2012, and having gathered the rankings on April 6, the company found that URLs with over 500 signatures (tweets) had an average ranking of 46. The average ranking improved to 41 for those with over 1,000 signatures, 31 for those with over 5,000 signatures, and 5 for those with over 7,500 signatures.
Other Factors Could Play a Role
Of course, while the finding is intriguing, some other variables could be at play, a fact acknowledged by Branded3. For example, the report operates under the rather sizable assumption that the number of signatures on the site equaled the number of tweets about that specific URL. Additionally, URLs with a high tweet volume are more likely to have media coverage and links from blogs and news sites, and could rank highly regardless of links or shares. Similarly, many of the Twititions with more than 1,000 signatures were specialized and possibly rank higher due to existing in a small niche. Finally, the study did not look look at the impact of longevity.
Given that it was not possible to isolate those factors, it is important not to confuse correlation with causation, despite Branded3 claiming that it has proved that tweet volume does affect ranking. Indeed, the presence of these factors call into question the conclusions of the study.? Given the attention currently paid to the relationship between social media and search rankings, though, the study’s attempt to correlate them is interesting.
Correlation High For Small Volume; Absent for Middle
The study found that URL rankings tended to rise alongside the volume of tweets when those numbered between 1 and 50, but then leveled out until around 500 tweets. From then on, the positive correlation was apparent again, and much stronger.
In terms of the sample size, 5,322 of the Twititions had between 1 and 99 signatures, 1,382 had between 100 and 499 signatures, 1,824 had more than 500 signatures, and 770 had at least 1,000 signatures.