I finally got around to exploring the Twitter analytics data and wanted to see whether I could find anything useful. My dataset contained 831 tweets, every single one since October 2013, as well as the text, the number of impressions, and the number of engagements. Just by loading the data into Excel, calculating a few values, and generating a pivot table it’s easy to investigate a few ideas. I’ve included some of the pivot tables below along with the various items that stood out.

• Avg impressions vs hashtag and mention (excluding @replies): Idea was to investigate whether tweets with hastags or mentions end up being due to the fact that they are more likely to appear in search results. The results are a bit weird since it seems as if having a hashtag only helps if there wasn’t also an @ mention. Otherwise it hurts.

• I suspected that sending someone an @reply would reduce total impressions but increase the engagement rate since it’s directed at someone. It does reduce the average impressions and only leads to a slight increase in engagement rate - and only when looking at the average of rates, not the total engagements over total impressions. I suspect most people don’t differentiate between an @reply and an @mention which doesn’t lead to a significant difference in actual engagement rates.

• If we exclude @replies and compare tweets with and without mentions, the tweets with mentions have both a higher average number of impressions and a higher engagement. Nothing surprising here - @mentions are good since they draw attention to a tweet while @replies hurt since they limit total reach. Still nice to have some data to confirm.

• Tweets with links have higher engagement - most likely since there’s a stronger call to action. Again this isn’t surprising but nice to see it backed up by a bit of data.