Twitter recently unveiled a free (and very powerful) analytics tool that gives ordinary users a chance to see which of their tweets work, and which ones don’t. I highly recommend that everyone who cares about the effectiveness of their Twitter account spend some time reviewing the analytics tool, located at analytics.twitter.com.
I spent about 30 minutes reviewing my results the other day. Here’s what I did and what I learned:
To get started
- Using the “export data” button in the top-right corner, I downloaded all of my tweets into a nice excel spreadsheet.
- I then deleted everything that wasn’t in the past 90 days to avoid suffering from analysis paralysis.
My first question: When are my tweets seen the most? I sorted the column by engagement rate and determined that, for me, the answer was the morning, easily. I also learned that weekends are not nearly as effective, for me, as tweeting during the week. This says something to me about my audience: it tends to use Twitter more as a professional tool, rather than a personal one.
Engagement – What works and what doesn’t
Next, I reviewed my highest engaged tweets for ideas on what content works, and what doesn’t. This took some subjective review and roughly categorizing tweets into categories (“controversial political statements,” “blog updates,” etc). For me, what worked the most:
- Tweets that referred to conversations that were already occurring (people wanted to jump in!).
- Social media scandals that were occurring as I tweeted them (Takeaway: Live content builds engagement).
- Social media scandals.
- Use of hashtags, pictures or links.
Remember, Twitter defines engagement as anytime a tweet is clicked on, expanded or replied to…engagement, simply put, means interaction.
As for what didn’t work:
- Wistful tweets…the equivalent of me talking to myself. Noted.
- One-off comments that have nothing to do with anything and appear out of place in my timeline. Again, I should have figured that one out on my own!
- Controversial information
- Cool personal information
- Breaking news items
I did more analysis of other categories as well, including:
- Content that got the most replies
- Content that got the most favorites
- Profile clicks
- URL clicks
- Hashtag clicks
- Detail expands
- Media clicks
By the way, please do not use my analysis in place of yours. Every Twitter audience – and every Twitter account, will be different. I offer this as a template for you to conduct your own analysis.
Have you given this tool a review? What do you think – how did your analysis go?