Caleb Rule (@CRuleSportsGuy) is the Social Media Coordinator at Georgia Regents University.
Shooting for an accurate-yet-brief social media analytics report
Over the last few months, I’ve tinkered with how I report on analytics to sure the most accurate-yet-brief report possible. Sure, it’s nice letting my director know our main Facebook page (of which I’m an admin) is doing 3-4x better than the average engagement for a page – but how much supporting evidence do I need to provide for that to make tangible sense, and is that statistic even terribly relevant?
A Google search on the subject brings up plenty of suggestions, especially from companies with a product to sell. Whether you need the services is up to you, but it can be difficult finding somebody to showcase their numbers and how they got them.
A simple formula for social media engagement
After poking around the wonderful world of the Internet, I’ve also noticed most posts I come across don’t have specific numbers for pages – and so as a thank you for reading, I’ve included some hard numbers you can use to help your benchmarks at the end, straight from two presences I manage today.
Here’s the primary formula I use:
Let’s break this concept down – if you already know what I’m about to explain, skip ahead five paragraphs! (And yes, I will have actual numbers for you to use however you wish at the end of this post!)
First, average post clicks. I use “Engaged Users” in Facebook Insights after exporting data at the post level, and it’s “Engagements” in Twitter Analytics data. For Facebook, I take the average of the entire column, but also take averages of each type of post –note that Facebook often categorizes a post incorrectly, such as photo galleries listed as a status update. This provides a more detailed picture at what types of posts are working better than others, even if it can be a bit tedious.
One thing I also do is delete all of the @replies I’ve made on Twitter. This is because replies are simply that: Replies to a person, and I don’t put a . before the reply because I’m usually not trying to broadcast it to everyone as I would with a normal tweet. Thus, replies will skew the numbers down if you don’t delete them (but it IS worth marking how many individual interactions you have with people on Twitter).
Now do the same with the impressions, which are categorized as “Organic Reach” for Facebook and as “Impressions” for Twitter. You’ll already have sorted the post type and what-not, so this should move along quickly.
With a few months’ worth of data, you should be able to set up rough benchmarks that can be refined with a larger sample size down the road.
Why use these numbers? Facebook already suppresses reach by a large margin – and if you have 1,000 followers on Twitter, you aren’t getting 1,000 impressions per tweet you send. This sidesteps the issue and focuses on the rate of engagement with people who are already seeing your stuff.
Now, for a few supporting details that can add or showcase blips in the data:
- Lifetime negative feedback – always good to glance over
- In FB Insights data, the third tab is “Lifetime Post Stories by Action” and you can see if something was shared a high amount, or just commented/liked by a large group
- The fifth tab shows how many times a link you post was clicked on – or if all of the clicks were on photos you had up.
- For Twitter, retweets are a gold mine – did something have a high amount of impressions because it was retweeted a few times by people with large audiences? (This could actually deflate the engagement rate, too!)
It’s worth noting everything above applies for paid reach as well – and if you to boost a post, you can mark down the number of clicks before putting in the money to have a somewhat-accurate idea of how your engagements rated before and after the boost happens.
Feedback from the community
Two colleagues in a Higher Ed social media Google+ group gave this feedback, in response to a question I posed to see what others in my field do for measurement:
Scott brings up a great key for benchmarking – having a comparable data set. It’ll also help refine those numbers you use to determine success and failure of your goals. (And if you were wondering, yes, he uses Sprout Social to assist his reporting.)
One more – responding to my query to how she calculates engagement stats:
Monica points out two things: First, that knowing and highlighting your most popular content is always a good thing. Second, it can be very easy to get a snapshot-sense of how a presence has performed, just with two statistics.
For Instagram, if you don’t have the new Business Suite they’re rolling out (slowly, I might add), then you can’t measure using post impressions – so for now, using (likes + comments)/total followers will have to do. Keep in mind the number will be lower: Brafton reports the average engagement using this metric is 4.21%.
I’ll show you mine
As a thanks for reading, and after poking around the Internet and being unable to find many people willing to post numbers, here are some relevant statistics I used in my report from this summer.
There are lots of fun conclusions you can find – such as Twitter organically reaching ~25% of our audience, as opposed to Facebook reaching ~14% – but that’s another topic entirely!
Hopefully, these will help you in setting a few benchmarks for your social media presences, and give you an idea of how to streamline your reporting process. Feel free to connect with me on Twitter if you have any comments or questions, and thanks for reading!
Will you show me yours?
I’d love to see your numbers, feel free to share them in the comment so we can learn together.