Engagement rate lets you compare posts fairly
The raw number of reactions (likes, replies, shares and so on) grows with the number of people who saw the post. If you compare posts by raw counts alone, a post that happened to be shown widely looks better than it really is. Engagement rate divides the reactions by impressions or follower count, so you compare posts as a percentage. For a short definition, see engagement rate.
One thing to keep in mind: engagement rate has no single agreed formula. What you divide by, and which reactions you count, differs between each network's analytics screen and each analytics tool. If you compare "an engagement rate of X%" from another company or article with your own number, check that the formulas match. Otherwise it isn't a real comparison.
Three common ways to calculate it
The divisor (denominator) is usually one of these three.
| Divisor | Formula | Best for | Watch out for |
|---|---|---|---|
| Impressions | Reactions ÷ impressions × 100 | Comparing how much viewers reacted, post by post | If the same person sees a post many times, the divisor grows |
| Reach (people reached) | Reactions ÷ reach × 100 | Comparing reactions per person who saw the post | Some networks don't show reach |
| Followers | Reactions ÷ followers × 100 | Comparing with other accounts using public numbers | Reactions from people outside your followers are counted too, so it can exceed 100% |
Dividing by impressions
This is the easiest method for comparing individual posts. Impressions are shown per post on most networks, so they work well for seeing how posting time, day of the week, or having an image changes results. If you only compare your own posts with each other, this method is enough to start with.
Dividing by reach
Dividing by reach removes the effect of the same person seeing a post several times, so you see reactions per viewer. It is convenient on networks like Instagram that show reach, but it can't be calculated on networks that don't.
Dividing by followers
This is often used to compare with other companies' accounts. Follower counts and reaction counts are visible from the outside, so you can calculate it under the same conditions even for accounts whose analytics you can't see. However, a post that spread through recommendations can get more reactions than you have followers, so the number reflects how far a post spread more than how good it was.
What to count as a reaction
What goes in the numerator matters as much as the divisor. Counting only likes gives a very different result from also counting replies, shares, saves, link clicks and profile visits. Even the "engagement" figure in a network's own analytics includes different things on each network. When you use a network's number, check its help pages to see what it counts.
If you calculate your own, count only reactions that every network provides, so you can use the same yardstick across networks. For example, limit it to three: likes, replies and shares.
Decide your own formula
When you use engagement rate inside your company, decide these three things and write them down.
- What you use as the divisor (impressions, reach or followers)
- Which reactions you count in the numerator
- Which day's numbers you use (the number measured how many days after posting)
The third is easy to overlook. Impressions and reactions are different right after posting and a week later. If you fix the timing, for example "compare numbers on day 7 after posting," you can compare posts published on different days.
Example calculation
Take a post with 2,000 impressions, 30 likes, 5 replies and 5 shares. Dividing by impressions, the engagement rate is (30 + 5 + 5) ÷ 2,000 × 100 = 2.0%. If the same post was published on an account with 800 followers, dividing by followers gives 40 ÷ 800 × 100 = 5.0%. The same post gives a result 2.5 times higher depending on the divisor.
Cautions when reading the numbers
A higher engagement rate isn't always better. A post with few impressions is seen only by your most loyal followers, so its rate tends to look high. A post that was recommended and shown to many people also reaches people with little interest, so its rate tends to drop. Look at the rate together with impressions to find posts that reached many people and also earned reactions.
Benchmarks like "the industry average is X%" usually don't tell you the formula or which accounts were measured, so comparing against them doesn't help you decide anything. Compare against last month's number calculated with the same formula, and against your own other posts. For the overall approach to analytics, see the basics of social media analytics. For ways to raise the rate, see how to increase engagement.
Steps for calculating it in a spreadsheet
The easiest way to keep this going is to enter the numbers into a spreadsheet in the same format every month. Use one row per post, with columns for "date posted," "network," "impressions," "likes," "replies" and "shares," then add "total reactions" and "engagement rate" columns to the right. If total reactions adds up only the reactions you chose, and engagement rate divides that total by impressions, you only need to paste in the numbers to get the rate.
Once the table is built, don't just sort by engagement rate. Also compare it with a version sorted by impressions. Sort posts into three groups: high rate but few impressions, many impressions but low rate, and high on both. Each group teaches something different. A post with a high rate and few impressions hit a topic your loyal followers care about, and might spread if you change the posting time. A post with many impressions and a low rate reached outside your followers but may not have given people a reason to react.
When you work out a monthly average, don't average the per-post rates. Divide the month's total reactions by the month's total impressions instead. That way one extreme post doesn't pull the average. Decide which method you use together with the rest of the formula.
What you can do with SNS TOWN
Analytics in SNS TOWN doesn't boil engagement rate down to a single formula. Instead, it lists the raw numbers you need for the calculation, post by post. You can view a table of impressions, views, likes, reposts and replies for each post over the last 30 days, and the CSV includes impressions, views, likes, reposts, replies, quotes and bookmarks for each post.
Open the CSV in a spreadsheet and you can calculate engagement rate with the divisor and numerator you chose. Numbers are read every day until day 3 after posting and once more on day 7, so they fit a rule like "compare numbers on day 7" as they are. You can also see the average reactions per post by day of the week and time of day on screen. It supports X, Instagram, Facebook Pages and YouTube. Check pricing for the plans that include it.
7 दिन मुफ़्त आज़माएँ। क्रेडिट कार्ड की ज़रूरत नहीं।
मुफ़्त में शुरू करेंFrequently asked questions
What engagement rate counts as good?
There is no single benchmark, because it varies widely by formula, network, industry and follower count. Compare against last month's number calculated with the same formula, or against your own other posts, and see whether it went up or down.
The rate in a network's analytics differs from my own calculation. Why?
The two count different things in the divisor and numerator. If you use the analytics screen's number, check the network's help to see how it is calculated. If you calculate it yourself, use your chosen formula the same way every time. The important thing is not to mix the two.
I divided by followers and got over 100%. Is that wrong?
No. When recommendations show your post to people who don't follow you, their reactions are counted too, so reactions can exceed your follower count. You can read it as a sign that the post spread beyond your followers.
Should video views count as reactions?
Views can be counted just because a video was shown, so including them makes the rate look large. It is easier to read if you look at views separately and count only actions that take effort, such as replies and shares, as reactions.
Should I set a target for engagement rate?
If you do, work out your own average for the past three months or so and aim slightly above it. A target based on the rate alone can push you to lower impressions just to raise the rate, so pair it with a target for impressions.
Is there a way to check other companies' engagement rates?
You can't see other companies' analytics, so use the follower-based method with the visible follower count and each post's likes, replies and reposts. It uses a different formula from your impression-based rate, so don't compare it directly with your own. Use it to compare other accounts with each other under the same formula.
Which is better, a post with many replies or one with many likes?
It depends on your goal. If you want to start a conversation or hear from customers, replies are the clue. If you first want to be noticed, likes and shares are. Look at the count for each type of reaction as well as the total engagement rate to see how posts differ.
