Interaction Rate for a single post
For a post, the interaction rate is relatively simple: it's the number of interactions (e.g. reactions, comments, shares), accounting for weights (e.g. if you give shares 3x the weight, they get 3x the interactions), divided by the account size (followers/page likes) at the time of posting.
So let's assume our dashboard has even weights for all interaction types. A Facebook post gets 10 interactions and its Page had 800 page likes when the post came out, that's: 10 / 800 * 100 = 1.25%
Interaction Rate for multiple posts by a single account
This is often seen in a leaderboard or Intelligence, or in search, and the goal is to find an average interaction rate from the many posts in that timeframe. To do that, we add up the interactions of all the posts (again, accounting for weights), divide by the number of posts, and then divide again by the average account size (followers/page likes) in question over the timeframe.
The example below is for all posts in a timeframe by a single Facebook Page:
Let's say we're looking for the interaction rate of a Facebook page over a week. It has 7 posts totaling 100 interactions (with even weights across all interaction types), and the page grew from 1,000 to 1,500 page likes that week (for an average of 1,250). That's: 100 / 7 / 1,250 * 100 = 1.14%
Interaction Rate for multiple posts by multiple accounts
For lists in Intelligence, we try to do the same kind of thing we do for an account: We add up all the interactions on every post from every account in the list (again, accounting for weights), and then we divide by that number of posts. Then we divide by the average size (follower count/page likes) of the accounts in the list. (In this case, we pick the latest size value for each account, not an average for each over the timeframe.)
Let's say we're looking at a Facebook list with 5 pages in it, each with 6 posts over the week, totaling 100 interactions (with even weights across all interaction types). The page sizes on the last day of the week are 1000, 1100, 1200, 1300, 1400, which average to 1200. That gives us:
1,200 average size (followers/page likes)
500 / 30 / 1200 * 100 = 1.39%
We can only get account size (followers/page likes) from the moment we start tracking the account within CrowdTangle. If you look for an interaction rate before we started tracking it, we can't create an interaction rate over that timeframe.
For Facebook, we only track page likes, not followers.
We track Instagram followers and Reddit subscribers.
Weights describe how much influence you want to give to one particular interaction type. They are on the bottom left of your screen.