Why the account that posts most is rarely the one that gets seen
Impressions per account is one division away from any export and it reorders every ranking in a report. What it tells you, and the trap inside the average.

The account that posts most in a conversation is rarely the account that gets seen most. Everybody knows this and almost every report still ranks accounts by how much they published.
The figure that says something is impressions per account. Take everything an account posted, add the impressions, divide by the number of posts. One line of arithmetic, and it reorders the room.
Volume is an effort metric
Counting posts measures how hard somebody worked. It says nothing about whether the work landed. An account that published forty times and was seen four thousand times did worse, per attempt, than an account that published twice and was seen a thousand times.
In the first case somebody is busy. In the second somebody has an audience.
Where the two rankings disagree
Put the ranking by volume next to the ranking by impressions per post and read where they disagree. That gap is the interesting part of the report.
An account high on volume and low on efficiency is shouting. An account low on volume and high on efficiency is being listened to, and it is usually the one worth contacting, quoting or hiring.
It is also how you find the account you did not know about. The big names are on both lists. The surprise is almost always in the second.
Averages have a trap in them
One post that went enormous drags an account's average up and makes a single lucky day look like a pattern. Before you act on a high average, look at how many posts it came from. Two is an anecdote. Twenty is a habit.
The report gives you both numbers side by side, which is the point of showing them together.
What to do with it
For a campaign report, this is the list of accounts that actually carried your message, in order. Not the ones that mentioned you most.
For competitor work, it tells you whether a rival's visibility comes from posting constantly or from a few pieces that travel. Those are different strategies and they cost different amounts of money.
For influencer work, it is the closest thing to a shortlist you can defend, because it is built on what the platform reported rather than on a follower count that anybody can buy.
The question it cannot answer
Whether being seen was good for you. A critical thread seen half a million times scores beautifully on this metric, and so does an announcement everybody loved.
Read the tweets. That advice never gets old, and this metric is the fastest way to decide which twenty to read.
Every report ships with the dataset behind it, gathered within the Twitter compliance rules, so any of these averages can be rebuilt by hand. See the plans or tell us what you need to measure.