HASHTRACKER

X and Twitter analytics 3 min read

Real reach against potential, and why the gap is so wide

Summing followers gives a number in the millions. Dividing the impressions X counted by that number gives the one you can defend in a meeting.

Cover drawn as a column chart

Add up the followers of every account that mentioned your brand and you get a number in the millions. Put it in a slide and somebody will ask whether that many people really saw it.

They did not, and the real answer is one division.

The two numbers

Potential reach is the sum of the followers of every account that took part, counting each account once. It is the audience that could theoretically have been exposed to something.

Real impressions is how many times X actually showed those posts, as X counts it for each post. It is not an estimate and nothing about it is modelled.

Divide the second by the first and you have the figure that ends the argument. A conversation whose real impressions come to eleven per cent of its potential reach is a conversation that reached about a ninth of the audience the follower count suggested.

See both on a sample report

Why the gap is so large

Nobody sees everything the accounts they follow publish. Timelines are ranked, people are asleep, and most posts are shown to a fraction of the people who could see them.

Potential reach also double counts in the way that matters least. If five accounts with a hundred thousand followers each mention you, their audiences overlap heavily, and the sum pretends they do not.

None of that makes potential reach useless. It makes it a ceiling, and a ceiling is worth knowing as long as nobody reports it as a floor.

What the ratio is good for

Comparing campaigns. Two campaigns with the same potential reach and very different ratios were not equally effective, and the follower totals hid it.

Comparing accounts. An account whose posts reach a high share of its own audience is worth more to you than a bigger account whose posts reach almost nobody.

Spotting the bought audience. A very low ratio on a specific account, repeated across months, is one of the few signals that its followers are not real people who read it.

What it is not good for

A target. There is no good number here. It moves with the platform, the subject, the time of day and the format, and chasing it produces reporting that optimises for a ratio instead of for an audience.

Say what it was, say what it was last time, and move on to the reason.

How to report it without a fight

Give both numbers in the same line, in that order. "The posts were shown 448,000 times. The accounts behind them have four million followers between them, so about eleven per cent."

That sentence is impossible to argue with, which is the entire point of using a counted number instead of an estimated one. If you want the longer version of why one is counted and the other is not, we wrote it in reach against impressions.

Every report carries both figures and the dataset behind them, gathered within the Twitter compliance rules. Tell us what you need to measure and we will tell you what a report would show.

Tell us what you need to measure

Choosing a subscription

Every subscription carries the same product. The only decision is how many tweets you need to analyse.

See the subscriptions