A number without a range is a guess you've stopped admitting

"Paid social returns £1.90 for every pound." It's a clean sentence. It fits on a slide. And on its own, it's not enough to make a decision with.

Every marketing estimate comes with uncertainty, whether anyone shows it or not. The data is finite, channels overlap, the world changes. A model that reports £1.90 might equally have reported £1.10 or £2.70 with slightly different data. Leaving that out doesn't make the number more certain. It just moves the risk to whoever signs off the budget.

What a range tells you

A range answers the question every budget holder is really asking: how sure are we?

"Most likely £1.90, somewhere between £1.10 and £2.70" says the channel probably pays back, but not by much, and possibly not at all. "Most likely £1.90, between £1.70 and £2.10" says something much stronger, with the same headline number.

Reading ranges in a meeting

Three habits make ranges useful rather than confusing.

1. Compare against break-even, not against each other. A channel whose whole range sits above £1 is a safer bet than one with a higher middle but a range that dips below £1.
2. Read overlapping ranges as "we can't tell these apart". If search is £2.10 to £3.40 and video is £1.90 to £3.20, the data doesn't say which is better. Don't move budget between them on the middle numbers.
3. Treat a wide range as a to-do, not a verdict. Wide usually means the data couldn't separate the channel from something else. That's exactly where a test earns its keep.

Saying it in plain words

Ranges don't need statistical language. "Each pound most likely came back as £1.90, and we're fairly confident it was between £1.10 and £2.70" is understood by everyone in the room. So is "we can't yet tell whether this channel pays back". Both are more useful than a precise-looking number that turns out to be wrong.

What changes when you show them

Teams that show ranges tend to make smaller, more frequent budget moves, test more, and get fewer surprises. Finance tends to trust the numbers more, because they can see where the confidence is and where it isn't.

The alternative is a single number that everyone treats as fact until the quarter it turns out not to be. A guess you've stopped admitting is still a guess.