What's a deal worth before it closes?

A sensible estimate is enough

The commonest mistake is waiting to fill the figure in until you know it exactly. The field then stays empty, and a pipeline with half its values missing is useless for the one job it has.

Put in what you think now, rounded. If you mostly sell annual contracts around the £1,000 mark, put £1,000. What matters is the total, not the rounding: whether a deal sits at £900 or £1,100 makes no difference at all to the question of where to go next.

What does matter is updating it the moment you know. A quote that's gone out has a figure on it — that isn't an estimate any more.

The trap: probability by stage

Plenty of CRMs hang a percentage on each stage. Lead 10%, contacted 25%, quoted 50%, negotiating 75%. Multiply by the value and you get a "weighted pipeline".

That looks precise and usually isn't. Those percentages come from nowhere. They ship with the product, they aren't based on your numbers, and at the volumes a small firm deals in you can't derive them reliably either — that needs hundreds of deals, not twenty.

What you end up with is a number everyone looks at and nobody dares act on. Worse: it feels like knowing something, when all you've done is a multiplication.

Two numbers that do say something

If you want to do some arithmetic anyway, these are the two you can keep without a dashboard, and they don't pretend to know more than they do.

Your total open value, by stage. Not weighted — just the sum. A lot sitting at the start means you'll have a problem in a month. A lot sitting at the end means you need to sort out capacity. That's usable, and you don't have to believe anything to use it.

Your own win rate over the past year. How many quotes did you send, and how many landed? That's one division, it's about your work, and it's far more trustworthy than a default percentage out of a box. Twenty quotes of which six went ahead is thirty per cent — and thirty per cent is a figure you're allowed to apply to your quoted stage.

The date matters more than the figure

An expected close date looks like the softest field of the lot, because it's almost never right. That's exactly what makes it useful: a date that's been pushed back three times is itself the signal.

That's information a figure won't give you. A £2,000 deal that's been closing "next month" for four months isn't an opportunity any more, it's noise — and it belongs in the lost column, with a reason. More on that in the piece on lost deals.

One deal is one decision

A common mistake that quietly inflates all your numbers: creating a deal per product or per invoice. You then have three cards that always move together, and every figure you pull out is three times too high.

The rule: one deal per decision the customer makes. If they decide on the whole package in one go, that's one deal. If they decide on a trial first and the rollout later, that's two — because the second can fall through while the first went fine. See also the piece on setting up a pipeline.