A false forecast is worse than no forecast. The problem is almost never the tool: it's the absence of an objective exit criterion at each stage.
You forecast $400,000 for the quarter. You do $180,000. Next quarter you forecast $350,000 “being conservative”, and you do $200,000. The problem isn't your optimism. It's that your pipeline stages don't mean anything.
The mechanism.
Ask three salespeople what the Proposal stage means. You'll get three answers: I sent a document, they saw pricing, we discussed it on a call. Three definitions, three different implied probabilities, and a forecast that aggregates things that aren't comparable.
Made worse by a second mechanism: opportunities move forward but never backward. A deal that reached Negotiation stays there for six months, because moving it back looks like an admission of failure. Your pipeline inflates mechanically.
The exit criterion.
A stage only means something if you can answer yes or no to: is this condition met? No judgement, no “it's coming along”.
Qualified: a problem named by the client, a confirmed budget, an identified decision-maker. Discovery: the quantified consequence of the problem, said by the client, not by you. Proposal: the decision-maker has seen the price and hasn't said no. Negotiation: one remaining obstacle, named, with a date. Won: signature.
If two salespeople can disagree about which stage a deal is in, the criterion isn't precise enough.●
Two rules to apply on Monday.
The reversal rule: an opportunity with no client activity — theirs, not yours — for 21 days moves back one stage automatically. No discussion, no exceptions. A pipeline that never moves backwards is a pipeline that lies, and this rule alone closes half the forecast gap.
The decision-maker rule: a deal where you know one person doesn't pass Discovery. A single-contact deal is lost when that person changes job — and they change job more often than you plan for.
Weighting, done properly.
Don't weight by the percentages your tool ships with: they come from a global average that says nothing about your company. Take your last twelve months. For each stage, count how many deals entered it and how many were won. That's your real rate.
The first time, it's always lower than expected — and that's precisely the information you were after. Recalculate quarterly: a rate that moves sharply signals either a market shift or a change in the team's behaviour. Both deserve a conversation.
Why the tool isn't the subject.
We're often asked which CRM to choose. On this particular problem the answer barely matters. A properly defined pipeline works in a spreadsheet; a badly defined one stays wrong in the most expensive tool on the market. The cost of a CRM isn't the licence, it's the team's time filling it in — and that time is only justified if the data produces a decision.
What we don't do.
We don't configure a mandatory field nobody will ever read. Every field imposed on a salesperson has to feed an identifiable decision. Otherwise it produces low-quality filling-in, and low-quality filling-in contaminates everything else.