What should you track in your sales pipeline?

A forecast needs five things per deal: who owns it, what it is, what it is worth, the odds of it closing, and the month it should close in. That is the whole list. Everything else in your CRM is reporting or process, and every extra field you make mandatory lowers the probability that the five that matter are filled in honestly.

The five that carry the forecast

Owner. Not for accountability theatre — because forecast accuracy is a property of a person. You cannot tell whose calls hold up without knowing whose calls they were, and that is the single most valuable thing a pipeline record eventually tells you.

Deal name. Enough to recognise it in a meeting. It exists so that a conversation about the number can become a conversation about a specific thing.

Value. The amount that lands in the period if it closes. Not total contract value if you are forecasting a quarter, which is the most common way a pipeline quietly doubles.

Odds. The owner's own number, not the stage's. A stage probability is a statement about your company's history applied flat to a deal it knows nothing about. A person's number can be scored; a stage's cannot.

Close month. Month, not day. A specific date is false precision in almost every B2B sale, and asking for it trains people to pick the last day of the quarter.

Five fields is roughly the point at which a rep will keep a pipeline accurate during a meeting rather than cleaning it up the night before a review. That is the real constraint: not what would be nice to know, but what will still be true in six weeks.

Useful, but not for the forecast

• Stage is good for diagnosing where deals stall and for coaching. It is a poor forecasting input and a worse probability.
• Source and segment matter for deciding where to spend next year. They tell you nothing about this quarter.
• Next step and next step date are the best activity fields, because both are falsifiable. "Technical review on the 14th" can be checked on the 15th.
• Competitor, lost reason, close date history. Valuable in aggregate, after a year. Worth collecting when you have the discipline, worth skipping when you do not.

The field most pipelines are missing

What actually happened. Most CRMs mark a deal won or lost and move on, overwriting the forecast that preceded it. That destroys the only record that could tell you whether your forecasting works.

Keep one row per deal per month: what the odds were when you called it, and what happened in the end. Twelve months of that answers questions no report can — whether you run hot, whether a particular rep's 70% means 70%, whether deals above a certain size behave differently.

Why fewer fields produce better data

Every required field is a tax paid at the worst possible moment, usually by someone who is mid-sale and does not care about reporting. The predictable result is not blank fields; it is plausible-looking defaults. A pipeline where every deal sits at 50% and every close date is the last day of the quarter is not an empty pipeline, it is a compliant one, and it is worse than empty because it looks like data.

Five fields, filled in honestly during a review while the team talks through each deal, will out-forecast twenty fields filled in on a Thursday night to clear a dashboard warning.