What Is a Sales Forecast? A Plain Guide for Founders
What is a sales forecast? A statement of the revenue you expect to close, built to guide decisions, wrong in a known direction by an amount you can measure.
A sales forecast is a statement about the revenue you expect to close, built from the deals you can see today, and its job is to tell you what the company can afford to commit to. It is not a report on how sales is going, and it does not have to be right. It has to be wrong in a direction you know, by an amount you can measure.
That second sentence is the whole idea. Most of what follows is about why it matters and what has to be true for it to hold.
What is a sales forecast, and what is it for?
Ask most teams what their forecast is for and you hear that it gives visibility. That is the weakest possible answer. Visibility is what a report gives you, and a report is judged on whether it looks good.
A forecast answers a different question: what can we afford to commit to? Every hire, every marketing budget and every timeline for money rests on an assumption about revenue that has not arrived yet. The forecast is the place where that assumption lives. Because it feeds so many expensive choices, a soft forecast makes every one of those choices soft, and the people making them often do not know it, because the number arrived looking like a fact.
So there are two standards you can hold a forecast to.
| Treated as | Judged on | What tends to happen |
|---|---|---|
| A report | Whether it looks good | The number drifts towards what people want to hear |
| A decision instrument | Whether the decisions made from it held up | The number is checked against outcomes and corrected |
The second standard is harder, and it is the only one that matters.
Why most forecasts are optimism with decimal places
The usual method goes like this. Someone adds up the deals that feel likely, multiplies a few of them by a percentage that came with the software, rounds towards the number the board wants, and presents the result with a precision it has not earned. The period ends, the number is missed, and the explanation is always the same: two big deals slipped.
They did. They were always going to. Nothing in the way the forecast was built could have told anyone that in advance.
Notice what the problem is not. The forecast being wrong is not the problem, because every forecast is wrong. It is a statement about the future made with partial information. The problem is that nobody knew how wrong, or in which direction, or which deals carried the error.
Compare two kinds of forecast:
- One is reliably ten percent optimistic. You discount it and steer. That is a useful instrument.
- One is sometimes right, sometimes badly high and occasionally low, for reasons nobody can explain. That is not an instrument at all. It is weather.
The six jobs a trustworthy forecast does
A disciplined forecast earns its place through six mechanisms. None of them is simply better visibility. Each is something the company either cannot do without it or does badly.
| Job | The idea in one line |
|---|---|
| Decision instrument | It answers what you can afford to commit to, not how sales is going |
| Probability by stage | Probability belongs to the stage a deal has reached, not to anyone's mood |
| Two different numbers | The weighted pipeline and the commit answer different questions |
| Time | A real deal that closes next quarter still wrecks this quarter |
| Lost deals | Every loss comes with a reason, and the reason is cheap data |
| The bridge to cash | A deal earns its way from hoped to probable to contracted |
Each row deserves its own treatment, and several have one. The short versions follow.
Probability belongs to the stage, not the mood
Ask a salesperson how likely a deal is and you get a reading of their last conversation. The person closest to a deal is the person most attached to it, so their judgement is optimistic. Ask the pipeline instead: of all the deals that ever reached this stage, how many closed? That answer is far more stable. It will be wrong about any single deal and surprisingly accurate across a larger group. How to build those numbers is covered in how to set stage probabilities.
Two numbers that answer two questions
Every open deal multiplied by its stage probability gives the weighted pipeline. It tells you what the pipeline is worth on average over time, which makes it the right number for asking whether you have enough pipeline for next quarter. It is the wrong number for predicting this month, because a month holds too few deals for averages to behave. For the near term you use the commit: the specific deals you would bet on, named one by one. The difference is explained in weighted pipeline versus commit. Founders who blur the two either panic about a healthy pipeline or hire against an unhealthy one.
Time is the variable most forecasts leave out
A deal can be real, qualified and genuinely going to close, and still ruin the forecast by closing next quarter instead of this one. Slippage, not loss, is the most common reason forecasts miss, and it is almost invisible if you only track value and stage. A forecast that ignores close dates and deal age is forecasting whether, when the company needs to know when. Cash does not arrive in the quarter a deal deserved to close in. The practical response is in how to handle slipping deals.
Lost deals are the cheapest data you own
Every lost deal comes with a reason, paid for in time and attention. Collected consistently, the reasons show whether you lose on price, timing, fit, a competitor or plain indecision, and each of those has a different fix in a different part of the company. Collected carelessly, they evaporate and the company keeps losing the same way while believing it loses at random.
The bridge between sales and cash
If you plan cash in three honest buckets, contracted, probable and hoped, and build the runway on the first two, then the forecast is the discipline by which a deal moves out of hoped and into probable. Without it, the cash plan either ignores the pipeline entirely, which is too pessimistic, or quietly includes all of it, which is how runway numbers lie. The mechanics are in how to turn a sales forecast into a cash plan.
The honest limit: statistics need numbers
Forecasting is statistics, and statistics need volume. With a dozen open deals, a weighted pipeline is a polite fiction, because no deal closes for a fraction of its value. One large deal can outweigh everything else and make every average meaningless. Early probabilities are set from thin history and will be wrong.
None of that is a reason not to forecast. It is a reason to forecast humbly:
- Name the big deals individually, rather than letting them hide inside a weighted total.
- Write down which probabilities are guesses.
- Treat the first few quarters as calibration rather than prediction.
The discipline pays off slowly and then all at once, usually in the quarter you most needed it.
What has to be true before the number means anything
A forecast is built on definitions, kept alive by a routine and improved by a comparison with what happened. If any of the three is missing, the number still appears on the page with the same confidence as before, which is the dangerous part.
| Layer | What it contains | Without it |
|---|---|---|
| Definitions | Stages defined by something the buyer did, stage probabilities from your own history, a written rule for what counts as commit, a buyer's close date, a defined lost reason list | People argue about the word commit instead of the deals |
| Routine | A short weekly review that looks at what changed, walks the commit and checks the age of deals | The pipeline goes stale and nobody notices |
| Comparison | A quarterly check of the forecast against what actually closed | Confidence accumulates but accuracy does not |
A stage defined by a fact, such as a discovery call that was held or a proposal that was reviewed with the buyer, can be checked. A stage that can be entered because the seller feels hopeful will see its probability drift upwards until it means nothing.
How to tell a forecast is working
You will not know from the number itself. You will know from what happens after it is produced.
- You can say how far commit sat from what actually closed, and in which direction.
- You know whether each stage converted at roughly the rate its probability claimed.
- Decisions have changed because of it. A hire was delayed, or a marketing push was started early, because of what the number showed.
The forecast earns its keep in the quarter after the review, when a decision went differently because of it. The method for checking it is set out in how to calibrate a sales forecast.
Where to start if you have nothing yet
You do not need software to begin. Everything described here can be done with a spreadsheet and a disciplined half hour a week. Start with the definitions, because they cost nothing and they decide whether later numbers mean anything: define the stages, label your probability guesses as guesses, write down what commit means, and give every open deal a close date that is the buyer's own.
The forecast is meant for a company past the first handful of customers, with a motion and a pipeline that has real stages and enough deals that predicting them is worth the effort. If you are earlier than that, the work is finding customers, not forecasting them.
Frequently asked questions
Does a sales forecast have to be accurate?
No. It has to be wrong in a direction you know, by an amount you can measure. A forecast that is reliably a little optimistic can be discounted and steered by. A forecast that is wrong at random cannot be used for anything.
Is a sales forecast the same as a sales target?
No. A target is what you want to happen. A forecast is what the deals you can see suggest will happen. Rounding the forecast towards the target is exactly how a forecast turns into optimism with decimal places.
Why do forecasts keep missing even when the deals are real?
Mostly because of timing. A deal that is qualified and genuinely going to close can still land in the next period, and a forecast that tracks only value and stage cannot see that coming.
Can I forecast with only a few open deals?
You can, but humbly. With a small number of deals the weighted figure is not reliable, so name the large deals one by one and treat the early quarters as calibration rather than prediction.
The full method, with the checklists, is in The Sales Forecast Playbook, a free 33-page playbook.