How to build a reliable sales forecast

A sales forecast answers one question: how much revenue are we likely to close this month or this quarter? It is not a target. A target is what you want. A forecast is what you expect, based on what is in the pipeline right now.

That sounds obvious, but in many teams the two blur together. A familiar pattern: in January leadership names a number for the quarter. The team arranges the pipeline so the number looks reachable. By mid February half of those deals have not moved for weeks, and in March people push hard on deals that are not ready to close. With discounts, and with a buyer who feels rushed.

A good forecast has three qualities. It is repeatable: someone else gets the same number from the same data. It is explainable: you can say per deal why it counts. And it is updated every week, not once a quarter.

The base is always the pipeline. Every deal has a value, a stage and an expected close date. Without those three fields you cannot forecast, only voice a feeling. That is why this guide starts with data hygiene, not formulas. Then come the methods, a worked example, pipeline coverage and a weekly routine.

Remember that a forecast is a range, not an exact number. The most useful conversation is about the gap between what is certain and what is possible, and which deals decide that gap. A sales manager who says "we will land between 80 and 95, and deals B and E decide where" has more to work with than someone who says "92".

A forecast is not a verdict on reps either. A number below target is information. It tells you to act now, with more leads, help on a big deal or an honest conversation about the target. The earlier you know, the more you can do.

The weighted forecast in this example is € 95,750. But look at what is behind it. Deal C contributes € 24,000 while it is only in the demo stage with two stages to go. If that deal slips a quarter, the forecast drops by a quarter. That is the weakness of a weighted pipeline with few deals: one large deal can decide the total.

So you put the categories next to it. Owners place their deals in three groups:

• Commit: the buyer has said they will sign and only details are being discussed. In the example deals B and E, together € 55,000.
• Best case: the deal can close this quarter if things go well. Deal A, € 40,000. Commit plus best case is € 95,000.
• Pipeline: the deal is active, but closing this quarter is unlikely. Deals C and D.

Now you have a range. The floor is € 55,000, the weighted number € 95,750, and commit plus best case € 95,000. In this example the last two are close. That is a good sign: the win probabilities and the reps' judgement say roughly the same thing.

Then add the historical run rate. Suppose the team closed € 70,000, € 85,000, € 90,000 and € 78,000 in the last four quarters. An average of € 80,750. A forecast of € 95,750 is higher than usual. Is that possible? Yes, if something changed, such as more leads or a new product. If nothing changed, the win probabilities are probably too optimistic.

So the number you send to leadership in this example is not 95,750 but a range with an explanation: "between 55 and 95, expected around 85, and deal A matters most". That is a number you can defend.

A forecast says what you expect to close. Pipeline coverage says whether there is enough in the pipeline to reach your target at all. You calculate it as open deal value with a close date in the period, divided by the target.

A made up example. The target for the quarter is € 100,000. Open pipeline with a close date this quarter is € 170,000. Coverage is 1.7.

Is that enough? It depends on your own win rate by value. Suppose you historically win 45 percent of open deal value in a quarter. Then you need coverage of 1 divided by 0.45, roughly 2.2. At 1.7 you are below that. The expected outcome is 170,000 times 0.45, so € 76,500.

You see that gap at the start of the quarter, not at the end. Then you can still act: follow up more leads, pull deals forward from next quarter where that is realistic, or adjust the target. Calculate coverage per month and per rep, so you also see who needs help.

There is no fixed rule for the right coverage. You set it from your own numbers. The guide on sales KPIs explains how to measure win rate and cycle time.

A forecast only becomes reliable when you compare it afterwards with what actually happened. Do that every period, and keep it simple.

A made up example. In week 1 of the quarter the expectation was € 95,750. At the end, € 82,000 was closed. The deviation is 95,750 minus 82,000, divided by 82,000. That is about 17 percent too high.

Then look at where the gap came from:

• Deals that slipped. Deal C did not close this quarter and is still open. That tells you something about close dates.
• Deals that were lost. A deal in Negotiation was lost after all. That tells you something about the exit criterion of that stage.
• Deals that were not there. A deal that did not exist in week 1 closed within the quarter. That tells you something about your cycle time.

Track it per quarter in a simple list: expectation in week 1, week 5 and week 9, and the outcome. After three or four quarters you see whether you are consistently too high or too low. Then adjust win probabilities per stage, or tighten the exit criteria.

If you sell subscriptions, retainers or service contracts, you need two numbers and you should not mix them up.

Contract value. What the deal is worth over an agreed period, for example a year. This is usually the deal value in your pipeline and the number you compare reps on.

Revenue in the period. What actually comes in from that deal this quarter. An annual contract of € 12,000 that starts in mid February brings in roughly € 1,500 in the first quarter.

A made up example. A team signs three subscriptions of € 1,000 a month each in March. Contract value is € 36,000 per year. Revenue in the first quarter is zero, because the subscriptions start in April.

So agree which number your forecast is, and label it in the report. For the sales meeting, contract value is usually the most useful. For finance, revenue in the period is what counts. Two reports side by side are better than one number that means something different to everyone.

Treating the forecast as the target

If the team knows the forecast has to equal the target, it will. Keep the two separate and discuss the gap.

Close dates that are never right

A close date that moves a month every month makes every forecast too high. Agree that a deal pushed twice goes back a stage or gets a conversation with the manager.

Win probabilities that never change

Template win probabilities are a start. Adjust them to your own numbers after six months. Otherwise you calculate with numbers that are not about your team.

A few large deals

With few deals, one large deal decides the total. Look at the five largest deals separately in every forecast and discuss them one by one.

Counting recurring revenue differently

If one rep enters an annual contract as one month and another as three years, no forecast adds up. Agree on one deal value rule. See how to set up a sales pipeline.

Looking only at the total

A total that adds up can hide one rep far above target and two far below. Look per rep and per pipeline as well.