How to forecast sales from the data you already have

A sales forecast is your best estimate of how much you'll sell in a coming period: next month, next quarter or the rest of the year. Founders plan hires from it, managers set targets with it, and investors ask for it before anything else about the quarter.

A forecast isn't a target. The target is what you want to sell. The forecast is what your deals and your activity say you will sell, and the gap between the two is the part worth acting on.

A forecast doesn't need to be exact to be useful. It needs to be close, honest about how sure it is, and early enough to act on. A forecast that says in the first week of March that the quarter will come in short gives you three weeks to fix it. One that says it on the 29th is a report.

Big vendors sell forecasting software with AI models and dozens of inputs. A team of 3 to 30 can do well with a spreadsheet and three methods, each checking the others.

Here's the whole job, in order:

1. Pick the period. A month suits most small teams. Use a quarter if your deals take months to close.
2. Pull what you sold in the same period last year, and in each of the last few months.
3. List every open deal with its value, its stage and the date it's expected to close.
4. Work out your own rates: how often deals close from each stage, how many calls book a meeting, how many meetings become a deal.
5. Run the three methods and put the numbers side by side.
6. Give the forecast as a range with a most likely figure.
7. Check it every Monday, and when the period ends, compare it with what you sold.

If you'd rather not build the sheet yourself, the sales forecast template weights the month's deals by your own stage chances and shows the gap to target.

The simplest forecast takes what you sold in the same period before and adjusts it for what's changed.

It's quick, and it's the right starting point when your sales are steady. It misses anything new: a big deal in the pipeline, a slow month of prospecting, a price change. That's what the next two methods catch.

With two years or more of monthly sales in a spreadsheet, Excel can run this method for you. Put the months in one column and the sales in the next, select both, then pick Forecast Sheet in the Forecast group of the Data tab. Behind the button, Microsoft says, is a method called exponential smoothing. It spots a seasonal pattern on its own and draws a range around each future month, at 95% confidence unless you change it. It needs at least two full cycles of history to find the pattern, so a yearly one takes 24 months of data.

If your sales climb or fall in a straight line, the FORECAST.LINEAR function fits that line through your past months and reads next month off it. Both read the past, so both share this method's blind spot.

Take every open deal expected to close in the period, and multiply its value by the chance it closes at the stage it's in. The chance should come from your own history: of the deals that reached this stage before, how many closed?

The pipeline value page has the formula and a calculator. Two rules keep the number honest. Use your own close rates, not a vendor's default percentages. And ask reps to move a deal's close date when it slips, not leave it sitting in October.

The third method works forward from what the team is doing. It needs three numbers from your own past: how many calls it takes to book a meeting, how many meetings turn into a deal, and your average deal size.

This method is the early warning. If calls drop in the first two weeks of a month, the activity forecast drops with them, weeks before the pipeline shows it. It's also the one a manager can do something about this week.

Lay the three numbers side by side for the same period:

When the first two agree within a tenth of each other, forecast somewhere between them and say how sure you are. When they don't, the gap is the conversation. A pipeline well under your past months means the team hasn't built enough deals. Pipeline coverage tells you how much open pipeline the target needs.

Give the forecast as a range with a middle figure, like "$120,000 to $135,000, most likely $125,000", because a single number sounds more certain than any forecast is.

Forecasting guides list more methods than these three, and some of them need years of records before they're any use. Here's what each one does, and when a small team has a use for it.

The rep's judgement deserves one more line. It's the method every team uses without calling it one, and it's the one that drifts most with mood. Ask for it, write it down, and compare it with the weighted pipeline each week. A rep whose number always beats the pipeline is telling you something about their deals or about their optimism.

A new business, a new product or a new rep has no past months and no close rates of its own. You can still forecast, as long as every number is labelled as a guess and replaced as soon as you have real ones.

1. Write the guesses down: calls a week, how many book a meeting, how many meetings become a deal, and the average deal.
2. Work for two to four weeks and count what happens.
3. Swap each guess for the rate you counted.
4. Forecast one month ahead, not a year, until you have a quarter of your own numbers.

If the forecast is for a business plan, the U.S. Small Business Administration's business plan guide asks for a five-year outlook, with the first year in quarterly or even monthly projections, each one explained. Show the guesses behind the number, because a lender can judge 30 calls a week and one meeting in six, and a bare revenue figure tells them nothing.

When a forecast misses, check the inputs before you check the sum. These six go wrong on small teams all the time, and each one has a fix you can make:

• Dead deals stay open. A proposal with no reply in six weeks isn't a deal. Close it, or push its date out.
• Stages move on hope. A deal moves to "proposal sent" when the proposal goes out, not when the call went well.
• One big deal swings the month. Forecast without it, and put it on a line of its own.
• The season gets ignored. Check the same month last year before trusting the last three.
• Close dates never move. A deal promised for October that slips still sits in October, and October looks better than it is.
• Nobody checks. Without comparing the forecast with what you sold, the same mistake repeats every month.

On a team, each rep also reads the stages a little differently. Agree in writing what has to happen for a deal to reach each stage, so "verbal yes" means the same thing on everyone's list.

A monthly forecast made once is out of date by the second week. Five minutes every Monday keeps it useful:

1. Update the weighted pipeline.
2. Compare last week's calls and meetings with what the activity method assumed.
3. Write the new range next to last week's, and one line on what changed.
4. If the range has dropped below target, decide this week what to do about it: a call block, help on a stuck deal, a push on proposals out.

When the month ends, write down the miss. A team that said $125,000 and sold $110,000 was $15,000 high. Keep that one line for every month in the same sheet as the forecast itself, and after a quarter the misses show you which way your method leans.

After a few months, if you're always high, your close rates are too generous. If you're always low, your reps are sandbagging, or the team is better than your history says.

• Microsoft Support, Create a forecast in Excel for Windows, read 27 October 2026
• Microsoft Support, FORECAST and FORECAST.LINEAR functions, read 27 October 2026
• U.S. Small Business Administration, Write your business plan, read 27 October 2026