How to Forecast Sales: Methods, Formulas & Steps

To forecast sales, follow five steps: (1) pick a time period (month, quarter, or year); (2) gather clean historical data and a list of exactly what you sell; (3) choose a forecasting method that matches your data and sales cycle; (4) run the numbers using that method's formula; and (5) adjust the baseline for known variables like seasonality, price changes, and new product launches, then review and update on a regular cadence. The core idea is simple: multiply what you expect to sell by what you expect to charge, then correct that figure for everything you already know is changing.

The right method depends on your situation. If you have a defined pipeline, use weighted (opportunity-stage) forecasting: forecast = value of open deals x probability of closing. If you have steady history, project forward from past sales with a growth rate. If you are brand new with no history, build a bottom-up forecast from estimated customers x average purchase value, or model off comparable businesses. Below you will find each method's formula, worked examples, how to do it in Excel, and how to forecast when you have no past data at all.

What is a sales forecast (and what it isn't)?

A sales forecast is a data-driven estimate of the revenue you expect to generate over a specific future period. It answers one question: how much will we realistically sell between now and a chosen date? A good forecast is built from inputs you can defend, units, deals, conversion rates, average prices, not a number you wish were true.

Two distinctions trip people up. First, a forecast is not a goal or a quota. A quota is what you want a rep to hit; a forecast is what you honestly expect to happen, which is often lower. Second, forecasting is not the same as pipeline management. Pipeline management is the day-to-day work of moving individual deals forward; forecasting is the big-picture projection of where those deals, plus everything else, will land in aggregate.

How do you forecast sales? (5 steps)

Whatever method you choose, the workflow is the same. These five steps turn raw data into a number you can plan a budget around.

1. Define the time period and unit. Decide whether you are forecasting a week, month, quarter, or year, and at what level: total company, product line, region, or rep. Forecasting by category (for a restaurant, drinks separately from entrees) is almost always more accurate than one company-wide guess.
2. Gather and clean your data. Pull historical sales by product, subtract returns and cancellations, and list every good or service you sell with its price. Your forecast can only be as good as the data behind it, so dirty CRM records or missing months will quietly wreck the result.
3. Choose a forecasting method. Match the method to your data and sales cycle (see the methods section below). A SaaS team with a CRM should weight its pipeline; a retailer with years of seasonal data should use a moving average or time-series model.
4. Run the calculation. Apply the formula for your chosen method to produce a baseline number for each period and category.
5. Adjust for variables, then review. Layer in what the raw math cannot see, seasonality, a price increase, a new product launch, a big marketing push, an economic shift, and round conservatively. Then set a cadence (weekly, biweekly, or monthly) to compare forecast vs. actuals and refine.

What are the main sales forecasting methods?

Forecasting methods fall into two families. Quantitative methods use numbers and statistics (history, regression, time series). Qualitative methods use informed judgment (expert panels, rep intuition) and are used when hard data is thin. Most strong teams blend the two. Here are the methods worth knowing, with the formula each one runs on.

• Historical forecasting: assume the next period mirrors the same period last year, optionally adjusted for a growth rate. Formula: Forecast = last year's sales x (1 + growth rate). Best for stable, seasonal businesses; weak during rapid change.
• Weighted pipeline / opportunity-stage: multiply each open deal by its stage's probability of closing. Formula: Forecast = value of deals in pipeline x close rate. Best for teams with a defined pipeline and a CRM.
• Lead-driven forecasting: work forward from the top of the funnel. Formula: Forecast = number of qualified leads x conversion rate x average deal size. Best for teams with reliable funnel metrics.
• Bottom-up forecasting: aggregate estimates from each rep, product, or region into a company total. Formula: Forecast = sum of (average deal size x deals per rep). Realistic and field-informed, but can drift optimistic.
• Top-down forecasting: start from market size and work down. Formula: Forecast = total market size x target market share. Fast for planning, but disconnected from frontline reality, best paired with bottom-up.
• Moving average: average the last N periods to smooth out noise. Formula: Forecast = (period 1 + period 2 + ... + period N) / N. Good for spotting trends through seasonal swings.
• Regression analysis: model how a driver (ad spend, price, season) mathematically predicts sales. Formula: Sales = a + bX. Powerful when you know what moves your numbers.
• Time-series models (e.g., ARIMA / exponential smoothing): decompose history into trend, seasonality, and noise to project forward. Highly accurate with rich, consistent data.
• AI / machine-learning forecasting: algorithms learn complex patterns from structured and unstructured data and update as new data arrives. Increasingly built into modern CRMs.

What is the sales forecast formula?

There is no single formula, the right one depends on your method, but two are the everyday workhorses.

Pipeline (deal-based) forecast: Sales forecast = total value of open deals x weighted close rate. If you have $500,000 of open deals and your historical close rate at the current stages is 30%, your forecast is $150,000. Apply different probabilities per stage for more precision, a deal at 'proposal sent' is far likelier to close than one at 'first call.'

Run-rate (history-based) forecast: average monthly sales rate = total revenue so far / number of months so far; then forecast = that rate x months remaining, added to revenue already booked. This assumes sales stay roughly stable, so use it as a baseline and adjust for growth or seasonality.

Can you walk through a sales forecast example?

Here are three worked examples using the run-rate method, the simplest place to start.

1. New farm seeking a grant: By May, the business has booked $5,000 this year. $5,000 / 5 months = $1,000/month average. $1,000 x 7 remaining months = $7,000. Annual forecast = $5,000 + $7,000 = $12,000.
2. Coffee shop, two categories with inflation: Last year, food earned $45,000 and coffee $55,000, with 0.5% inflation. Food forecast = $45,000 + ($45,000 x 0.005) = $45,225. Coffee forecast = $55,000 + ($55,000 x 0.005) = $55,275. Forecasting each category separately gives a sharper picture than one blended number.
3. SaaS pipeline example: You have $800,000 of open deals, but they sit at different stages, $300,000 at 20%, $300,000 at 50%, and $200,000 at 80%. Weighted forecast = (300k x 0.2) + (300k x 0.5) + (200k x 0.8) = $60k + $150k + $160k = $370,000. This stage-weighting is why a pipeline forecast beats simply assuming every open deal closes.

How do you forecast sales for a new business with no historical data?

This is the hardest case and the most common question, because you cannot project from a past you do not have. The trick is to build the forecast from the bottom up out of assumptions you can defend, and to triangulate from outside data.

• Build a bottom-up unit model: estimate how many customers you can realistically reach in a period, then multiply by average purchase value. Forecast = estimated customers x average order value. Ground each input, foot traffic, ad clicks, outreach capacity, in a real constraint, not a hope.
• Use comparables (analog forecasting): find similar businesses in your industry, location, or model and borrow their early ramp, revenue per location, per seat, per square foot, then adjust for your differences.
• Run a test market: launch to one region or segment, measure real results, and extrapolate. Forecast = (test-market sales / test-market share) x total market. This replaces guesses with evidence before you scale.
• Apply the Delphi method: poll several experienced people (advisors, would-be customers, industry veterans) for independent estimates, then converge over a couple of rounds. Useful when no numbers exist yet.
• Keep it conservative and scenario-based: build low, expected, and high cases rather than a single number, and revisit monthly as your own real data starts to accumulate, at which point you switch to data-driven methods.

How do you forecast sales in Excel?

A spreadsheet is enough to start, and you do not need to be a statistician. Three approaches cover most needs.

1. Simple growth model: list historical periods in one column and sales in the next, then project forward with a formula like =previous_cell*(1+growth_rate). Change the growth rate cell to run scenarios instantly.
2. Moving average: use =AVERAGE() over the last three to twelve periods to smooth seasonality, then carry that average forward as your baseline.
3. Built-in forecasting: highlight your date and sales columns and use the FORECAST.ETS function or the Data tab's 'Forecast Sheet' button, which applies exponential smoothing and even draws confidence intervals automatically. For pipeline forecasting, build a deal table with value and stage-probability columns and use =SUMPRODUCT(value_range, probability_range) to get a weighted total.

What makes a sales forecast accurate (and what breaks it)?

Even good methods produce bad forecasts when the inputs or habits are wrong. Accuracy comes less from a fancier model than from clean data and disciplined process. Watch for these accuracy killers and fixes.

• Dirty or incomplete data: the single biggest cause of bad forecasts. A forecast is only as reliable as the CRM behind it, so enforce data hygiene and require reps to keep deal stages current.
• Happy ears (optimism bias): reps and managers routinely overrate deal probability. Weight stages from historical close rates, not gut feeling, and track each rep's forecast vs. actuals to calibrate.
• Ignoring external shocks: seasonality, price changes, new competitors, tariffs, or economic shifts can flip a forecast overnight. Build these in deliberately and forecast more frequently in volatile periods.
• Set-and-forget: a forecast made once a quarter and never revisited goes stale fast. Update on a fixed cadence and measure accuracy each cycle (e.g., with mean absolute percentage error) so the process improves over time.
• Over-reliance on one method: blend quantitative and qualitative views, and consider layered forecasting, reps submit, managers adjust, leadership applies strategic context, to balance frontline detail with the big picture.

What tools help you forecast sales?

You can start in a spreadsheet, and many small teams should. But spreadsheets stop scaling once deal volume grows: they break, they go stale, and they depend on someone manually updating them. A CRM centralizes deal data, applies stage probabilities automatically, and keeps the forecast live as deals move, which is why forecasting is consistently more accurate in a CRM than a spreadsheet once you are past the earliest stage.

Modern CRMs increasingly layer AI on top, detecting patterns, flagging at-risk deals early, and updating projections in real time as new data arrives. An all-in-one platform like MapleConnect, for example, combines a CRM with built-in pipeline tracking, AI, and automation so the data feeding your forecast stays current without extra manual work. Whatever tool you choose, the principle holds: the technology does not replace judgment, it just makes the inputs cleaner and the updates faster so your forecast stays trustworthy.