Sales forecasting

Why it matters

Most decisions that go wrong in a growing business go wrong because of a revenue assumption. Hiring a second technician for work that arrives three months late, buying stock for a season that is softer than expected, or taking on a lease on the strength of a verbal yes. A written forecast with known accuracy turns those gambles into managed risks, and it is the bridge between your sales pipeline and your cash flow budget.

Three methods, and when each fits

The historical method projects forward from past results: last year’s same quarter adjusted for growth and known changes. It works well for businesses with many small transactions and seasonal patterns, such as retail, hospitality and trades with steady residential demand, where individual deals are too small to track one by one.

The pipeline method, usually weighted by stage, adds up the value of open deals multiplied by the probability of winning from their current stage. It suits B2B businesses with fewer, larger deals and a defined sales process. The judgement method asks the person closest to each deal to call it: commit, likely, possible or unlikely. It is quick and often surprisingly good, but depends on honesty and gets harder as deals multiply. Most growing businesses combine two: pipeline or judgement for the committed deals, history for the run-rate of small repeat business underneath.

Getting the probabilities right

Stage probabilities should come from your own data, not from software defaults. Take the last six to twelve months of closed deals and ask: of all the deals that reached "proposal", what share eventually won? Repeat for each stage. The numbers will be lower than you expect, and that is the point. A forecast built on honest probabilities is conservative where your instinct is optimistic.

Refine with segments once you have enough deals. Win rates often differ sharply between new and existing customers, between deal sizes, or between lead sources. A proposal to an existing client might win three times in four; the same proposal to a cold prospect might win one in four. If the difference is that large, forecast the two groups separately.

A forecasting rhythm

Forecasting is a habit more than a calculation. Update the forecast weekly as part of the pipeline review, when close dates and stages change. Produce a formal monthly forecast for the next three months that feeds the cash flow budget, and a quarterly view for the year ahead that drives hiring and investment decisions. Each time, record what you forecast so that later you can compare it with what happened.

Keep three versions in view: commit (deals the buyer has verbally agreed, plus the contracted run-rate), likely (commit plus weighted pipeline) and upside (everything that could possibly land). Plan costs against commit, plan capacity against likely, and treat upside as the case for extra prospecting rather than extra spending.

Measuring and improving accuracy

The only way to improve a forecast is to measure how wrong it was. Each month compare the forecast you made for that month with actual revenue won, and record the error as a percentage. Then look at why: deals that slipped to the next period, deals lost that were forecast as likely, deals that landed that were not in the pipeline at all. Each cause points to a different fix.

Slipping deals mean close dates are set by hope; ask buyers what has to happen by when. Losses forecast as likely mean probabilities are too generous or stages are advanced too early. Deals landing from nowhere mean leads are not being entered until late, which hides pipeline risk. Within two or three quarters most businesses get their monthly forecast within a range they can plan around.

How to build a sales forecast

1. Separate repeat revenue from new deals Contracted, subscription or steady repeat revenue is forecast from history; new deals from pipeline.
2. Calculate stage win probabilities from your own closed deals Six to twelve months of data. Round the results and note the sample size.
3. Weight each open deal Value × stage probability, including only deals with a realistic close date in the period.
4. Add the run-rate baseline Last year’s equivalent period, adjusted for growth, price changes and known customer losses.
5. Produce commit, likely and upside views Plan spending on commit, capacity on likely, prospecting effort on the gap to target.
6. Record the forecast and compare monthly Track error and its causes. Adjust probabilities and date discipline accordingly.

Real-world example

A Melbourne recruitment agency forecasts a hiring decision

Latitude Recruitment places finance and administration staff for mid-sized Melbourne businesses. The owner, Grace, wants to hire a third consultant but a bad quarter would make the salary a strain. She has a CRM with two years of closed deals and a steady base of repeat clients worth about $45,000 a month in fees.

From her history she finds that roles reaching "shortlist presented" fill about 60 per cent of the time, roles at "brief taken" about 35 per cent, and "qualified enquiry" about 15 per cent. Her open pipeline for the next quarter weights out at roughly $95,000 of new-client fees. Add the repeat baseline of $135,000 and her likely forecast is about $230,000 for the quarter, against a commit view of about $175,000. The new consultant costs around $35,000 a quarter fully loaded. Because even the commit view covers existing costs and most of the new salary, and the likely view covers it comfortably, she hires, with a clear plan to lift prospecting if the pipeline weight drops below $80,000 at any monthly check.

Common mistakes

• Using software default probabilities Vendor defaults are guesses about an average business. Calculate yours from your own closed deals.
• Forecasting the target A forecast that always equals the target is not a forecast. Let the number be uncomfortable; that is its job.
• Counting lifetime value in the period A three-year contract does not land in the quarter it is signed. Forecast the revenue that will actually be invoiced in the period.
• Never checking accuracy Without a monthly comparison of forecast and actual, probabilities never improve and the same errors repeat.
• Ignoring seasonality Many Australian businesses see January and the weeks around end of financial year behave differently. Build it into the baseline.

KPIs to track

• Forecast accuracy Actual revenue divided by forecast for the same period, tracked monthly. Many small businesses aim to get within 10 to 20 per cent.
• Slip rate The share of deals forecast to close in a period that moved to a later one. High slip means close dates are not grounded in buyer commitments.
• Commit conversion How often deals flagged as commit actually close in the period. It should be very high; if not, commit is being used loosely.
• Pipeline weight versus target Weighted pipeline for the period compared with target. The gap tells you how much prospecting is needed now.

Frequently asked questions

How far ahead should a small business forecast?

Three months in detail, twelve months in outline. Beyond a quarter, pipeline becomes guesswork for most businesses and the historical run-rate carries the forecast.

What accuracy is realistic?

It depends on deal size and volume. Businesses with many small sales can often forecast within a few per cent; businesses with a handful of large deals will see bigger swings. The goal is a known error range you can plan around, not perfection.

Should the forecast feed my budget?

Yes. The commit view should underpin cost commitments in the cash flow budget, with the likely view used for capacity planning. Keep the two documents connected so a change in one shows up in the other.

Can AI tools forecast for me?

Some CRMs offer predictive forecasting that learns from your history. They can help once you have a few hundred closed deals of clean data; before that, a simple weighted pipeline you understand is more reliable.

Key takeaways

• A forecast is a prediction from pipeline and history, not a target or a hope.
• Use run-rate history for repeat revenue and weighted pipeline for new deals.
• Calculate stage probabilities from your own closed deals and review them quarterly.
• Keep commit, likely and upside views, and plan costs against commit.
• Measure forecast error monthly and fix the cause: slipping dates, generous stages or late-entered leads.