Sales Forecasting Methods Compared: Which One Fits Your Business

Run rate

Average the last three months of closed revenue, adjust for seasonality and any known change in capacity, and publish it. Its great virtue is that it cannot be gamed by optimism in the pipeline, because it never looks at the pipeline. Its weakness is that it is blind to anything new: a large deal about to land, a new product, a channel that just started working.

Weighted pipeline

Multiply each open deal value by the historical conversion rate of its stage and sum the result. This is the standard method and it is only as good as two inputs: whether your stages mean the same thing to everyone, and whether the percentages come from your own history. Get either wrong and it produces a number that looks analytical and is essentially decorative.

Commit based

Ask each salesperson what will close, then hold them to it over time. In small teams this works surprisingly well because the deals are individually visible and exaggeration is socially costly. The critical rule is separation: a commit is an estimate and a target is a goal, and the moment leadership treats a commit as a promise, commits become negotiations and their information value disappears entirely.

Cohort or source based

Work backwards from enquiries. If a given source produces a stable enquiry to customer ratio and a stable average value, then this month enquiry count forecasts revenue a cycle from now. This is the most useful method for marketing-led businesses because it forecasts far enough ahead to be actionable, and it degrades gracefully as long as lead quality is stable.

If your sales cycle is under two weeks and volume is steady, use run rate and stop there. If your cycle is one to three months and your stages are honest, use weighted pipeline with your own percentages. If you have fewer than about thirty open deals at a time, use commit forecasting and skip the mathematics, because at that volume statistics are noise and individual knowledge is signal. If your revenue depends on marketing spend, add source based forecasting alongside whichever you chose, because it is the only method that gives you warning.

Publish a range

Three numbers: a low case that assumes only committed deals land, a base case, and a high case. As an illustrative example, a forecast published as somewhere between twelve and eighteen with a base of fifteen is far more useful for planning than a confident fifteen, because it tells the person making a stock or hiring decision how much room to leave.

Exclude the deals nobody is working

Any deal with no dated next action should be excluded from the forecast entirely. It is not being progressed by anyone, so its inclusion is a statement about hope rather than about work in progress. This single rule usually removes a meaningful chunk of the pipeline and improves accuracy immediately, which tells you something uncomfortable about the pipeline.

Handle the big deal separately

In a small business one deal can be a large share of the quarter. Averaging it into a weighted total hides the fact that your forecast is really a bet on one signature. Report the base separately from a named list of large deals, each with its expected date and what it is waiting on. Then the conversation becomes about the specific dependency rather than about a percentage.

Keep a simple record: what you forecast for each month, what actually happened, and the difference as a percentage. After six months you will know two things that matter more than the method you chose. First, your bias, meaning whether you consistently run high or low, which can simply be subtracted. Second, your spread, meaning how wide the honest range is.

That record also settles the recurring argument about whether the forecast is any good, because it converts a matter of opinion into a short table. Whatever system holds your pipeline should make this easy: a snapshot of the forecast at the start of each month, compared with closed revenue at the end of it.

Related reading: what a CRM is, lead management software, sales automation, reporting features, CRM for small business, and use cases.