All field notes

How-to ·

Sales forecasting methods compared: how each works, what it needs and when to use it

Seven sales forecasting methods compared: historical, weighted pipeline, forecast categories, cycle length, regression, AI and blends, with an example.

The main sales forecasting methods are historical, weighted pipeline, forecast categories, sales cycle length, regression and AI or predictive models. Most B2B teams blend two: a weighted pipeline checked against the reps' commit call.

Each method answers the same question, how much will we close and when, from a different starting point. Some start from the past, some from the open deals, and some from the reps' judgment. This guide compares them, shows how to choose, and works through one quarter's numbers with four of them side by side.

What the job is

A sales forecast is a prediction of the revenue a team will close in a period, usually the current quarter, with a number for each rep, team and region. Finance uses it to plan spending and hiring, and leadership uses it to tell the board what is coming.

The front-line manager builds the first version each week from the reps' deals, the sales leader rolls the teams up, and RevOps runs the reports and the models. A weekly forecast takes a manager one to two hours, most of it spent checking deals in the forecast call. Choosing a method is a once-a-year decision, revisited when the forecast misses badly two quarters running.

Sales forecasting methods compared

Method How it works Data it needs Best for Main weakness
Historical Takes the same period last year and adjusts for growth Two or more years of closed revenue Stable businesses with seasonal patterns Blind to this quarter's deals
Weighted pipeline Multiplies each deal's value by its stage's close probability Open deals with stage and amount Teams with defined stages Only as good as the stage data
Forecast categories Reps call each deal Commit, Best case or Pipeline Rep and manager judgment Most B2B teams, every week Optimism and sandbagging
Sales cycle length Checks whether each deal can close in the time left Deal age and average cycle by segment Long or variable cycles Ignores what is happening in the deal
Regression Models sales on drivers such as pipeline created or headcount Several years of clean data Annual planning Needs data most teams lack
AI and predictive Learns patterns from CRM, email and call activity Large volumes of clean CRM data Larger teams with good data Hard to explain a single deal's score
Blended Runs two methods and investigates the gap Whatever the two methods need Almost every team Takes discipline to act on the gap

Each method in brief

All seven use the same inputs in different proportions: the past, the open deals and someone's judgment. The differences show up in which deals they trust and how they treat a deal that has stalled.

Seven ways to build the quarter's number
The quarter's forecast
  1. HistoricalLast year's quarter, adjusted for growth
  2. Weighted pipelineDeal value times stage probability
  3. Forecast categoriesCommit, best case and pipeline
  4. Sales cycle lengthCan the deal close in the time left?
  5. RegressionSales modeled on the drivers behind them
  6. AI and predictivePatterns in CRM and activity data
  7. BlendedTwo methods checked against each other
The quarter's forecast
  1. HistoricalLast year's quarter, adjusted for growth
  2. Weighted pipelineDeal value times stage probability
  3. Forecast categoriesCommit, best case and pipeline
  4. Sales cycle lengthCan the deal close in the time left?
  5. RegressionSales modeled on the drivers behind them
  6. AI and predictivePatterns in CRM and activity data
  7. BlendedTwo methods checked against each other

Historical forecasting

Historical forecasting takes revenue from the same period last year and adjusts it for growth, pricing changes and headcount. It is quick and gives finance a stable baseline. It cannot see a large deal that has slipped or a new product that is selling faster than expected, so it works best as a check on the other methods.

Weighted pipeline (opportunity stage)

A weighted pipeline multiplies each open deal's amount by a probability tied to its stage, then adds them up. A $100,000 deal in a stage with a 40% probability counts as $40,000. The probabilities should come from the team's own history: of the deals that reached this stage, what share closed? The method breaks when deals sit in stages they have not earned, because every inflated stage adds weight to the total.

Forecast categories (rep judgment)

Reps place each deal in a forecast category, usually Commit, Best case or Pipeline, and the manager adjusts. A commit forecast is the number the team stands behind. This is the most common weekly method in B2B, and it is only as good as the definition of Commit. "I feel good about it" produces a different number from "the signer has agreed a date and the approval steps are mapped."

Sales cycle length

This method asks whether each deal can close in the time left. If the average cycle for a segment is 110 days and a deal entered the pipeline 40 days ago, it is unlikely to close in a quarter with 30 days remaining, whatever its stage says. It is a useful filter on the other methods, especially for long enterprise cycles.

Regression analysis

Regression models sales against the drivers that seem to cause them: pipeline created, number of ramped reps, marketing spend, price changes. It suits annual planning and capacity models. It needs several years of clean, consistent data, and the relationships it finds can change when the market does.

AI and predictive forecasting

Predictive tools learn from patterns in CRM fields, emails, meetings and call activity, then score each deal and roll the scores into a forecast. They are fast and can flag deals the team has overlooked. They inherit the quality of the CRM data, and a deal score rarely tells a rep what to do next on that deal.

Blended forecasting

A blended forecast runs two methods, usually a weighted pipeline and the reps' commit, and treats the gap between them as the agenda. When the two agree, the number is probably sound. When they disagree by more than about a tenth, someone has to look at the deals behind the difference.

How to choose a forecasting method

  1. Start from what the forecast is for. A weekly number for the board needs a method that reacts to this quarter's deals, such as forecast categories or weighted pipeline. An annual plan can lean on historical and regression methods.
  2. Check what data you can trust. If close dates and stages are unreliable, fix the data before adding a model on top of it. The pipeline hygiene checks are the quickest test.
  3. Measure your stage probabilities. Pull the last four quarters and work out the share of deals that closed from each stage. Use those numbers in the weighted pipeline in place of the CRM defaults.
  4. Write down what Commit means. Define each category by what the buyer has done, such as the signer joining a call or the paper process being mapped. The deal risk guide lists the evidence each stage should carry.
  5. Pick a second method to check the first. Weighted pipeline against commit is the common pair. Historical works as a sanity check for the quarter as a whole.
  6. Review the gap every week. Any deal that accounts for a large share of the difference gets discussed in the forecast call, with a named next step.
  7. Track accuracy by method. At quarter end, compare each method's number with the result. After two or three quarters the team knows which one to lean on. The guide to improving forecast accuracy covers how to measure it.

A worked example: four methods, one quarter

This example is illustrative. Arrowmere, a 600-person logistics firm, sells freight contracts to retailers. Its sales team has a target of $3.0M for the quarter and $6.0M of open pipeline due to close in it.

Historical. The same quarter last year closed $2.6M. The business is growing about 10% a year, so the historical method says $2.86M.

Weighted pipeline. The team's own history gives three stage probabilities: Discover 20%, Propose 40%, Negotiate 70%. The pipeline holds $2.0M in Discover, $2.5M in Propose and $1.5M in Negotiate. Weighted, that is $0.4M plus $1.0M plus $1.05M, or $2.45M.

Forecast categories. The reps call $3.1M as Commit.

Sales cycle length. The average cycle is 110 days and 45 days remain. Most of the $2.0M in Discover entered the pipeline in the last month, so very little of it can close this quarter.

The methods range from $2.45M to $3.1M. The gap between commit and weighted pipeline is $0.65M, which is too large to ignore. The sales director asks each manager to go through the Commit deals and check one piece of evidence: has the person who signs joined a call? Deals worth $0.8M fail the check. They move to Best case, which brings Commit to $2.3M. The director submits $2.5M with $0.6M of named upside, each upside deal carrying a planned conversation to bring in the signer.

Where it goes wrong

  • Using the CRM's default probabilities. Out-of-the-box stage weights describe nobody's team. Measure your own, or the weighted number is a guess with decimals.
  • A Commit category with no rule. If each rep decides what Commit means, the forecast is the sum of seven moods. Define it by buyer evidence and apply it in every forecast call.
  • Trusting a model on dirty data. A predictive tool trained on stale close dates learns to predict stale close dates. Fix the data first.
  • Explaining the gap and moving on. When two methods disagree, the gap names deals that need work this week. A forecast call that ends with a number and no next steps has wasted that information.
  • Changing methods every quarter. Accuracy only becomes measurable when the method stays still long enough to compare.

The evidence behind the number

Every method above depends on the stage and category each deal carries, and those depend on what the buyer has done. Tools that flag risky deals are useful, and the work starts after the flag: deciding the next conversation and helping the rep have it. Through Bryce, managers get a Monday read of which deals carry buyer evidence. Reps prepare the conversation that fills a gap with Caty by voice or video on a call. The for leaders page shows the full picture, and a ready-made sales forecast spreadsheet template gives the weighted and category numbers in one sheet. To see it on your own pipeline, talk to us.

FAQ

What are the main sales forecasting methods?

The main methods are historical, weighted pipeline, forecast categories, sales cycle length, regression and AI or predictive forecasting. Many teams blend two of them and investigate the gap between the results.

Which sales forecasting method is most accurate?

The most accurate method is usually a blend: a weighted pipeline built on the team's own stage history, checked against a commit call with a written definition. Accuracy depends more on the deal data than on the method chosen.

What is the difference between qualitative and quantitative forecasting?

Qualitative forecasting relies on judgment, such as reps' forecast categories or a panel of experts. Quantitative forecasting relies on data, such as historical revenue, weighted pipeline or regression models.

How do you calculate a weighted pipeline forecast?

Multiply each open deal's value by the close probability for its stage, then add the results. A $100,000 deal at a stage with a 40% probability contributes $40,000.

How often should a sales forecast be updated?

Update the sales forecast every week during the quarter, after reps have refreshed their deals and before the forecast call. Annual and quarterly plans can use slower methods such as historical or regression.

Put an expert on your deals.

Talk to Keenan about a live deal for free, or talk to us about experts for your team, your product or your coaching business.

Talk to us