Sign in

Blog · Forecast and pipeline

Four sales forecasting methods compared, and the bridge that shows where they disagree

The four common ways to forecast a quarter, rep roll-up, stage-weighted pipeline, run rate and historical conversion, what each needs, what each is good and bad at, a worked quarter where they give four different numbers, and why the right output is not one of them but the bridge between them, with the deals that explain the gaps.

The short answerRep roll-up is what reps commit; stage-weighted pipeline is open deals times CRM probability; run rate is recent revenue scaled; historical conversion is pipeline at this week times what such pipeline closed at before. They give four numbers, and each is right about something. The useful output is the bridge between them: the roll-up exceeds historical conversion by the deals reps are committing above their track record, and the deal list names them. A forecast call that starts from the bridge argues about deals, not methods.

Four forecasts, four numbers, one quarter. The sales leader picks the one that matches the target and the CFO picks the one that does not. The useful output is neither; it is the bridge between them, made of deals. This guide sets out the four methods, what each needs, a worked quarter, and the bridge.

The four methods

Method Formula Needs Good at Bad at
Rep roll-up Σ deals reps commit Rep judgement Known changes; deal-level knowledge Bias; optimism at week one
Stage-weighted pipeline Σ open deal value × CRM probability CRM stages Simplicity Probabilities nobody calibrated
Run rate Recent revenue scaled to the period Ledger The floor; stability Anything that changes
Historical conversion Σ pipeline at this week × conversion of such pipeline in past quarters Weekly snapshots and outcomes Calibration to the team's own record Changes the record has not seen

The rows you need

  • Pipeline snapshots: opportunity, rep, stage, value, close date, commit flag, week.
  • Outcomes: opportunity, won or lost, date, value.
  • Ledger: customer, period, revenue.

Opportunity and rep identifiers only.

A worked quarter, at week four

Method Forecast
Rep roll-up $4.6m
Stage-weighted pipeline $4.1m
Historical conversion $3.7m
Run rate $3.4m

Four numbers, a spread of one point two million. Each is defensible. None is the answer.

The bridge

From To Line Value Deals
Run rate Historical conversion Pipeline above run-rate customers' typical repeat +$0.3m 22 deals at new or growing accounts
Historical conversion Stage-weighted CRM probabilities above the team's track record, by stage +$0.4m Proposal stage at 60% in CRM, 42% in history
Stage-weighted Rep roll-up Deals committed above their stage weight +$0.5m 9 deals, listed, 6 of them with rep R-04

The roll-up is nine hundred thousand above the team's own track record, and the bridge names the nine deals and the stage whose CRM probability is inflated. The forecast call is about those nine deals and whether the proposal stage really converts at sixty percent this quarter.

The identity

each method's total = Σ its deal-level or customer-level components bridge lines sum from one method to the next, with every deal on exactly one line

A deal on two bridge lines, or on none, fails it and is listed.

Where it goes wrong

Picking one method. The argument is about the method instead of the deals.

Probabilities uncalibrated. Stage-weighted is a guess with decimals.

No snapshots. Historical conversion cannot be computed; the calibrated method is unavailable.

Bridge without deals. A haircut line explains nothing.

Every week, four numbers and the bridge

Mapped once, the snapshots, the outcomes and the ledger produce all four forecasts, the bridge and the deal lists every week. Covirage builds this from the exports as they are. The forecast analysis solution describes the setup, and the forecast bias guide covers the rep-level calibration that explains most of the gap between the roll-up and history.

Questions people ask

Which method is most accurate?

At most teams, historical conversion, because it uses the team's own track record rather than reps' judgement or CRM probabilities. But it is blind to a known change, a new product or a lost customer, that the rep roll-up carries. The bridge is where the two meet.

Is run rate a forecast at all?

It is the floor: what happens if nothing changes. A forecast below run rate needs a reason; a forecast far above it needs pipeline to justify it, and the bridge shows whether the pipeline is there.

What does the bridge contain?

Lines from one method's number to the next, each attributed to named deals or a stated assumption: deals committed above historical conversion, deals in the pipeline with no rep commitment, run-rate customers with no open deal. Every line sums, and every deal is on one line.