Forecasting and pacing

Define what the forecast measures

Pacing uses closed-won bookings, quota, the calendar, and the expected booking pattern. A deal forecast adds estimated open-deal contribution to booked revenue. Show the sales leader's submitted forecast alongside both so the review can identify where their estimates differ.

• Agree on period, review cutoff, motion, currency, amount definition, owner keys, and eligible population.
• Record when the source was exported, imported, and reviewed. Importing today's deals does not update the historical win rates or booking curve.
• Use a consistent selling-day convention, including any holidays. Avoid applying ramp twice when quota already includes it.
• Show unavailable comparisons when quota, calendar, or source coverage is missing.

Compare progress, required pace, and forecast

Let A be booked revenue, Q the target, and c the share of full-period bookings expected by the review date. Use A / c to project a finish from bookings so far. Use Q − A to calculate what still needs to close. The amount required to hit target is not evidence that those deals will close.

• Attainment = A / Q. Seasonal expected bookings = Q × c. Linear expected bookings = Q × elapsed selling-day share.
• Curve-based finish = A / c, when the curve has reached the chosen minimum for displaying a projection. Early in the period, a small booking can swing the projection sharply because c is small.
• Remaining gap = Q − A. Required daily pace = gap / remaining selling days.
• Expected remaining bookings = Q × (1 − c). Required lift = gap / expected remaining bookings; 1.0 means the normal remaining pattern would suffice.
• Show target met for a nonpositive gap; show the outstanding gap at period close. A zero denominator makes the corresponding ratio unavailable.

Save the forecast and its assumptions

Document whether the curve is based on deal counts or booking dollars, its cohort and completed periods, calendar coordinate, blend, fallback, and version. A count curve limits deal-value influence but does not exclude large deals unless a filter does. Test whether it suits dollar pacing.

Compare the estimates based on stage, forecast category, rep history, and close timing with the sales leader's submitted forecast. Find which deals account for the difference. Only combine estimates after testing their accuracy on comparable past periods.

• Keep the production curve fixed for an agreed evaluation period; log shadow candidates separately.
• Save source and calibration cutoffs, actuals, call, curve version, exceptional-deal treatment, and decision.
• Preserve prior deal states to explain amount, timing, and category changes. A missing snapshot does not prove a newly created opportunity.
• When a buying process has ended, the seller closes it directly in the CRM and records the reason.

Challenge commit and coach the forecast

I challenge commit calls by asking whether the rep would bet their commission on those opportunities closing. The discussion then needs to establish what the buyer has committed to, what could prevent the purchase, and whether the timing supports the call.

I also use a forecast-accuracy challenge. Reps try to beat my forecast for their own number or the team. We review the closest calls in the weekly forecast meeting, then compare them with the quarter-end result for the overall prize. The aim is to build forecasting judgment and make accuracy worth paying attention to.

• Save the rep’s call, the manager’s adjustment, the date, and the forecast period before updating the CRM or assumptions.
• Agree on how closeness will be measured before the challenge. Compare calls made at comparable points in the period.
• Use a missed forecast to investigate the underlying deals and assumptions. Identify the buyer question, coaching need, or forecast change that follows.
• Keep quota attainment on its own scorecard. A strong sales result does not make every earlier forecast accurate.

Review the manager’s contribution to forecast error

Forecasting behavior is shaped by the manager as well as the rep. I spend time with sales managers on why accuracy matters and how overconfidence or inflated calls affect the forecast.

Retain the original rep call when the manager changes it. At period close, compare both with the same actual result and discuss the evidence behind the adjustment.

Worked calculation: bookings are behind the seasonal plan

Assume a $1.2 million quarterly new ARR bookings target, $300,000 booked, and 30 selling days left. Half the selling days have elapsed, but the chosen booking pattern expects only 35% of the quarter’s bookings by this date.

• Expected bookings by this date: $1.2 million × 35% = $420,000. Bookings are $120,000 behind that pattern. A straight-line comparison would expect $600,000.
• Projected finish from the booking pattern: $300,000 / 35% ≈ $857,000. This calculation does not use open opportunities.
• Bookings still needed: $1.2 million − $300,000 = $900,000, or $30,000 per remaining selling day.
• The normal remaining pattern accounts for $780,000. Reaching target requires about 1.15 times that amount.
• If the submitted forecast is $960,000, reaching target needs another $240,000 supported by deals that can close before quarter-end.

Measure forecast accuracy over time

Compare each saved forecast with the final bookings result using the same lead time, such as four weeks before quarter-end. Test whether it beats a simple baseline. Keep the model version and large-deal treatment fixed so the comparison shows forecast performance rather than a change in method.

• Track signed and absolute error in the same currency and measure; percentage errors can fail when actuals are near zero.
• Check interval coverage against later outcomes before attaching a confidence claim.
• Do not treat repeated snapshots of one quarter as independent quarter validations.
• Explain whether a changed result came from the cohort, count-versus-dollar basis, blend, freshness, or another assumption.