Every revenue decision — the rate you load, the group you accept, the staffing the GM signs off — leans on one number: how many rooms you expect to sell. Yet "the forecast" is often a spreadsheet someone inherited, updated by feel. There are only three core methods working revenue managers use, and they're all buildable in a spreadsheet. Here is each one, with the math, plus the adjustments and the accuracy check that separate a forecast from a guess.
A rooms forecast predicts, for each future stay date: room nights sold (occupancy), ADR, and therefore revenue and RevPAR. Two distinctions matter before any method:
Pickup forecasting answers: from this many days out, how many more rooms do we typically add? For each stay date you look at rooms OTB today, then add the average pickup your hotel has historically generated from this many days before arrival (DBA):
Forecast = OTB today + historical pickup from this DBA to arrival
Example: a Tuesday stay date is 14 days out with 52 rooms OTB. Over the last eight comparable Tuesdays, the hotel added an average of 21 rooms inside the final 14 days. The pickup forecast is 52 + 21 = 73 rooms. Track pickup at standard checkpoints — 90/60/30/14/7/3/1 DBA — and separate weekday from weekend (their booking curves are different animals), and you have a pickup model. The discipline is in the reference set: comparable days only, recent enough to reflect current demand, with event-distorted dates excluded or handled separately.
Pace forecasting compares where you are now to where you were at the same point before the same date last year (STLY — same time last year), and assumes the relationship holds to the finish:
Forecast = LY final × (OTB today ÷ STLY OTB)
If you have 60 rooms OTB where last year you had 50 at this point, you're pacing at 120%; if that date finished at 80 rooms last year, the pace forecast is 80 × 1.20 = 96 rooms. Pace is fast, intuitive, and excellent at flagging dates that are ahead or behind. Its weakness is baked in: it assumes this year books like last year. A calendar shift (Easter moving, a citywide event that changed dates), a new supply opening, or a channel change all break the ratio — which is why pace is best used as a cross-check and an alarm, not as the only number.
Different business books differently, so forecast it differently and sum the parts:
| Segment | How to forecast it |
|---|---|
| Transient (retail, OTA, direct) | Pickup method — it books on a curve |
| Corporate negotiated | Recent weekly production per account, adjusted for known travel patterns |
| Group | Definite blocks × expected materialization (wash), plus tentatives × conversion probability |
| Contract / crew / wholesale | Contracted allotments — near-fixed, forecast at commitment |
A 120-room hotel might build a night as: transient pickup 58 + corporate 22 + group block 40 at 85% expected materialization (34) + crew contract 6 = 120 forecast against 126 nominal — which also tells you the night is a sellout risk and the last transient rooms should not go cheap. Segment build-up takes the most maintenance and repays it with the most insight: when the forecast misses, you can see which business missed, and ADR follows naturally because each segment carries its own rate.
A forecast nobody scores never improves. After each month, compare forecast to actual at a fixed checkpoint (e.g. the 30-days-out snapshot) using absolute percentage error, averaged into MAPE:
MAPE = average of |actual − forecast| ÷ actual
A common working standard is within ±5% at 30 days out on rooms; tighter as arrival approaches. Score it by day of week and by segment — a 3% total miss can hide a group segment that missed by 20% and a transient segment that covered for it. The bias matters too: consistently forecasting low (sandbagging) misprices your peak nights just as surely as optimism overstaffs your soft ones.
All three, in proportion to your data. A small independent can start with pace vs LY this week (it needs only OTB and last year's finals), add pickup checkpoints within a month, and grow into a segment build-up as tracking matures. A hotel with meaningful group business needs the segment structure — wash is unmodelable any other way. Whatever the blend, the cadence matters more than sophistication: a simple forecast updated every week beats an elaborate one updated when someone remembers. A working rhythm: daily glance at the next 30 days' pickup, a weekly 90-day forecast update, and a monthly rolling 12-month refresh for budget and staffing.
The Forecast & Pickup Workbook in the toolkit is exactly these methods as working spreadsheets: a 90-day pickup tracker with DBA checkpoints, a segment build-up forecast that rolls RN × ADR into revenue, occupancy and RevPAR, and a pace-vs-LY sheet — formulas live, ready for your numbers. The forecast review step-by-step (what to check, in what order, before the morning meeting) is part of the daily RM routine, and the full SOP set around it is in the RM Playbook.