How to forecast hotel demand without a data team

You need to know, roughly, how many rooms you'll sell three weekends from now before you set next month's schedule, order linens, or decide whether Saturday's rate should hold or move. Most hotel forecasting content assumes you have a revenue manager reading a dashboard built for exactly that job. If you're the owner, the GM, and the person who answers the front desk phone, the real question isn't which forecasting software to buy. It's how to forecast hotel demand using the booking data already sitting in your property management system.

What a hotel demand forecast actually needs

A forecast is a specific claim about the future, not a feeling: how many rooms will sell on a given date, and how confident are you? Get it wrong high and you've overstaffed an empty night; get it wrong low and you're short-staffed on your best weekend of the month. Every forecasting method, from a notepad to an enterprise RMS, is built from the same three raw ingredients.

On-the-books (OTB) is what's already reserved for a future date, straight from your property management system. Pickup is how that OTB number changes between two points in time — the net rooms added or lost since your last check. Pace compares how fast pickup is accumulating this year against the same point in the booking window last year, which is what tells you whether a date is running ahead of or behind its own history. Comp set — the handful of competitors whose rates and occupancy you benchmark against — adds outside context once your own numbers are in hand.

None of that requires buying anything. A PMS export, a spreadsheet, and a repeatable schedule for pulling both are the entire toolkit — the same terms the industry's own rate-shopping and forecasting guides use, like Lighthouse's occupancy forecasting guide, worth a read if you want the longer version before building your own.

The pickup-and-pace method, explained simply

Start with the version every independent hotel can run today: same-date-last-year, adjusted for current pace.

Say a Saturday six weeks out sold 62 rooms by this point in its booking window last year, and finished the night at 71 total. If this year's OTB for that same Saturday is running 12% ahead of where last year's OTB stood at the same lead time, your baseline forecast is 71 × 1.12 ≈ 79 rooms — before adjusting for anything local that last year didn't have.

That's the classic "additive pickup" method, and it has stayed in use for decades because it's cheap to run and rarely produces a wild answer. Its weak point shows up at longer lead times, where one noisy week can throw off the whole number. A study in the Journal of Revenue and Pricing Management tested an extended version on a small independent hotel's reservation data — folding in a rolling average of recent pickup history instead of anchoring to a single comparison point. Forecasting up to 28 days out, the extension beat the classic method, and the gap was widest at the longer horizons and on dates the hotel wasn't full. The improvement wasn't new software; it was using more of the history already sitting in the PMS instead of trusting one data point. Worth knowing the test conditions, though: a single property, over a stretch when travel demand was badly disrupted. Encouraging for the method, not a promise about your hotel.

How to forecast hotel demand in 30 minutes a week

Here's a version that runs on what you already have.

Pull on-the-books for the next 90 days. Most PMSs export this in a couple of clicks — rooms and revenue booked, by arrival date. About 15 minutes.

Pull the same window from last year, at the same lead time — what was booked 90 days out from that date last year, not just how the date eventually finished. A few minutes if you saved last year's export; longer the first time you build the habit.

Calculate pace for each date: this year's OTB divided by last year's OTB at the same lead time. Anything outside roughly 0.85–1.15 deserves a second look — either real demand movement or a data problem, like a group block that fell off the books or a channel that stopped mapping correctly.

Apply pace to last year's final result, not to this year's partial OTB, to get a working forecast for the date.

Log it, one row per date, refreshed weekly:

Date	Days Out	LY Final	This-Year OTB	Pace	Local Event?	Forecast
Aug 15	26	71	58	1.12	Convention move-in	80
Aug 22	33	64	49	0.98	—	63

Total time: 25–35 minutes for a single property with a manageable comp set. Longer the first week; shorter every week after, once the pull becomes routine.

If you outgrow the single-date view — several rate classes, multiple lengths of stay — Cornell's Center for Hospitality Research publishes a free forecasting aggregation tool built for exactly that next step, no subscription required.

Add your local events calendar before you trust the number

Pace tells you how this year compares to last year. It doesn't tell you why — and the why is usually sitting in a calendar somewhere, not in your booking engine.

Before acting on a forecast, check it against what's actually happening in your market on those dates: a convention move-in, a festival, a graduation weekend, road construction cutting off your best parking, a new hotel three blocks away opening with launch pricing. Permit filings and zoning notices are usually public months before a shovel goes in the ground, which is the only warning you'll get before a competitor's added room count changes the whole market's occupancy math. Building a demand calendar from these signals is worth doing once and updating monthly, rather than reconstructing it from memory every time a number looks wrong.

The adjustment doesn't need to be precise. If pace says a date is running flat but a 4,000-person conference just added your city to its rotation for the first time, you don't need a model — you need to move that forecast up and put the date on your rate-review list before your comp set gets there first.

How far ahead should you forecast, and how often should you check it?

A forecast that only lives in your head decays the moment the property gets busy — which is exactly when the market is moving fastest. Two habits keep it honest.

Match resolution to horizon. Daily detail for the next 30 days, where staffing and rate calls are immediate. Weekly buckets for days 31–60, where you're watching pace trends more than single dates. A broader view for days 61–90, mostly to catch anything unusual — a large group hold, a citywide event — early enough to still react.

Reconcile every week, not just when a number feels off. Pull last week's forecast, compare it to what actually happened, and note the miss. A forecast that consistently runs high or low in the same direction isn't bad luck; it's a sign your pace baseline or event weighting needs adjusting.

This is also where an STR report earns its keep. Your own OTB and pace explain your property. Occupancy and RevPAR index against your comp set tell you whether a miss was you, or the whole market moving under everyone at once.

When a manual forecast reaches its ceiling

The system above has an honest ceiling. It scales fine to one property, a handful of rate classes, and a comp set small enough to watch by eye. It strains once you're forecasting a second location, more segments than fit in a spreadsheet, or a market where new supply and events move faster than a monthly calendar review catches them.

Even before that ceiling, the harder half of this system usually isn't the math — it's staying current on the local signals that explain a pace shift before your rate decision is already late: a competitor announcing a renovation, a conference confirming your city a year out, a new property pulling permits three blocks away. The Intel Club is built for that half: a daily briefing that tracks competitor moves, market shifts, and local signals in your market and opens with a recommended next step, so the events-and-competitors side of your forecast shows up on its own instead of requiring a separate search every week. Membership is $99/month with a 7-day trial; if you'd rather see the shape of it first, see what a daily briefing covers for hotels.

Either way, the forecast itself — on-the-books, pace, last year's numbers — stays yours to run. Thirty minutes a week, one log, reconciled honestly, beats a number nobody checks.