What demand forecasting means for a vacation rental manager
Every manager forecasts, whether or not they call it that. The question at the Monday meeting — "are we going to fill July?" — is a forecast, and so is the decision to drop rates on a slow October week. The difference between a manager who forecasts well and one who does not is rarely intelligence. It is whether the answer comes from a consistent, comparable set of numbers or from someone scrolling through the reservation calendar and forming an impression.
A working forecast for a rental portfolio answers three questions on a fixed schedule. Where do we stand for each upcoming week compared with this time last year? Which units, weeks or channels are running behind? And given how bookings normally arrive from here to the stay date, where are we likely to finish? The first two questions are descriptive and can be answered precisely. The third is a prediction, and its honesty depends entirely on how well the first two are built.
Revenue management in vacation rentals borrows its vocabulary from hotels, but the portfolio behaves differently. Each unit is unique, owners block their own dates, weekly stays still dominate some coastal and mountain markets, and a single three-bedroom condo might be booked eleven months ahead while a studio down the hall fills in the final ten days. That is why forecasting has to be done at the level of the unit and the arrival week, not only for the portfolio as a whole.
Booking pace: the core of vacation rental forecasting
Booking pace compares what is on the books (OTB) for a future period today with what was on the books for the equivalent period on the same day last year, usually called same time last year (STLY). If the week of the Fourth of July has 610 booked nights today and had 540 at this point a year ago, you are pacing roughly 13% ahead, even though the week has not happened yet.
The power of pace is that it turns a calendar into a trend early enough to act on. Being 13% ahead with sixty days to go suggests rates for the remaining inventory may be too low. Being 20% behind with three weeks to go is a signal to look at price, visibility and minimum-stay rules before the window closes, not after.
| Measure | This year | Same time last year | Variance |
|---|---|---|---|
| Nights on the books | 610 | 540 | +13% |
| Revenue on the books | $198,300 | $171,900 | +15% |
| Average daily rate | $325 | $318 | +2% |
| Pickup in the last 7 days (nights) | 44 | 61 | −28% |
Notice that the example is ahead on the books but slowing on recent pickup. That combination usually means the early bookers arrived on schedule and the late-booking segment is softer than last year — a very different situation from simply being "ahead." Reading pace and pickup together is what turns a number into a decision.
Comparisons must be aligned. Holidays move between days of the week, Easter moves between months, and a school calendar change can shift a peak week entirely. A good pace report compares like-for-like arrival weeks and day-of-week patterns rather than raw calendar dates.
Why unit, week and channel matter more than the portfolio average
A portfolio that is 4% ahead of last year can contain forty units that are 25% ahead and thirty that are 30% behind. The average looks healthy while a third of your owners are about to have a bad summer. Pace by unit exposes that, and it is also the view that matters most for owner relationships, because owners judge you on their own property, not on your portfolio.
Arrival week is the second essential cut. Demand in a vacation market is lumpy: holiday weeks, school breaks, event weekends and shoulder seasons behave like separate markets with different booking windows. Rolling everything into a monthly figure hides exactly the weeks where intervention would pay.
Channel is the third. If pace is flat overall but Airbnb is up and your own website is down, the forecast is fine and the marketing has a problem. If direct bookings are up and the OTAs are down, you may be seeing a change in the platforms' search ranking rather than in traveler demand. For more on reading channel economics, see our guide to direct booking intelligence.
The data a vacation rental forecast needs
Most managers already own everything required. The work is extracting it in a consistent shape. The essential inputs are:
- Reservations with the booking-created date, not only arrival and departure. Without the date each booking was made you cannot reconstruct what was on the books a year ago, and pace becomes impossible.
- Cancellations and modifications, with their dates, so that last year's on-the-books figure reflects what was actually held on that day.
- Rates and revenue by night, separated from cleaning fees, taxes and pass-through charges.
- Available inventory: units added or lost, owner blocks and maintenance holds. A unit that left the program last spring should not make this year look worse.
- Channel and source for every reservation.
- Demand signals from your own website: searches by arrival date, availability checks and abandoned checkouts often move before bookings do.
- External context such as event calendars and, in coastal markets, the tropical weather season tracked by the National Hurricane Center.
The property management system holds most of this, whether that is Streamline, Track, Escapia, LiveRez, Hostaway, Guesty or another platform. Pulling it out cleanly and on a schedule is covered in our article on vacation rental data integration.
How forecasting works alongside dynamic pricing tools
Dynamic pricing tools — PriceLabs, Wheelhouse, Beyond, or Airbnb's own Smart Pricing — recommend or set nightly rates from market data, seasonality and your occupancy. They are useful, and most professional managers should use one. But they answer a narrower question than a forecast does.
A pricing tool sees the market and the calendar. It does not necessarily see your direct-booking funnel, the owner who is threatening to leave, the twelve units joining next quarter, or the fact that last year's July was inflated by a one-off event. A forecast built from your own data is the check on the tool: it shows whether the rates it set are producing the pace you need, and it gives revenue managers grounds to override it when they should. We go deeper on the tools themselves in vacation rental dynamic pricing.
The pricing tool decides tonight's rate. The forecast decides whether the strategy behind it is working.
Common forecasting mistakes
Most broken forecasts fail in predictable ways. The most common is comparing against last year's final result instead of last year's position on the same day, which makes every future week look behind. The second is ignoring inventory change, so a portfolio that grew by 15% looks like it is booming when same-store demand is flat. Third, cancellations are left in last year's numbers but removed from this year's, which flatters the comparison.
Others are subtler: treating occupancy as the goal when revenue per available night is what owners are paid on; reading monthly totals that blend peak and shoulder weeks; and jumping to machine-learning models before the descriptive numbers are trusted. A sophisticated model trained on inconsistent data produces confident nonsense. Get pace right first, then predict.
The forecasting metrics that matter
Define each of these once, in writing, and use the same definition everywhere — dashboards, owner reports and meetings.
- On the books (OTB)
- Nights and revenue already reserved for a future period as of today, net of cancellations.
- Same time last year (STLY)
- What was on the books for the equivalent period on the same day a year earlier, aligned by day of week and holiday.
- Pickup
- New nights or revenue booked for a future period during a recent window, usually the last 7 or 14 days.
- Occupancy
- Booked nights divided by available nights, after removing owner and maintenance blocks.
- Average daily rate (ADR)
- Rental revenue divided by booked nights, excluding cleaning fees, taxes and pass-through charges.
- RevPAR
- Revenue per available rental night — occupancy multiplied by ADR. The best single measure of how well a unit earns.
- Booking window
- Days between the booking date and arrival. Its shape tells you how much of a week is still to come.
- Same-store portfolio
- Only the units present in both periods, so growth or attrition does not distort the comparison.
How to build a booking pace forecast
A first version can be running within weeks. Each step makes the next one trustworthy.
- Export two to three years of reservations. Include booking-created dates, cancellation dates, channel, unit, nightly revenue and stay dates. If your system cannot export booking dates, fix that first.
- Define available nights and the same-store set. Remove owner blocks and maintenance holds, and flag units that joined or left so comparisons stay fair.
- Build pace curves for each arrival week. For every future week, plot nights and revenue on the books against the same point last year, aligned by day of week and holiday.
- Cut by unit and channel. Rank units by variance to last year and show channel mix, so the weekly meeting starts with the outliers.
- Agree decision rules. Write down what happens when a week is behind or ahead by a set margin at a set distance from arrival: rate review, minimum-stay change, marketing push.
- Add a projection. Once history is clean, project the final position from the normal booking curve, then measure forecast error every week and improve it.