Who this is for. An independent hotel where the forecast is built in a spreadsheet, without a revenue management system. The method is the same one large chains use; what changes is the number of segments you can realistically handle.
What you will be able to do. Produce a daily forecast of roomnights and ADR broken down by segment, channel and room type; flag the dates where last year stops being useful; project groups on what actually gets occupied rather than what gets blocked; net off cancellations and no-shows, and set your overbooking level; read your weekly pickup and pace; answer how many rooms you will still sell and at what rate; carry it through to TRevPAR if you sell anything beyond the room; and turn all of that into a decision, which is the only part that makes money.
Three acronyms to fix before starting, because they appear in every report in this industry: OTB (on the books, what you hold right now), LY (last year), STLY (same time last year, read at the same lead time).
What is the difference between a hotel budget and a forecast?
A budget is built once a year, approved, and does not change. A forecast is dynamic: it is revised as new data arrives and it is what you use to correct the plan while the year is still running. The budget is, in effect, your first forecast. Every forecast rests on two components: a long-term trend, which is what seasonality and last year tell you, and a short-term trend, which is what the booking curve of recent weeks is telling you now. The core formula is Forecast = OTB + (Closing LY - OTB STLY), applied separately to roomnights and ADR, and always by segment. A worked example on ADR: OTB today is 84, last year at the same point you were at 90, and you closed at 101. Expected pickup is 11, so the ADR forecast is 95.
Part 1. Foundations
Module 1. What a forecast is, and what it is not
The budget is approved in September or October and stays fixed. It is the reference you are measured against. The forecast moves every week with real data, and its job is to tell you, early enough to act, that the budget will not be met, or that it will be beaten.
Two components drive it. The long-term trend is seasonality and last year. The short-term trend is the booking curve of the last few weeks. A forecast built on only one of the two fails in a predictable direction: rely only on history and you miss this year's shift; rely only on current pace and every quiet fortnight looks like a catastrophe.
Exercise. Write in one line what your budget says for next month, and in another what your current OTB says. The gap between the two is what this guide teaches you to explain.
You have understood this module when you can say why a forecast that never moves is not a forecast.
Module 2. The data you need, and cleaning it first
Four figures per day: current OTB, OTB STLY, last year's closing figure, and budget.
Before projecting anything, clean the base. Remove cancellations and forecast on what is genuinely left to sell. Check that every block and every reservation is attached to the correct rate code. This is unglamorous work and it is where most bad forecasts are born: a base with rate errors produces a confident projection of the wrong number.
Two checks worth doing every time. First, that group blocks are real: a block held for a group that never confirms is inventory you are not selling and volume you are not forecasting. Second, that no-shows and early departures are reflected in last year's closing figure, because if they are not, your history is optimistic and every projection built on it inherits that optimism.
Exercise. Take next month, pull the four figures per day, and flag in a separate column any day where you had to correct a rate or release a block. Those flagged days are the ones to watch.
You have understood this module when your OTB matches the PMS after cleaning, and you can explain any day where it did not.
Part 2. Segment before you project
This is the part most independent hotels skip, and it is why their forecast ends up as a single number nobody can act on. A total occupancy figure does not tell you what to do. A figure broken down by segment, channel and room type tells you exactly which lever to pull.
Module 3. Segmentation: the taxonomy the industry uses
The standard classification splits demand into three families.
Transient, meaning individual business: Retail (rack, BAR, OTA gross), Discount (non-refundable, semi-flexible, long stay, OTA net, loyalty), Negotiated (corporate, consortia), Qualified Discount (government and similar programmes), and Wholesaler or FIT.
Groups: corporate, association and convention (MICE), government, tour operator groups, SMERF (social, military, educational, religious, fraternal), and contracts, which covers crews and blocks longer than thirty days.
Others: no-shows, day use, early departure fees, late check-out, extra beds, upgrades.
On top of that sits the distinction that decides what you can actually manage, called yieldability:
- Yieldable: you control both quantity and price. Transient and MICE. This is where yield management applies.
- Semi-yieldable: supplements already negotiated. Your margin lies in managing room types and services during high occupancy. Corporate accounts, tour operators.
- Non-yieldable: in theory not managed, but it delivers annual volume. Crews, tour operator series, government programmes. If it suits you, you release more rooms at the negotiated rate.
Do not try to run fifteen segments. Keep the five or six that genuinely behave differently in your property and group the rest. A segment you cannot explain is a segment you cannot forecast.
Exercise. List last year's segments with their roomnights and revenue. Mark each as yieldable, semi-yieldable or non-yieldable. If two behave identically in volume and rate, merge them.
You have understood this module when you can say which of your segments you can genuinely price, and which only bring volume.
Module 4. Channels: mix, net ADR and cost of acquisition
Forecasting by channel is not an accounting exercise. It is how you decide where to push.
The channel families are OTAs (net, opaque, flash sales), B2B (net, gross, FIT, opaque), GDS (consortia, negotiated, FIT, static or dynamic), your own web engine (direct, SEM, metasearch), groups, and direct contact (walk-in, email, telephone).
Three rules that change decisions:
Work with net ADR, never gross. Each channel carries its own commission and acquisition cost. Two channels showing the same ADR can leave very different amounts of money in the property. The typical net ADR hierarchy runs direct above IDS, IDS above groups, and groups above corporate and FIT, but treat that as a starting point to verify in your own numbers, not a law.
Direct is not automatically the best channel. It carries costs sitting in four different budget lines: booking engine fees, metasearch and paid search, loyalty discounts, chargebacks, and the hours your team spends on the phone. The goal is the balance point between channels while you optimise the direct one, not a war against OTAs.
Have the acquisition cost per channel written down and agreed. Not estimated in a meeting. The classic failure is the property that says twenty per cent all year and discovers at year end that the real figure was twenty-eight.
Now the part that decides actions. If a channel performed well last year in a given period, that is a hypothesis for this year, not a certainty. Forecast it by channel, compare against your current pace by channel, and if one channel is behind while the others hold, that is exactly where a targeted action belongs: a promotion visible only in that channel, a rate plan opened there, an allotment released. Acting on the total when the problem sits in one channel spends budget and moves nothing.
Exercise. Build a table with one row per channel: roomnights LY, revenue LY, commission, net ADR, and current OTB for the same period. The channel with the widest gap between LY and current OTB is your first candidate for action.
You have understood this module when you can name the channel that leaves you the most money per room, which is rarely the one with the highest gross ADR.
Module 5. Room types: differentials and how to forecast them
Forecasting only total rooms hides the most common problem in independent hotels: you sell out your cheapest room type on every peak and leave the expensive ones empty, then conclude the hotel was full when what actually collapsed was the mix.
The differential between room types is set with one of four methods: absolute value, absolute value by season, percentage, or percentage by season. Choose one and keep it consistent, because a differential that moves without a reason breaks both guest trust and your own reporting.
For the forecast, work at two levels at least: total, and by room type on your peak dates. If one room type runs ahead of last year while another lags, the lever is not price, it is availability. Closing a room type on a peak date, or opening it, is one of the fastest levers you have.
Exercise. For your ten highest-occupancy dates last year, write down occupancy by room type. Identify which type sells out first. That type is your bottleneck and its differential deserves a review.
You have understood this module when you can explain, for a date that sold out, which room type filled first and what that cost you.
Module 6. Lead time: when demand actually arrives
Lead time is the distance between booking date and stay date, and each segment has its own curve. The industry training material cites these examples: leisure at around five days, corporate at two, a large group about a year out, a sporting event further still, an airline contract around ninety days. Treat them as what they are, examples from specific properties. Your curve looks like nobody else's and it is the only one that decides.
The practical consequence is that inventory control and pricing must be segmented by pace curve rather than applied uniformly. The industry usually groups lead time into buckets: in period, 0-3 days, 4-7, 8-14, 15-29, 30-59, 60-89, and 90 or more.
The useful question is simple: at what point in the curve does the bulk of your business arrive? If most of your volume comes in inside fifteen days, holding a firm rate ninety days out costs you nothing and protects you. If your volume arrives at sixty days, a last-minute discount lands after your guest has already decided.
Exercise. Take one month of last year's reservations and classify them into the buckets above. Draw the curve. That curve is your calendar for taking decisions.
You have understood this module when you can say how many days before arrival the rate decision for a given date should already be taken.
Part 3. Project
Module 7. The demand calendar: when last year lies to you
Everything that follows rests on one idea: last year tells you what will happen this year. That is true most of the time, and when it is not, it does not warn you. Which is why the first step of projecting is not a formula, it is a calendar.
Three cases break the comparison with last year:
- A new event. A congress, a concert, a trade fair that exists this year and did not exist last year. Your history does not contain it, so the formula will project an ordinary date where you are about to get compression.
- An event that has gone. The most expensive of the three, because nobody looks for it. Last year you had the congress and this year it has moved to another city. Your history is inflated and you will project demand that no longer exists.
- A holiday that moves. Easter changes month. Long weekends change week. Two months with the same name may not hold the same Fridays and Saturdays, and in a hotel that is half the business.
From this comes a rule repeated throughout this guide: comparisons are made day of the week against day of the week, never date against date.
Classify every date in the period as one of six types: low, shoulder, normal, high, compression and special event. Compression earns its own label because it is not simply a busy date: it is when the whole market fills up and your rate can move far more than you would dare on an ordinary day.
And one warning that saves money: do not multiply factors blindly when they overlap. If an event repeats every year, it already lives inside your history. Adjusting for it again counts it twice and leaves you with a projection nobody can explain when it misses.
Exercise. Take the next ninety days and mark every date with its type, its event and its holiday. Then flag the dates where last year is no use, and write beside each one what percentage you will adjust by, and why.
You have understood this module when you can point at a date in your calendar where the formula is going to lie to you, and say in which direction.
Module 8. The core formula, applied by segment
Forecast = OTB + expected pickup
Expected pickup = Closing LY - OTB STLY
Forecast = OTB + (Closing LY - OTB STLY)
Apply it separately to roomnights and to ADR, and always by segment. A worked example on ADR: OTB today is 84, last year at the same point you held 90, and you closed at 101. Expected pickup is 11, so your ADR forecast is 95.
ADR needs one extra step that roomnights does not: adjust it with what you specifically know. A contract renewed at last year's rate, a channel policy that changed in January, a segment you decided to stop taking. The formula gives the starting point, not the final answer.
Two cautions when segmenting. First, a segment with very few roomnights produces a volatile forecast, so group the small ones. Second, if you changed your segmentation mid-year, last year's figures are not comparable, and pretending otherwise is how a forecast turns into fiction without anybody lying.
Exercise. Run the formula for one week, by segment, for roomnights and ADR. Keep the segments you actually sell, not the ones the PMS offers by default.
You have understood this module when you can produce both numbers for any day without looking up the method.