MAPE in hotel forecasting: definition, formula and benchmarks

MAPE stands for Mean Absolute Percentage Error.

It measures how far your forecasts deviate from actual results, expressed as a percentage so you can compare accuracy across different periods, metrics and property sizes. The lower the MAPE, the more accurate your forecasting.

For hotels, MAPE is most commonly applied to occupancy, average daily rate (ADR) and revenue per available room (RevPAR) forecasts, though it works across any quantifiable metric you forecast regularly.

Understanding your MAPE (and tracking it consistently over time) gives you a clear diagnostic signal: not just whether your forecasts are off, but by how much and in what direction.

In this short guide, we’ll break down what MAPE is, why it matters for hotels of every size, how to calculate it, what ‘good’ looks like, and most importantly, how you can use it to sharpen your forecasting accuracy.

Key takeaways

• MAPE (Mean Absolute Percentage Error) measures how far your forecasts deviate from actual results, expressed as a percentage — the lower the figure, the more accurate your forecasting

• For most hotels, a MAPE of 10–20% is acceptable; below 10% is excellent, and 15–25% is realistic for properties with high seasonality

• Consistent errors in specific periods — weekends, shoulder season, event weeks — tell you exactly where your forecasting inputs need work

• Tracking MAPE over time matters more than hitting a target number; a steadily improving MAPE is a stronger signal than landing at an arbitrary benchmark

Why MAPE is important for hotel forecasting

Forecasting errors have direct operational consequences. Overestimate demand and you overstaff, over-purchase and end up discounting to fill rooms. Underestimate it and you leave revenue on the table, run short on resources and damage the guest experience.

MAPE turns the abstract concept of "forecast accuracy" into one number you can track, benchmark and improve. It's equally useful for a revenue manager running a sophisticated forecasting model and an independent hotelier working from a spreadsheet; the scale differs, but the diagnostic value is the same.

It's also a better benchmarking tool than most hoteliers realize. Rather than comparing your forecasting performance against other properties – which is difficult given the differences in market type, demand profile and available systems – you compare against your own historical MAPE.

A MAPE of 22% that's down from 30% represents genuine progress. One that's holding steady at 18% despite a more volatile market is a sign your process is working.

How to calculate MAPE

Used by statisticians in many fields, the formula for MAPE might appear slightly trickier than those usually associated with revenue management, but break it down and it’s actually fairly straightforward:

MAPE = (1/n) × Σ (|actual – forecast| ÷ Actual) × 100

Where:

• n = the number of data points (days, weeks or months)

• Actual = the real performance figure

• Forecast = the predicted figure

• Σ = the sum of the results in the brackets

• | | = take the absolute value — treat the difference as positive regardless of direction

In practice: for each period, calculate the percentage difference between your forecast and what actually happened. Add those percentages together, then divide by the number of periods. That gives you the average percentage error across the time window you're measuring.

Let’s take a three-night example:

MAPE = (10% + 20% + 20%) ÷ 3 = 16.7%

On average, your forecasts were off by about 17% across those three nights.

However, MAPE tells you the magnitude of your forecast error but not the direction. A MAPE of 16.7% could mean you're consistently over-forecasting, consistently under-forecasting or bouncing between both. Tracking which way your errors tend to fall – alongside the MAPE figure – gives you the full diagnostic picture.

Common pitfalls to avoid

• Zero actual values: If occupancy or revenue is zero for a period, the MAPE formula divides by zero, which breaks the calculation. To avoid this, your options are to:
  – exclude those periods from the calculation
  – substitute a small placeholder value
  – switch to a different metric, like mean absolute error (MAE), for periods where zero actuals are likely.

• Too few data points: One or two nights won't give you a meaningful picture. Use at least a full month, or a complete season, for a reliable MAPE reading.

• Context without caveats: A high MAPE during a city-wide event week or an unprecedented cancellation wave doesn't necessarily mean your forecasting process is broken. Isolate outlier periods before drawing conclusions about your overall accuracy.

What’s a good MAPE for hotels?

Across industries, a MAPE below 10% is generally considered excellent. For hotels, that benchmark is achievable in some markets and property types, but it isn't the universal standard.

Hotel demand is inherently more variable than many other industries: seasonality, local events, weather and last-minute cancellations all introduce noise that makes tight forecasting harder. Most hotels operate with MAPEs in the 10–25% range, depending on market conditions and forecasting method.

What counts as acceptable depends heavily on your property type and situation.

Hotel MAPE benchmarks by property type

These ranges are indicative. Your market, your systems and how consistently you apply your forecasting methodology all influence where you land.

The goal isn't to hit a target MAPE. It's to improve your MAPE over time. A consistent downward trend – even from 28% to 22% over 12 months – is a more meaningful signal than landing at an arbitrary benchmark.

MAPE vs. other forecast accuracy metrics

MAPE is the most widely used forecast accuracy metric in hotel revenue management, but it isn't the only one. Depending on your property type and what you're measuring, alternatives may serve you better.

When to consider WMAPE instead of MAPE

If your property has significant seasonality – peak weekends, holiday periods, major local events – standard MAPE gives equal weight to a quiet Tuesday in January and your busiest Saturday in August. WMAPE corrects for this by weighting each period's error by how much business actually materialized. For revenue managers who care about accuracy where it counts, WMAPE is often the more honest metric.

How to use MAPE to improve hotel forecasting

Calculating MAPE regularly reveals patterns that a single forecast result never would.

You might find that your midweek forecasts are consistently within 8% of actual but your weekend forecasts routinely miss by 20% or more. Consistent weekend errors often point to something specific: underestimating event-driven demand, overestimating leisure pickup, or both. That diagnosis tells you exactly where to focus — better event tracking, a different baseline for leisure segments, earlier rate adjustments on high-demand dates. Without MAPE as a consistent measure, those patterns stay invisible.

MAPE is also a useful quality check on your forecasting tools. If you use a revenue management system (RMS) or a dedicated forecasting model, your MAPE trend tells you whether the system's outputs are actually improving over time or simply generating numbers with false precision. A rising MAPE from an automated model is a signal to revisit its inputs, recalibrate its parameters or supplement it with manual adjustments.

For properties without dedicated forecasting software – where forecasting still happens in spreadsheets or from PMS reports – MAPE provides a lightweight but rigorous way to make the process measurable. Calculate it monthly. Track it season by season. Use it to run controlled experiments: try a different weighting approach, incorporate a new data source, adjust for local events, then measure whether the number moves.

The discipline of measurement is what turns forecasting from an estimate into a process.

FAQs

What does MAPE stand for?
MAPE stands for Mean Absolute Percentage Error. It's the standard metric for measuring forecast accuracy – calculated by finding the average percentage difference between your predicted values and your actual results across a set of data points.

What is a good MAPE for hotels?
For hotels, a MAPE below 10% is generally excellent. A MAPE of 10–20% is acceptable for most properties, and 15–25% is realistic for hotels with significant seasonality or irregular demand. The most important benchmark is improvement over time: a MAPE trending downward from 28% to 20% over a year is more meaningful than hitting an arbitrary target.

How do you calculate MAPE?
MAPE = (1/n) × Σ (|actual – forecast| ÷ actual) × 100. For each period, divide the absolute difference between actual and forecast by the actual figure, multiply by 100 for a percentage, then average the results across all periods. See the worked example in the calculation section above.

What is the difference between MAPE and MAE?
MAPE expresses forecast error as a percentage, which makes it easy to compare across different metrics and time periods. MAE (Mean Absolute Error) expresses error in the original units – room nights, revenue dollars – which is more useful when you want to understand the practical scale of the error rather than its relative size.

What does a high MAPE mean?
A high MAPE means your forecasts are deviating significantly from actual results. The deviation could be consistent over-forecasting (predicting more demand than materializes) or under-forecasting (underestimating demand), or a combination of both. Tracking the direction of your errors alongside the MAPE value helps identify which pattern applies.

What is WMAPE and when should hotels use it?
WMAPE (Weighted Mean Absolute Percentage Error) weights each period's error by the actual volume for that period. Hotels with strong seasonality often prefer WMAPE because it avoids giving the same weight to a low-occupancy Tuesday in January and a peak-weekend night in August. If some periods matter significantly more to your revenue than others, WMAPE gives a more operationally relevant accuracy read.