Metrics Libraryby Semaphor
Library / Hospitality
Hospitality · Owner economics

Projection accuracy (projected vs actual revenue)

Projection accuracy compares the revenue projected for a home when the owner signed with what it actually earned. It ties sales promises to owner trust.

Plan your metrics

What it is

Managers often give prospective owners a revenue projection, sometimes built from market data and comparable homes. Once the home is live, actual revenue can be compared to it.

This is not a widely standardized metric. It is included because owners judge managers against what they were told.

Why operators track it

  • Owner trust. A home that earns well below its projection is an at-risk owner.
  • Sales discipline. Shows whether projections are consistently optimistic.
  • Model improvement. Errors by market and home type show where the model is weak.
  • Retention link. Read it with owner churn.

How it is calculated

Projection accuracy = actual revenue ÷ projected revenue. Error per home = (actual − projected) ÷ projected.

  • Use the same revenue basis as the projection (gross rent, GBV, or owner net).
  • Compare the same period: usually the first 12 months live.
  • Report the portfolio ratio and the mean absolute error per home.

Worked example

Example data The same made-up portfolio is used on every hospitality page: a manager with 120 homes, four of them shown in detail.

Five example homes onboarded in 2025 were projected to earn $485k in their first year and earned $442k: 91.1% of projection. Per-home errors ranged from −18.7% to +3.3%. Mean absolute error was 9.7%.

First-year actual vs projection at signupBeach house−13.3%Condo+3.3%Cabin−18.7%Studio−3.3%Lake house−10.0%$104k vs $120k$93k vs $90k$61k vs $75k$58k vs $60k$126k vs $140kPortfolio: $442k actual vs $485k projected = 91.1% of projection. Mean absolute error 9.7%.Four of five homes came in under projection. Averaging signed errors would hide how far off each one was.
Example data Projection accuracy compares what owners were told at signup with what the home earned.

How to read it

  • Look at direction and size. Consistent under-delivery points to optimistic projections.
  • Adjust for owner use. If the owner blocked more nights than assumed, note it.
  • Market shocks affect all homes in a market.
  • We do not publish a benchmark. Levels vary by market, season, and property type.

Definitional choices and common pitfalls

  • Revenue basis. Must match the projection.
  • Period. First 12 months live, or calendar year.
  • Owner nights. Adjust the projection for actual owner use, or not. Show both.
  • Error measure. Portfolio ratio plus mean absolute error.

How different roles use it

Role How they use it
Owner relations Prepares conversations with owners whose homes are under projection.
Revenue manager Finds homes where pricing or setup is the cause.
Finance / trust accounting Supplies actuals.
Product / engineering lead Stores the projection as data at signup.

Questions it answers

  • Are our projections too optimistic?
  • Which markets miss most?
  • Do owners under projection churn more?

How to present it in a report

  • Per-home table: projected, actual, error.
  • Portfolio ratio and mean absolute error.
  • Split by market and signup cohort.
Customer report mockupExample data
Projection accuracyHomes onboarded in 2025, first 12 months
Actual vs projected91.1%$442k of $485k
Mean absolute error9.7%5 homes
Under projection4 of 5
HomeProjectedActualError
Beach house$120,000$104,000−13.3%
Condo$90,000$93,000+3.3%
Cabin$75,000$61,000−18.7%
Studio$60,000$58,000−3.3%
Lake house$140,000$126,000−10.0%

Data you need

Fields

Field What it means Example value
Projection Amount, basis, period, model version $120,000 gross rent, first 12 months
Go-live date When the home started taking bookings 2025-04-01
Actual revenue Same basis and period $104,000
Owner nights Assumed and actual 14 assumed, 30 actual

What the raw data looks like

Example data A few rows as they might come out of the source system.

Home Projected Basis Go-live Actual 12 months Error
Beach house $120,000 gross rent 2025-04-01 $104,000 −13.3%
Condo $90,000 gross rent 2025-06-01 $93,000 +3.3%

Where this data usually lives

  • CRM and guest tools hold owner contracts and leads, guest messages, reviews, and upsells. Sales projections often live in spreadsheets.
  • Property management system (PMS) holds units, the calendar with block types, reservations, rates, fees, and often owner statements and trust accounting. For example: Guesty, Hostaway, OwnerRez, Track, Streamline, and Escapia. Some larger managers run their own in-house PMS.
  • Market data providers estimate market occupancy, ADR, and pace. For example: AirDNA, which uses listing data plus partner data (AirDNA), and Key Data, which works with property manager data (Key Data). Their definitions may differ from yours.

The same field can live in more than one system. Agree with the customer which one is the source of truth.

Common data gaps and traps

  • Projections live in sales decks, not in data.
  • Basis not recorded.
  • Go-live date unclear.

FAQ

Is this a standard industry metric?
No. It is a practical check that ties sales projections to results.
Should owner stays adjust the projection?
Show both: raw, and adjusted for the owner's actual use.
How accurate should projections be?
We do not publish a benchmark. Levels vary by market, season, and property type.
For your data teamFormula variants, tables, SQL sketch, and dashboard spec

Formula details

ratio   = SUM(actual) / SUM(projected)
error_i = (actual_i - projected_i) / projected_i
mape    = AVG(ABS(error_i))

Worked example detail

Home Projected Actual Error
Beach house $120k $104k −13.3%
Condo $90k $93k +3.3%
Cabin $75k $61k −18.7%
Studio $60k $58k −3.3%
Lake house $140k $126k −10.0%
Total $485k $442k 91.1% of projection

Tables and fields

Table Fields Notes
projections unit_id, amount, basis, period_months, model_version, created_at
units unit_id, go_live_date

SQL sketch

SELECT p.unit_id, p.amount AS projected, SUM(n.rent_amount) AS actual,
       (SUM(n.rent_amount) - p.amount) / p.amount AS error
FROM projections p JOIN units u USING (unit_id)
JOIN reservation_nights n ON n.unit_id=p.unit_id
  AND n.night_date >= u.go_live_date AND n.night_date < u.go_live_date + INTERVAL '12 months'
GROUP BY p.unit_id, p.amount;

Build spec

Copy this into a ticket. It lists the grain, inputs, logic, and edge cases.

metric: hospitality.projection_accuracy
grain: one row per unit
inputs:
  - projection amount, basis, period
  - go-live date
  - actual revenue on the same basis
logic:
  - ratio = actual / projected
  - error per unit; mean absolute error
edge_cases:
  - less than 12 months live: exclude or prorate, label
  - basis mismatch: block
  - owner-use adjustment: optional column

As a dashboard card

  • Per-home error bars with portfolio ratio KPI.

Common mistakes

  • Different revenue basis for projection and actual.
  • Averaging signed errors only.
  • Comparing partial first years.

Version 1.0 · Last reviewed 2026-10-20 · Machine-readable definition