How to Choose the Hotels You Compare Yourself To
Your comp set decides every benchmark number you will ever read, and you pick it yourself. The selection criteria, the right size, and a worked example.
Your comp set — short for competitive set — is the small group of hotels you measure yourself against. Pick hotels a guest would genuinely choose instead of you: same catchment, same quality tier, average rates within roughly 30% of yours, same kind of guest. Six to ten properties is the normal size, and STR requires at least four reporting hotels before it will produce a benchmark report. The uncomfortable part is that you choose the list yourself, which means you can accidentally choose a flattering answer.
Why this one list decides everything
The three benchmark indexes every owner eventually asks about are all ratios against your comp set. Nothing else.
- MPI (Market Penetration Index) = (your occupancy ÷ comp set occupancy) × 100
- ARI (Average Rate Index) = (your ADR ÷ comp set ADR) × 100
- RGI (Revenue Generation Index) = (your RevPAR ÷ comp set RevPAR) × 100
An index of 100 means you captured exactly your fair share. Above 100 means more than your share, below means less. Those are STR's definitions and they are not controversial.
What is controversial is the denominator. Change the hotels in the list and every one of those three numbers changes, without a single thing changing at your property. That is the vulnerability.
The four criteria that actually matter
Location. Not distance in miles — substitutability. Would a guest searching for your dates seriously consider both properties? A hotel four minutes away on the wrong side of a highway, serving a different exit, may not compete with you at all.
Quality tier. Similar age, condition, star rating, amenity level. A property with a conference floor and a restaurant is not your competitor just because it is nearby.
Rate range. Comparable ADR — average daily rate, meaning room revenue divided by rooms sold. A common working rule is within about 30% of yours. Beyond that you are comparing different purchase decisions.
Segment. Who stays there. A hotel filled by corporate contracts and a hotel filled by weekend leisure travelers can sit on the same street and never compete for the same guest.
The size rule, and why it exists
Six to ten hotels is the working norm. STR will not produce a report unless at least four properties in the set are reporting data — that minimum exists to protect confidentiality, since with fewer contributors a hotel could reverse-engineer a specific competitor's numbers from the aggregate.
Too few and one competitor's renovation, closure, or bad month swings your whole index. Too many and the set blurs into "the market", which tells you about the town rather than about your competitors.
If you cannot find enough genuinely comparable hotels, widen carefully on one criterion — usually geography — and write down what you widened. You will need to remember that when the numbers look strange.
Selection criteria at a glance
| Criterion | Include if | Exclude if | Common error |
|---|---|---|---|
| Location | A guest would treat you as interchangeable | Different catchment, exit, or island | Using a radius instead of guest behavior |
| Quality tier | Comparable condition, rating, amenities | Full-service vs limited-service mismatch | Including the newest hotel in town "to aim high" |
| ADR | Within roughly 30% of yours | Twice your rate, or half of it | Adding a luxury property for prestige |
| Segment | Same guest type and booking window | Contract, group, or extended-stay dominant | Ignoring a corporate-contract house next door |
| Size | Broadly similar room count | 300 rooms vs your 45 | Assuming rooms count does not matter |
| Data | Reports to your benchmark provider | Non-reporting | Building a set that cannot produce a report |
A worked example
You run a 45-room independent boutique, ADR around $180, mostly weekend leisure with some midweek corporate. Six candidates nearby:
| Candidate | Rooms | Est. ADR | Verdict |
|---|---|---|---|
| Riverside Inn, 0.4 mi, boutique | 38 | $172 | ✅ In — same guest, same rate band |
| Harbour Hotel, 0.9 mi, boutique | 52 | $205 | ✅ In — top of the band, direct rival |
| Grand Central, 0.6 mi, full-service | 240 | $215 | ❌ Out — scale and segment differ |
| Budget Lodge, 0.3 mi, highway | 80 | $89 | ❌ Out — half your rate, different decision |
| The Old Mill, 2.1 mi, boutique | 30 | $190 | ✅ In — worth the extra distance |
| Airport Suites, 5 mi, extended stay | 110 | $145 | ❌ Out — different segment entirely |
Three qualifiers is not enough. You widen on distance — the criterion you are most confident about — to bring in two more boutique properties in the next town that show up beside you in search results. That gives you five, and you note in your file that the set is geographically wider than ideal.
Budget Lodge is the interesting exclusion. It is the closest hotel to you and you probably lose a handful of price-shopping bookings to it every month. It still does not belong, because including it would drag your comp set ADR down and make your ARI look better than it is.
Primary and aspirational sets
Keep two lists and be clear about what each one is for.
Your primary set is where you are now — the hotels you actually trade against, and the set your owner should judge you on. That number should be defensible.
Your aspirational set is where you want to be — a tier up, or the properties you will compete with once the refurbishment finishes. Track it, learn from it, and never report it as your performance benchmark. Chasing an index against a set you do not belong in produces one outcome: you look like you are failing while doing perfectly good work.
How often to review it — and when not to
Once a year, on a fixed date, is the right cadence. Also review immediately when something structural happens: a new hotel opens, a competitor closes or reflags, a nearby property finishes a major renovation, or your own positioning changes.
What you must not do is review the set because you did not like this month's RGI. That is how comp sets quietly drift downward until every index reads above 100 and none of them mean anything. Write down the reason for every change you make. If the reason is "the number was low", stop.
What the evidence says about where hotels sit
Cornell researchers analyzed 67,008 hotel observations from 2001 to 2007 (Enz, Canina & Lomanno, Competitive Hotel Pricing in Uncertain Times, Cornell Hospitality Report Vol. 9 No. 10). They found roughly 54% of hotels were priced below their comp set in both 2001 and 2004 — the majority of the industry sat under its own benchmark.
The same dataset shows what that costs. Hotels priced 20–30% below their comp set ran about 15 points higher occupancy and about 12% lower RevPAR. For independent hotels specifically, that 20–30% discount band showed a RevPAR shortfall of 16.31%, while sitting 5–10% above the comp set showed +4.12%.
Those numbers only mean something if the comp set is honest. The authors note their data shows correlation, not causation — but the arithmetic of the index is not in doubt: garbage denominator, garbage index.
The mistakes that make your indexes meaningless
Building the set from a map radius. Guests do not book by radius. They book by search results and substitutability.
Including the nicest hotel in town for ambition. It inflates comp set ADR, deflates your ARI, and produces a permanent sense of failure.
Including the cheapest hotel to feel better. The mirror-image error, and the more common one.
Quietly editing the set mid-year. Your year-over-year comparison is now broken and nobody will remember why.
Never revisiting it after a new opening. A 120-room arrival changes your market whether or not you add it to the list.
The bottom line
Choose hotels a real guest would genuinely choose instead of you — same catchment, same tier, rates within about 30%, same guest type. Six to ten of them, at least four of which report data. Keep a separate aspirational list and never confuse the two. Review annually and after structural changes, write down every edit and its reason, and refuse to edit the set because you disliked a number.
Do that and MPI, ARI, and RGI become the most useful numbers you have. Skip it and they are decoration.
If you would rather see your comp set as a day-by-day rate table — your rate against each competitor's best available rate, with sold-out flags and your daily rank — that is what our competitor analysis view does. Choosing the hotels is still your call, and it should be.