How to build a hotel comp set that does not lie to you in 2026

Arshad Kacchi · Founder, RevPerfect · 24 May 2026 · 12 min read

A GM I worked with at a 140-key CBD property pulled up the monthly benchmarking deck and pointed at a comp index reading 117. Big number, big smile. I asked who was in the comp set. Three of the five hotels had not refurbished since 2014. One sat on the wrong side of a railway line that had become a noise hotspot. The fifth had repositioned as long-stay corporate three quarters earlier and had stopped competing for our segment. The headline 117 was not a sign of outperformance — it was a sign the hotel comp set had drifted from operating reality, and the index was measuring a fight that no longer existed.

Below is what a hotel comp set is in 2026, the seven rules I use to build one, where it quietly breaks, and the audit cadence that keeps it honest.

What a hotel comp set actually means in 2026

A hotel comp set, short for competitive set, is five to seven hotels chosen because they compete with you for the same booking on a typical night. It is the denominator behind every competitive index a revenue manager reads. RGI, MPI, ARI, share-of-market — all of them divide your performance by the average of the comp set. Get it wrong and every downstream index lies for as long as the wrong list is in the data feed.

Two things have changed in recent years that make comp-set rigour more important than it has ever been. Micro-markets move faster — a new build, a brand conversion, or a major refurb can change the local substitution map in a single quarter. And owners are reading indices more often, not less. A stale comp set does not just produce bad pricing — it produces a bad performance review.

Operating versus aspirational

The most useful distinction is the split between an operating and an aspirational comp set. The operating set is the five to seven hotels you actually compete with — it drives pricing, the rate-shop, and the segment story. The aspirational set is a wider group you would like to position alongside over time, used for capital planning and owner narratives. Both are valid. Mixing them up — running daily pricing off an aspirational list — is one of the most common ways a comp set silently misleads a property for years.

The seven selection rules

Here are the seven rules I use to decide whether a property earns a place in the comp set. A candidate has to clear at least five. Clear all seven, it is a core member. Five or six, it goes in with a note. Below five is out, regardless of how much the GM likes the building.

1. Same micro-market. Within roughly two kilometres in a CBD, ten in a regional market, or the same airport precinct. If a guest searching online sees both properties on the same result page, the geographic test passes.
2. Same product tier. A four-star full-service property is not in the comp set of a three-star limited-service property, even on the same street. Product tier defines the willingness-to-pay band.
3. Comparable room count, within roughly fifty percent either side. A 200-key hotel benchmarked against a 40-key boutique reads distorted indices, because the small property's pricing is dominated by compression nights the larger property does not feel.
4. Similar segment mix. Stay within roughly fifteen percentage points on the largest segment. Otherwise the indices reflect the segment gap, not the rate-and-occupancy gap.
5. Active competitor on the typical booking. Run a meta-search query for three or four typical date ranges. If the candidate consistently appears alongside your property, it is a real competitor.
6. Stable operating model over the last twelve months. A hotel mid-refurb, mid-rebrand, or mid-segment-pivot is not benchmarkable. It rejoins after it settles.
7. Reasonable data availability. If you cannot read the candidate's published rates reliably, the comp set has a hole. Acknowledge it rather than papering over it with the next closest substitute.

The discipline is to write the rules down and apply them as a checklist. Hotels that do this annually produce comp sets that hold up. Hotels that decide the list in a room over coffee end up with the 117-with-the-wrong-properties story I opened with.

The math under the indices the comp set feeds

The indices the comp set feeds are simple division problems, and a division problem with the wrong denominator is unrescuable. Once they land on the slide, the indices do not carry a footnote that says the denominator includes one hotel that does not really compete with us anymore.

RGI (revenue generation index) = your RevPAR ÷ comp-set average RevPAR × 100. MPI (market penetration index) = your occupancy ÷ comp-set average occupancy × 100. ARI (average rate index) = your ADR ÷ comp-set average ADR × 100. The mathematical identity that ties the three together: MPI × ARI ÷ 100 ≈ RGI.

The deeper treatment of how to read the three indices in the correct order lives in the companion piece on RGI, MPI, and ARI explained. The point here is that every one of those indices is built on the comp set you chose. The comp set is the only input under your direct control. Choose it well and the indices tell you the truth. Choose it poorly and the indices become a slow-acting lie.

Worked example. A 120-key suburban property I worked with had a comp set including one large airport hotel three kilometres away — higher product tier, much higher corporate share. Comp-set ADR A$268, occupancy 78 percent, RevPAR A$209. Subject property: ADR A$224, occupancy 81 percent, RevPAR A$181 — MPI 104, ARI 84, RGI 87. The story: "you fill the building but you cannot hold rate." When we replaced the airport hotel with two suburban properties of the right tier and segment mix, comp-set ADR fell to A$232, occupancy held at 77 percent, comp-set RevPAR A$179. Subject property: MPI 105, ARI 97, RGI 101. Same building, same month, a different operator conversation. The first version would have driven a rate cut. The second drove a rate hold. The comp set was the lever.

Where the comp set breaks down

Even a well-chosen comp set will quietly degrade if no one is watching it. Here are the four failure modes I see most often in practice.

1. New supply that nobody added to the list

A new hotel opens within the micro-market. For six months it ramps occupancy with promotional rates, which is normal, and the natural response is to leave it out of the comp set. Twelve months in, the new property has stabilised at a higher rate than the comp-set average — and your indices are now reading a denominator that excludes the property your guests are most often comparing you to. Fix: add new supply the moment it crosses one full year of operation.

2. A comp-set hotel that quietly repositioned

A property in your comp set undergoes a refurbishment, a rebrand, or a segment pivot. From the outside, nothing visible has changed — same address, similar room count, updated but not radical photos. Internally, the property has moved from a four-star full-service positioning to three-and-a-half-star limited-service, or vice versa. Indices that include the repositioned property will drift for months before anyone notices.

3. Aspirational creep

Over time, comp sets tend to drift upward. Owners ask for a more expensive comparison hotel to be added and nobody removes one in exchange. After two or three years the comp set has migrated from "five hotels I really compete with" to "seven hotels I would like to position alongside." Indices steadily fall. The natural reading is "we are losing share." The actual reading is "we are now measuring ourselves against a different fight."

4. A comp set that uses different data sources

None of these failure modes are dramatic — each shifts the indices a few points at a time. Over twelve months, those few points compound into a meaningfully different operator conversation. The audit cadence below is what catches them.

What to do about it: the five-step playbook

Here is the routine I walk every revenue team through when they want to put their comp set on a sound footing and keep it there.

1. Run the seven-rule checklist as an annual audit. Block ninety minutes. Tick or cross each property against each rule. Below five out of seven leaves the comp set; new candidates scoring five or higher come in. Date the audit, archive the previous version.
2. Document the source for each property's data. One line per property: where the rate comes from, what tax basis, what cadence, who owns the feed. The discipline is to make data drift visible.
3. Maintain two comp sets in parallel. The operating set for daily pricing and the rate-shop. The aspirational set for owner narratives and capital decisions. Label them clearly.
4. Set a mid-cycle trigger list. New supply opening, a major refurb completing, a brand change, a segment pivot, or a sustained fifteen-percent-plus ADR movement at a comp-set member all trigger a same-week re-audit.
5. Bring the comp set forward on every owner deck. One slide near the front: the five properties, the rules they passed, the date of the last audit. The slide costs nothing to maintain and makes every downstream index footnote-free.

The hard part is institutional: getting the GM and the asset manager to accept that the comp set is not a permanent fixture, and that changing it is a sign of rigour rather than goalpost-moving. Once that is settled, the audit runs itself in ninety minutes a year.

A real scenario (anonymised): the 80-key boutique on a repositioning curve

An 80-key independent boutique I worked with sat on a high street in a regional capital. Comp set static for four years: three other boutiques and two branded mid-scale properties. Indices had drifted — RGI 104 four years ago, RGI 94 in the most recent quarter. The owner read the slide as a performance problem and was within a quarter of asking for a GM change.

We ran the seven-rule audit. Two of the three boutiques had quietly refurbished and repositioned to a higher product tier. One branded mid-scale property had moved its segment mix from sixty percent corporate to forty percent after a major contract loss. The subject property had not moved at all. The indices were reading a real gap, but the gap was that the comp set had drifted upward around the property.

We rebuilt the comp set: two originals remained, three new properties came in, all five cleared at least six of the seven rules. The subject property posted RGI 101, MPI 109, ARI 93. The story changed from "the building is losing share" to "the building fills well but is priced about seven percent below the segment-correct market" — which led to a methodical rate-floor reset rather than a GM change. Twelve months later RGI sat at 105.

Comp set and the broader benchmarking stack

A comp set is one input, not the whole stack. Pace, pickup, segment mix shift, and the forward demand calendar sit alongside it. The cleanest comp set will still mislead if the underlying forecasting routine is weak. The companion piece on hotel pickup and pace covers the daily reading order that pairs with a healthy comp set, and the wider revenue management strategies for 2026 piece places comp-set rigour into the broader operating routine.

FAQ

Closing

I built RevPerfect because the existing tools assume the comp set you start with is the comp set you should keep. We treat the comp set as a living document — versioned, audited, and footnoted on every chart it feeds. Try RevPerfect free → or book a 20-minute walkthrough and I will show you what a clean comp-set audit looks like on your own data.