Most operators spend hours configuring PriceLabs and ten minutes on their comp-set. That ratio is backwards. Your comp-set is the foundation everything else sits on. Get it wrong and your dynamic pricing engine is optimizing against phantom competition.

The default behavior is to grab whatever Airbnb or Key Data surfaces automatically, call it a day, and move on. What you get is a sample that mixes a two-bedroom apartment with a six-bedroom villa, a place that sleeps four with one that sleeps twelve, and a professionally managed property with a host who logs in twice a month. That is not a comp-set. That is noise with a spreadsheet attached.

Bad comp-set data does not just give you bad market intelligence. It actively misprices your calendar, causes you to undercut when demand is strong, and panic-hold rates when the market is actually soft. You are flying blind while thinking you have instruments.

What a Comp-Set Is Actually Supposed to Do

A comp-set is a proxy for the competitive pressure your property faces on any given night. It tells you what the market will bear for something comparable to yours, in your submarket, at this moment in the booking window. That is a specific and narrow job.

It is not a tool for benchmarking your average nightly rate against the entire destination. If you have a three-bedroom townhouse in Palma’s old town and your comp-set includes rural fincas twenty minutes outside the city, the signal is meaningless. Urban and rural inventory in Mallorca do not compete with each other. Guests searching for walkable bars and restaurants are not cross-shopping a countryside property. Your comp-set has to reflect actual substitutes, not geographic proximity.

Think about it from the guest’s perspective. What else would they book if your property was unavailable? That list is your comp-set. Everything else is irrelevant data you are paying to be confused by.

The Filters That Actually Matter

Bedroom count is the most important filter and the one most frequently applied too loosely. A one-bedroom difference in a smaller property matters enormously. A two-bedroom in Lisbon’s Bairro Alto and a three-bedroom in the same neighborhood are not interchangeable. They attract different party sizes, different budgets, and different booking windows. Start with an exact bedroom match, then consider loosening by one if your market is thin.

Guest capacity matters separately from bedroom count. Two properties can both have three bedrooms and sleep four people or twelve people. That gap changes everything about the rate they can charge. In Ibiza, a villa sleeping ten near Es Cavallet is competing against a completely different guest pool than a villa sleeping six in the same zone. Cap capacity at plus or minus twenty percent of your own property.

Quality tier is the filter operators are most reluctant to apply honestly. You need to be realistic about where your property sits. A well-maintained mid-tier apartment in the Algarve should not be comping against properties with professional photography, designer interiors, and a hundred five-star reviews. That is not humility, that is accuracy. PriceLabs lets you filter by review score. Use it. Set a floor at 4.6 and a ceiling that matches your tier. If you are a four-star-equivalent property, do not comp against the market’s top ten percent.

Amenities are worth filtering where they create genuine demand separation. Pool versus no pool in Tuscany or Marrakech is a binary market split. Guests searching for a pool villa are not seriously considering properties without one. If your property has a pool, your comp-set should require one. The same logic applies to beachfront in the Caribbean, ski-in access in alpine markets, or a rooftop terrace in Barcelona. One missing amenity filter can collapse your average comp rate by fifteen percent because you are now measuring against a cheaper product category.

If you have not revisited the amenity filters in your PriceLabs settings since you first set them up, that is almost certainly costing you money right now.

Building the Set by Hand

Automated comp-set tools are a starting point, not a final answer. Key Data and Wheelhouse will surface candidates based on algorithm logic you cannot fully see. That logic is reasonable, but it is not calibrated to your property’s specific competitive position. You need to spend time inside the actual listings.

Search Airbnb for your location with your exact dates, your bedroom count, and your guest capacity. Sort by total price. Look at the first thirty results. Open each one. Read the description, look at the photos, read recent reviews. You are asking one question: would a guest who is seriously considering my property also seriously consider this one? If yes, it belongs in consideration. If the answer is obviously no, remove it regardless of what the algorithm says.

Do this across three date windows: a peak weekend, a mid-week shoulder period, and a low-season window. The properties appearing consistently across all three are your actual competition. The ones that only show up in peak are either pulling from a different demand pool or are managed opportunistically and will not track your pricing challenges in the months that matter most.

Aim for fifteen to twenty-five properties in the final set. Below fifteen and a handful of outliers will skew your averages badly. Above thirty and you start diluting the signal with properties that are only loosely competitive. In thin markets like a specific village in Umbria or a niche zone in Dubai, fifteen may not be achievable and that is fine. A smaller, accurate set beats a large, polluted one every time.

When to Rebuild and Why Most Operators Never Do

The competitive landscape of any STR market shifts constantly. New inventory enters. Properties get delisted, change ownership, or dramatically raise or lower their quality through renovation or neglect. A comp-set you built eighteen months ago is describing a market that no longer exists.

Rebuild your comp-set at minimum twice a year. In high-growth markets, do it quarterly. Barcelona and the broader Catalonia market have seen enough regulatory pressure and supply changes over the last two years that a comp-set from 2024 is actively misleading in 2026. The same is true for Dubai, where new hotel-apartment supply has repositioned what mid-tier short-term rentals can realistically charge. If you are tracking Spain specifically, the STR market dynamics in 2026 make this non-negotiable.

Rebuild triggers that should prompt an immediate review: a major new competitor opens near you, a large property in your set disappears from the platforms, your occupancy diverges significantly from market occupancy reported in Key Data, or you complete a renovation that moves you up a quality tier. That last one is where operators leave the most money on the table. You upgraded the kitchen and added a hot tub, but you are still comping against the version of your market that existed before you made those changes.

What Bad Comp-Set Data Does to Your Pricing Over Time

A corrupted comp-set does not produce one bad night of pricing. It produces systematic mispricing that compounds across every week of the year. If your set is pulling down your market average because you have included cheaper, lower-quality properties, your dynamic pricing tool will interpret strong demand as normal and never push rates to where they could go. You leave ten to twenty percent on the table during your peak weeks and do not see it because your occupancy still looks healthy.

The inverse is equally damaging. If your comp-set skews toward premium properties above your tier, your market rate looks artificially high. You hold rates too long, fill too late, and your last-minute discounting ends up erasing the margin you thought you were protecting. This shows up clearly when you run a proper revenue audit: occupancy looks fine but revenue per available night is soft. The 42-point revenue audit framework treats comp-set accuracy as one of its first diagnostic checks for exactly this reason.

Beyond Pricing, Wheelhouse, and PriceLabs are all capable tools. The comparison between them matters less than most operators think. What matters more is whether the market data you are feeding any of them is accurate. A good algorithm running on a bad comp-set will optimize you confidently in the wrong direction. You can read the tool comparison for European markets after you have fixed the data problem underneath it.

The comp-set is not a setup task you do once. It is an ongoing position you maintain. If you cannot tell me right now who your top fifteen competitors are and when you last verified each of them is still on the market, you do not actually know what your pricing is competing against.