How is the pricing algorithm configured?
For each listing, the algorithm works in a few steps: it identifies the property, builds a set of comparable listings around it, calculates a market-based starting price from that set, applies adjustments for factors like day of week and season, and produces a full year of daily prices.
It refreshes automatically twice a day. Settings that define the comparable set — such as how closely bedrooms and guest count must match and how many comparables to include — are adjustable.
Who does the algorithm consider as competitors?
The algorithm looks at other comparable listings on the same booking platform as yours, matched by bedrooms and guest capacity, then ranked by proximity to your property. It excludes listings flagged as unusual or no longer active, and it does not use your own listing's asking price.
Does the algorithm factor in my unit's booking pace?
No — pacing is not a documented input. Pricing is built from comparable listings in your market, not from your own unit's booking pace. There are lead time and market occupancy factors, but those are derived from the comparable set, not your individual unit's performance.
How many competitors does the algorithm consider?
Up to 150 comparable listings by default (adjustable), with a minimum of about 10 before results are treated as low-confidence. It selects the closest matches with no fixed distance limit, so in thinner markets it will naturally reach further out.
Does the algorithm price for today or for future dates?
It produces a full year of forward-looking prices. It anticipates future conditions through day of week, season, year-over-year trends, and lead time — pricing for what is expected on each future date, not just current conditions. The lead-time factor also adjusts as a date gets closer.
How many data points does the algorithm consider?
Roughly 150,000 data points per run — approximately 200 comparable listings across roughly two years of daily rate and availability history. It analyzes a large historical picture rather than a single snapshot.
How does the algorithm determine seasonality?
Two ways:
- One factor captures weekly and seasonal demand patterns across comparable listings.
- Another compares this year's market rates against last year's for the same month.
The year-over-year comparison requires at least one year of history to take effect; if that history isn't available, this factor has no impact.
What data points and weights does the algorithm use?
The algorithm starts with a market-based price from comparable listings, then applies five adjustments:
- Day of week
- Season / week of year
- Lead time
- Year-over-year trend
- Market event allowance
The final price is the starting price adjusted up or down by the combined effect of those factors. Note: the exact calibration of each factor is still being finalized and specific weight values are not yet documented. For detailed questions, please contact our development team.
How does the algorithm forecast demand?
By modeling how booked comparable properties are over time and how lead time relates to occupancy. For dates beyond approximately six months out, availability signals from one platform are set aside as unreliable at that range.
How does the new algorithm differ from the old system?
The old system is a closed box that cannot be inspected. The new one:
- Produces results in the same format (no changes on your end)
- Is built on transparent market data with explainable, visible adjustments
- Bases the starting price on the market, not your own asking rate
During the transition, both systems run side by side for comparison.
Does the algorithm pull in my own portfolio units when calculating pricing?
The starting price never uses your own listing's asking rate. Whether your other portfolio units could appear among the comparables is not explicitly addressed in the documentation — they could appear if they meet the bedroom, guest count, and distance criteria. For a definitive answer, please contact our development team.
Can I add my own custom comp set?
Not currently. The comparable set is built automatically and there is no feature to hand-pick your own comparables. Some matching settings will soon be adjustable, but a custom comp set option is not available at this time.