One night. Two ways to arrive at a price.
A base price model adjusts a number you gave it — change the base, the answer moves with it. An optimization model uses a choice model to estimate the booking probability at every price, and stops at the price with the highest expected revenue (price × booking probability). Set both running on the same July 4th and watch what each one does.
Base price model
Answers: “How should I adjust the base rate you gave me?”
Optimization model
Answers: “What price makes the most money tonight?”
Same base rate. Same comp set. Four different nights of demand.
The base price model returns the identical number every time, because demand was never one of its inputs. Optimization moves with the night.
| Real demand for your night | Base price model | Optimized price | Expected revenue — base | Expected revenue — optimized | Difference |
|---|
How each model gets to its number
Same night, same data available to both. The difference is what each one does with it.
What a base price model actually is
A market-following multiplier. It averages what your competitors are asking, expresses that as a factor, and multiplies it by a base rate you chose. Its ceiling is the quality of that guess — and when the whole market misprices a night, it misprices it too, with full confidence.
What optimization does instead
It uses a choice model to estimate the booking probability of your specific listing at each candidate price, multiplies price by that probability, and returns the maximum. Comps are one input to demand rather than the answer itself, so the price reflects what the night is genuinely worth to you.
Illustrative demand curves for explanation, with 12 comps to keep the chart readable. In production Quibble attaches about 30 comps to each listing, puts the 10 best matches still available each night into the model, and estimates each night’s booking probability with a discrete choice model, using Bookable Search data, comps, events and seasonality.