What the pricing agent does

Amapas 353 is a condo in Puerto Vallarta that rents as either a two-bedroom or a one-bedroom. The pricing agent is a small tool with one job: help decide what to charge per night for the one-bedroom configuration.

It does three things:

• Watches the competition. It visits a list of comparable Airbnb rentals and records what each one charges for stays across the coming months.
• Keeps the history. Every price it has ever seen is kept, so a stay that is checked again later shows how the price moved — nothing is overwritten.
• Recommends a nightly rate. For each season of the year (Winter, Summer, holidays, festivals) it combines the competitor prices with your own price floors and produces a recommended rate in US dollars.

The one rule

The recommendation is always the higher of two numbers:

1. A market number, calculated from what competitors are charging that season.
2. Your floor — the minimum price you have set for that season.

The floor always wins. No recommendation ever comes out below a floor you have set. How to set floors is covered on the “Rates, floors, and the season calendar” page.

What it does not do

The tool never changes a price anywhere. It does not touch your Airbnb listing, your booking software, or any website. It ends at a number for you to read; what you do with that number is up to you. (A connection to the booking software is planned as separate work, but nothing here talks to it yet.)

Where you see the results

Results are mirrored into the project’s Google spreadsheet — the same one the old process used. After each run, two tabs are rewritten:

• Observations — the most recent price seen for every listing and stay: who, which dates, how many nights, the price per night in US dollars, and which season the stay falls in.
• Rate Recommendations — the latest recommended rate per season and year, alongside the competitor statistics and your floor, with a “Method” column saying exactly how each number was produced.

Treat those two tabs as read-only: the tool overwrites them completely on every run, so anything you type into them will be erased.

The weekly rhythm

Most weeks you do nothing. The tool is built to run automatically once a week, refresh the spreadsheet, and save its results. (The automatic run is switched off until its first manual run has been checked — see “The weekly run”.) Your recurring tasks are:

• Glance at the Monday run (or at the spreadsheet) to see it worked — see “The weekly run”.
• Once a month, glance at the proxy traffic used in the proxy company’s dashboard; expect about 1.5–2 GB. There is no Airbnb sign-in to look after: the tool reads only what anyone sees without an account.
• Occasionally, adjust your floors or the season calendar — see “Rates, floors, and the season calendar”.

Two honest caveats

The prices include Airbnb’s fees. What the tool records is the “$X for N nights” figure from Airbnb’s booking box. Airbnb now shows that figure under “Prices include all fees”, so the cleaning fee and Airbnb’s service fees are inside it. Whether taxes are inside it too is being confirmed with the first real run. Older prices, brought in from the original spreadsheet, may have been recorded before Airbnb showed prices this way, so a jump between those and the first new prices is not, on its own, a sign that the market moved.

Do not assume the automatic run is working until you have seen it work. Airbnb actively resists automated visitors, and the computers that run the Monday job are the kind Airbnb is most suspicious of. The tool is built to fail loudly rather than record garbage — a run where nearly everything comes back empty stops and reports failure instead of pretending. The “The weekly run” page explains how to do a small test run and what to do if it cannot get through.