Method
How Watch estimates occupancy and share of voice
We'd rather tell you exactly where our numbers come from. Here is the method, its strengths and its limits.
Everything is public, nothing is guessed in secret
Watch never connects to your Airbnb account or your neighbours'. We work 100% on public data: listing pages, public calendars, AI engine answers. This page explains how we turn that public data into useful estimates, and why we always say “estimated” occupancy.
Your neighbours' estimated occupancy
Nobody outside Airbnb knows a listing's true occupancy. We estimate it with two cross-checked methods, the same approach the big data providers use.
Method 1: reading calendars daily
Every day we snapshot your neighbours' public calendar (availability, price, minimum stay per date). A date that flips from available to unavailable as it approaches is a likely booking. A long block that appears all at once is more likely a host block. By stacking these flips we reconstruct an occupancy estimate.
Method 2: review velocity
On top of that, we watch how fast a listing accumulates reviews. On average, one stay in two leaves a review. From the pace of new reviews and the average stay length, we cross-check the calendar-based estimate.
Why “estimated”
Even data leaders can't perfectly tell a booked date from a host-blocked one, which can overstate occupancy by a few points. So we show “estimated occupancy” everywhere. Our promise isn't an exact number: it's the same method as the market references, applied to YOUR neighbours, and the trend over time.
Share of voice in AI answers
More and more travellers ask ChatGPT, Perplexity or Gemini where to stay. We measure whether your listing comes up in those answers.
A panel of realistic searches
For each market we build a panel of 15 to 30 realistic traveller questions (“where to stay in your city with family near a landmark”, in French and English). Each month we replay these questions against the real engines, several times per question because answers vary, and we look at whether your listing, your neighbours or your area are cited.
What this number is, and isn't
We always talk about “share of voice across our panel of realistic searches”, never the number of times real travellers saw you: that data exists nowhere. Most listings start at zero citations, and that's exactly the point. Moving from zero to one, then three citations is what Watch shifts, and we smooth the number over two months to cut the noise.
How your neighbours are chosen
The 5 tracked neighbours are the comparables your audit identified: same area, same range, same capacity. We watch them continuously and detect new entrants in the same search pass.
Measuring your progress
Beyond the radars, we track your BnBoost score and its visibility, click and booking sub-scores month over month. It's how we show, with numbers, that last month's actions paid off on your listing's public levers.
Privacy and data
We never store a neighbour host's name in your reports: your competitors show up as “Listing A, B, C”. Collected calendars are kept for 18 months to analyse seasonality, then purged. You can request deletion of your data at any time.
Ready to stay a step ahead?
Watch opens soon. Join the waitlist to reserve your spot and your founder price.