Skip to main content
All CollectionsData Methodology Data Methodology
How AirDNA Identifies Duplicate Listings
How AirDNA Identifies Duplicate Listings
T
Written by Tom Williams
Updated over a week ago

Many hosts chose to list their short-term rental listing on multiple listing channels, such as Airbnb or VRBO, giving their listing increased visibility and a higher likelihood of bookings. Having two separate listing pages for one listing means we pick up property performance twice for the same listing. To ensure the highest accuracy with our data, we match listings on multiple platforms and combine their performance figures.

How Does AirDNA's Property Matching Algorithm Work?

First, we identify every Airbnb and VRBO property within 600 meters of each other. We use the latitude and longitude in our database for each property to do this. This way, we can narrow the possible matches to a small subset.

We then extract each property’s information to determine similarity. We use booking behavior, price information, and property attributes. We employ more than ten attributes to ensure we capture as many signals as possible.

We then compare each attribute from the smaller subset between every close Airbnb and VRBO property. This way, properties that have similar characteristics will have higher similarity scores.

After this, our machine learning model considers all these similarity scores and gives us a probability of two properties being a match. We take the highest chance of all the possible matches. If the probability is too low, we won’t consider it a match.

Did this answer your question?