How did you obtain this data?
We have a collaboration with AirDNA. This company specializes in scraping Airbnb data: the systematical collection of website-published data. They have developed a system that allows them to scrape on a daily basis. BigData algorithms developed by AirDNA will deduce daily status from historical booking data for each Airbnb listing. As a result, they are able to visualize whether or not they are available, reserved or blocked.
How are the numbers calculated?
All calculations are based on full data files, not just on samples. This means that, apart from the restrictions that apply to working with "scraped" data, the calculations are completely reliable.
How reliable are the numbers?
The numbers are very reliable. Because AirDNA collects data on a daily basis, they have a complete insight as opposed to data scrapers that only make snapshots. Scraping is always an approximation of reality. Triangulation of our findings – comparison with results of studies using other methods - confirms the reliability. Our findings for Amsterdam 2015 were fully in line with other later published studies, and the numbers released by Airbnb itself.
How do you determine whether an apartment is booked?
Previously Airbnb used three status codes: A (Available), B (Blocked) and R (Reserved). In the ongoing "arms race" with data scrapers, Airbnb shielded this information. Now, scrapers can only see A (Available) or U (Unavailable). AirDNA developed an algorithm to extrapolate the status by weighting factors such as historical rental behavior at each listing, the length of unlisted periods and the time since the last booking. The Reserved status of an individual listing can be calculated this way. The data at market level prove to be completely accurate.
AirDNA's website shows different numbers? Why is that?
AirDNA’s website shows a snapshot at this moment in time. This means that the composition of the group may be different tomorrow. After all, Airbnb hosts do not post their listing every day on the site. Our figures represent all total figures because the data is collected daily.
How does AirDNA get these numbers?
AirDNA scrapes every day and collects both "Property Files" with the features of each listing, and "Daily Files" with status updates. The AirDNA files make it possible to calculate the totals over a period, and not just the average. It’s specifically important because listings are continuously retrieved from the platform and reactivated again. AirDNA took reservation information from Airbnb when it was displayed prior to October 2015 allowing them to train the model with this base data set. Furthermore AirDNA ingests real reservation information from data partners which continues to train the model.
What AirDNA’s purpose?
AirDNA is an independent data company that collects Airbnb data, both for commercial market researchers and other forms of research. AirDNA sells the collected data. Our research is the result of a collaboration between Hotelschool The Hague, Colliers International and AirDNA.
You indicate that a part is in the hands of professional landlords. How do you come to that conclusion?
The definition of "professional" can be discussed. In order to make the term measurable, we mainly do this with the number of listings a host offers. In case of a single listing, it is more likely to be the main residence. When a host has several active listings, it more likely involves an investor who is professionally active on Airbnb. Hosts with more than 10 listings are often mediators (so-called "concierge" companies) who have professionalized the exploitation of residential space for tourists.
There are more data scrapers that keep track of Airbnb activity. What makes your research different?
The main difference is that AirDNA scrapes the data on a daily basis and we use the entire dataset in our research. Many studies are based on snapshots and make estimates.
Airbnb claims that the actual numbers are lower. How do you explain the difference?
Airbnb indeed says that the actual numbers are lower. However they do not show any numbers that can support this claim. The difference can only be explained when we can see Airbnb's datasets. Since Airbnb does not fully disclose its numbers it is impossible right now. Our findings for Amsterdam 2015 were fully in line with other later published studies, and the few numbers released by Airbnb itself.
Why does not Airbnb share numbers?
Airbnb does not comment on this.
Do you also have Airbnb numbers of other cities?
Yes, we have the numbers for every city in Europe. We can report on every major city where newsworthy developments occur. In the past year we released report about Amsterdam, The Hague, Rotterdam, London, Berlin, Barcelona and Reykjavík. AirDNA has global coverage so data is available for each city.
Why do you research for Airbnb?
We research all major global developments in tourism. Airbnb is currently one of the most relevant developments that will impact the future of the hospitality industry.