Guests on Airbnb never meet their host before they book. Yet a single detail in the host’s profile picture, a smile, makes a measurable difference to how often a property gets booked.
A smile in the host’s photo increased demand for a property by 3.5% on average. For an average listing, the researchers calculate that comes to about $672.80 in extra revenue a year.
The study was carried out by Shunyuan Zhang of Harvard Business School, Elizabeth Friedman of Columbia University, Kannan Srinivasan of Carnegie Mellon University, Ravi Dhar of Yale University and Xupin Zhang of East China Normal University. It was published online in the Journal of Consumer Research on August 9, 2024.
Teaching a computer to spot a smile
The team trained a deep learning model, a ResNet-50 network, on public face datasets so it could recognise a smile, gender and age in a photo. They then ran it on Airbnb host pictures.
Their main dataset covered properties of 9,248 hosts in 359 zip codes across seven US cities: Austin, Boston, Los Angeles, New York City, San Diego, San Francisco and Seattle. Demand was measured as the number of days booked each month from January to August 2016. About a quarter of hosts did not smile in their photos.
To compare like with like, the researchers used a weighting method that balanced smiling and non-smiling hosts on host, property and neighbourhood features, and controlled for seasonality in each city.
Bigger payoff where guests feel less sure
The authors argue that a smile works by reducing uncertainty, about the quality of the stay or about dealing with the host. The pattern in the data fits that idea.
A smile did not significantly change demand for female hosts, but it did for male hosts. The authors link this to the common perception that men are less warm than women. In their sample, 66% of male hosts smiled in their photos, against 82% of female hosts.
The smile also mattered more for shared stays than private ones, and more for inexperienced hosts, where a smile generated an extra 6.3% increase in demand compared with an experienced host’s smile. It mattered more in zip codes with higher crime, and for properties without free parking, where the extra lift was about 3.8%.
Experiments point the same way
Because observational data cannot fully pin down cause and effect, the team added online experiments. In one, 1,600 participants on Prolific rated how certain they felt about different listings, confirming that factors such as a male host or a new host raised uncertainty.
Other experiments showed that a smile improved perceptions of the host’s warmth and competence, which raised certainty about the stay and in turn the likelihood of booking. A test that varied only the host’s gender and smile found the effect was larger for male hosts.
The final experiment, with 799 participants, moved outside Airbnb altogether. People rated a listing for either a small family-owned inn or a Hilton, with a photo of the co-owner or branch manager smiling or not. Smiling raised the likelihood of booking for both, from 4.59 to 5.72 out of 9 for the family inn and from 5.08 to 6.02 for the Hilton.
The authors note that a smile could even help level the field on Airbnb, since in their data female hosts did better on the platform overall.
Study Details:
- Title: Serving with a Smile on Airbnb: Analyzing the Economic Returns and Behavioral Underpinnings of the Host’s Smile
- Authors: Shunyuan Zhang, Elizabeth M. S. Friedman, Kannan Srinivasan, Ravi Dhar, Xupin Zhang
- Journal: Journal of Consumer Research
- Publication Date: August 9, 2024
- DOI: 10.1093/jcr/ucae049