Regional openness predicted earlier Airbnb, Uber and Flickr adoption across 349 U.S. metro areas

Researchers led by Tobias Ebert of the University of St. Gallen and Martin Obschonka of the University of Amsterdam examined why platform services spread faster in some U.S. cities than in others. In a paper in Research Policy, they link early adoption of Airbnb, Uber and Flickr across 349 metropolitan areas to the average psychological openness of residents, measured on personality data from 3,234,899 people sciencedirect.comResearch Policysciencedirect.comOpen source ↗. Openness is the personality trait that covers curiosity and willingness to try unfamiliar things.

Data and method

Between 2004 and 2015, respondents answered a 44-statement online personality survey and gave their zip code. The team averaged scores by metro area. The averages ran from 3.48 out of 5 in Bismarck, North Dakota, to 3.93 in Santa Fe, New Mexico. Counting a score of 4 or higher as open, the share of open residents ranged from 21% to 48%.

Adoption was estimated indirectly. For hosting and driving, the team used tax records of businesses without employees in accommodation and passenger transportation. Uber reached 142 regions in its first five years. For Flickr, the team sampled 80 million photos uploaded between 2005 and 2015 and counted how many different people in each region posted during the site's first five years.

Results and limits

Regions differed markedly in early adoption: in the five years after Airbnb launched, San Diego gained 9.5 new accommodation businesses per 100,000 residents, against 2.2 in Minneapolis. The correlation between openness and early adoption was about 0.4 for each platform and 0.49 for the three combined sciencedirect.comResearch Policysciencedirect.comOpen source ↗.

A region one standard deviation higher in openness gained about 2.4 more accommodation hosts per 100,000 residents and about 48 more ride-share drivers. For a region the size of New York, the authors calculated roughly 450 more hosts and more than 9,000 more drivers over five years.

The association weakened somewhat but remained after adjustment for demographic, socio-economic, infrastructural, social and other personality characteristics. Openness was the only variable linked to adoption of all three platforms.

The data are a snapshot, so they do not establish that openness drives adoption. Respondents were self-selected online, skewed toward women and younger people, with an average age of 27; reweighting to match the population did not change the results. The authors suggest studies that follow regions over time or treat a platform launch as a natural experiment sciencedirect.comResearch Policysciencedirect.comOpen source ↗.