If you’re trying to get through a busy railway station, you might think you’re planning your own route through the crowd, but you’re probably following a stranger — even when you don’t realise it, and even when it’s not the fastest route.
The unseen “crowdsourced” pathing was uncovered from watching how train passengers navigate around a platform obstruction at Eindhoven Centraal station.
Researchers analysed three years of anonymous tracking data from passengers leaving trains on one platform. By observing how people chose between two routes around a kiosk, they found a clear pattern: many passengers copied the path taken by the person immediately ahead of them, rather than choosing independently.
This behaviour wasn’t limited to friends or family.
Using mathematical techniques to identify people travelling together, the researchers filtered out the friends and families, leaving behind individuals travelling alone. What remained showed a strong “stranger-following” effect, where individuals still treated the person in front as an informal leader.
Once someone chose a particular route, others were more likely to follow, creating small cascades of people making the same decision. These knock-on effects meant crowds did not always spread out evenly or take the most efficient paths through the railway station.
The findings suggest that crowd movement is shaped less by rational choice and more by imitation.
For transport planners, this has practical implications. Clear signage, better sightlines, and early guidance can matter more than simply providing multiple routes. If the first few people move the “wrong” way, many others may follow.
That suggests that people at the front of a crowd need strong directional signals about where to go, since those behind them will ignore any signs that do exist. That could mean larger, more visible arrows and signs at the front of the space where crowds might congregate, but also fewer of them further back in the station. So fewer signs, but larger and more focused in placement.
Understanding how easily crowds fall in behind a single, unknown individual could eventually help design stations that move people more efficiently – and more safely – especially at busy times.
The research Z. Wang, A. Gabbana, & F. Toschi, Avalanches of choice: How stranger-to-stranger interactions shape crowd dynamics, Proc. Natl. Acad. Sci. U.S.A. 123 (8) e2528167123, was published in the Proceedings of the National Academy of Sciences.
