China’s bike rental startups are learning a lot about how people spend their free time

Next Story

Media Prima buys Rev Asia for $24M to create Malaysia’s largest digital media platform

Editor’s note: This post originally appeared on TechNode, an editorial partner of TechCrunch based in China.

China’s bike rental startups have hogged the startup limelight and investor dollars this year. The trend has been impossible to ignore, both on the ground and in reporting. Ofo became the industry’s first unicorn, while Mobike has pulled in over $300 million in funding this year alone. The influx of cash has cast much attention on the space including, principally, how these companies propose to become viable businesses when they charge barely a dollar for a ride.

Well, we have some color on that potential after Mobike released a fascinating report analyzing user data from the recent Labor Day holiday weekend. The insights provide one of the first pictures of what can be built from bike rental data and an idea of how the company wants to reveal what it knows. Indeed, market intelligence and insight could also be an area where the startup could monetize its business in the future.

The ‘May 1st Little Long Holiday Cycling Report’ was generated by the Mobike Big Data Artificial Intelligence Platform — or Mofang (魔方 ‘Magic Cube’) for short — and it illustrates a change in behavior from Mobike users over the holiday weekend, which ran April 29 to May 1, compared to regular weekends and commuting patterns.

The data falls roughly into two different categories: national trends and behavioral change, or places versus people.

National

Overall, during the three-day holiday weekend, usage was up 17.2 percent nationally from the weekend before, with the largest spikes seen in tourist destinations rather than first-tier cities, up 51 percent in Xiamen, 47 percent in Hangzhou. Tianjin saw the biggest jump among China’s big cities at 36 percent, down to 26 percent for Shanghai which was fifth on the top city list.

Chengdu had the highest proportion of female riders at 47.7 percent followed by 45.7 percent in Beijing, and 43.3 percent in Shanghai.

Growth spike for Labor Day weekend compared to previous weekend for Tier 1 cities: Tianjin, Guangzhou, Shenzhen, Beijing, Shanghai (top to bottom)

The report tracked the top five ranked cities and there is no mention of the locations which had either very small increases or falls. All in all, the national average was a 17.2 percent increase, but Mofang provided further interpretation of the figures.

The top five places that attracted the highest Mobike use was listed as Beijing, Shanghai, Guangzhou, Shenzhen and Chengdu.

The report includes the “top 5 hottest tourist destination sites” for visitors who registered their account in other parts of China: Beijing, Shanghai, Guangzhou, Chengdu, and Xiamen. This is where the user tracking starts to become more apparent.

Behavior

The report outlines trends of who is going from which city to where: Beijing and Shanghai users head to Chengdu, while Guangzhou and Shenzhen city folk head to Haikou and Xiamen for a change of pace.

Then whether users stay in their hometowns or travel to another location. They’re tracked as they head to scenic spots (in their tens of thousands in some cases) and their behavior monitored: they cycle longer than normal, later in the day, visit more places, and in larger groups.

Graph of use over time of day with the morning peak delayed over the holiday weekend

Superficially the data shows the impact the bikes are having on leisure time. The lifestyle advertising seems to have made headway as touristy areas thronged with cyclists. By revealing the data behind the trend, Mofang demonstrates how applying AI to massive amounts of data is going to create a valuable resource for Mobike as it builds a more detailed picture of its users’ lives.

Selling targeted data to travel companies (or setting up your own) is one thing, but the fact the company is showing that when you head out for a holiday cycle, it knows who you’re cycling with and when, where, at what speed, and whether you’ve traveled together from somewhere else. All this hints at just what else that Magic Cube might glean from its database.