Your Foursquare Check-in Can Open A New Starbucks Store

With the development of technology, companies are seeking new ways of cutting costs and increasing profits. Although new technologies such as ERPs and CRMs are mostly adopted by the companies, nowadays another good business practice for companies is utilizing the technology that their customers make use of; the social media. Ever since the social media replaced the face-to-face communication of people, it has been a target of harsh critiques. However, social media is not pure evil; it can be used in different ways as a strategic advantage for the companies. There are a lot of companies that would potentially pay tons of money for your social media data.

Starbucks might be one of those potential buyers. For Starbucks, deciding where to open a new coffee shop is both expensive and very risky business. If they went traditional, they would look at the demographics, the revenue, nearby competitors and the type of the neighborhood before opening a new store in that area. If the predictions were accurate, the store would probably attract thousands of customers daily, if not it could close within a month. Thus, it is really a critical choice for the company to select the area to open a new store, and those traditional methods are not always perfect. Now, there is an easy way to predict whether or not those new stores would be a successful business.

Researchers from University of Cambridge found out that Foursquare check-ins can be used to choose the best retail location. They mixed publicly available Foursquare data from New York City with machine learning algorithms to check if the data would be helpful. They also considered the location of the landmarks; people who checked in at a train station were most likely to check in at a nearby Starbucks. By analyzing 620,932 check-ins shared on Twitter over 5 months. The researchers then determined whether an area is popular, and the flow of the people from one area to another. With the area information, the researchers developed machine learning algorithms and trained them with 2/3 of the available data on popular areas. The algorithms they developed were able to predict the potential popularity of the remaining third areas. The prediction was then compared with the actual data for that 1/3; the algorithm they developed had 76% accuracy. This means that determining the location of a new store using this algorithm would bring 76% of success guarantee.

While the traditional methods would not give Starbucks any number for the probability of success, your Foursquare check-in can do so. By combining the traditional methods with social media data, companies can have competitive advantage, and have higher chances of having a successful business. So, would people have the power of attracting more stores to their neighborhoods by just checking in or would companies keep relying on the traditional methods while opening a new branch, just because they are “not ready” for a new and never before used method?

References:

Dmytro Karamshuk, Anastasios Noulas, Salvatore Scellato, Vincenzo Nicosia, Cecilia Mascolo,;“Geo-Spotting: Mining Online Location-based Services for Optimal Retail Store Placement”; Computer Laboratory, University of Cambridge, UK; June 2013.