Please view the main text area of the page by skipping the main menu.

End of the fixed cost era: AI analysis ushers in dynamic pricing

Mikiko Togashi is seen with one of the bags she rents from the Laxus luxury bag rental service, in Shibuya Ward, Tokyo, on April 6, 2019. (Mainichi/Akane Imamura)

"Just by going a few seats further in, you can get such a difference in the price?" Junichi Ochi, 70, a resident of Yokohama, got a surprise while searching online via a tablet for tickets to see the pro-baseball team Fukuoka SoftBank Hawks.

For the Wednesday April 10 game between the SoftBank Hawks and Hokkaido Nippon-Ham Fighters, first base tickets 10 rows from the front near the aisle were priced at 7,395 yen, but three seats further in on the same row cost 6,400 yen -- a 995 yen saving. In the same block, standard sales prices are 8,000 yen, creating a price discrepancy as big as 1,600 yen.

Ochi was looking at the Softbank Hawks' new "AI Ticket" system introduced in January. Prices change according to demand using technology called "dynamic pricing." The system decides on appropriate ticket costs for each seat based on Artificial Intelligence (AI) analysis of data including the game's day and time, the teams, the kind of seating and the starting pitcher. Seating further from the aisles is more likely to be left unsold, so the pricing system gradually cuts fees in increments as fine as a single yen, with prices slowly falling as match day comes closer allowing for detailed valuation.

The idea of one product having one price is being exploded by our ability now to absorb and analyze masses of data in real time. Until now the only businesses running pricing structures that changed in response to demand were those like the airline and hotel industries, which alter their fees to reflect factors such as the seasons. But now AI analysis is enabling us to alter prices on a product-by-product basis.

Hideto Hirata, CEO of Dynamic Plus Co., a dynamic pricing company established in 2018 by Mitsui & Co. and Yahoo Japan Corp., says, "Dynamic pricing is all about reading demand from data analysis."

Systems made by the company are used by sporting bodies like Nippon Professional Baseball and J-League professional soccer. The March 2 soccer match between Yokohama F. Marinos and Vegalta Sendai saw prices slowly rise in the month after tickets went on sale on Feb. 2. But on Feb. 21, with the J-League season poised to begin the next day, a large number of seats remained unsold. The AI used the timing to spur demand by placing huge discounts on some available tickets.

The main SSS seats, typically the third most expensive tier available, were at their unadjusted cost of 5,400 yen on the first day of sales. By Feb. 18 they climbed to their highest fee of 7,400 yen. But on Feb. 21 when the AI's price cut scheme took effect, they dropped in value by about 20%, and on match day itself could be obtained for as little as 4,000 yen, almost half the highest asking price over the sales period.

Dynamic pricing is spreading to other industries too. Universal Studios Japan in the western Japan city of Osaka introduced the scheme to its ticketing in January. Online retailer Amazon uses the software for products it sells directly, with the asking price for goods including pet products and diapers shifting on even an hourly basis each day.

At discount shops run by Trial Company Inc., based in Fukuoka in southwest Japan, all price labeling on the shelves is electronic and subject to change in real time. In the taxi business, Tokyo-based Nihon Kotsu Co. has been trialing variations in the charge for picking up a passenger who books a car based on how busy the service is. Dynamic Plus's Hirata says this trend is set to continue. "Business fields such as goods delivery, which are typically busiest over the year-end and New Year period, and those that change hugely with the seasons like the clothing sector, are probably going to start using this technology in the future."

Meanwhile, "unlimited use services" are also gaining in popularity. Mikiko Togashi, 39, a company employee living in Shibuya Ward, Tokyo, was out shopping with her branded Givenchy handbag in tow. Buying one will set you back hundreds of thousands of yen, but by using the luxury handbag rental service Laxus operated by Laxus Technologies Inc., Togashi is able to use the bags for a relatively low 7,344 yen a month.

Since she started using the service around a year and a half ago, Togashi has rented about six bags through the system. "Buying them is expensive. But this way I can try out bags that catch my eye. It's like I'm free from the price tag," says a smiling Togashi. Currently, Laxus has more than 300,000 subscribers and figures are rising.

Monthly fixed price unlimited use "subscription services" are also becoming more popular. Following the 2008 collapse of Lehman Brothers and the financial crisis, the trend took off in the U.S. as people became more conscious about saving money.

Video and music services such as Netflix, which allow viewers to watch as many movies and TV shows as they want for 864 yen a month, are already a widely accepted form of the subscription model. But recently other companies have expanded into tangible goods like clothes, electronics and cars.

Typically, subscription based services are under pressure to constantly renew their products and services to avoid customers losing interest. The key to keeping those monthly payments active is data. Through the Laxus smartphone app, the service can work out the user's current location. If, for example, it notices a customer is in close proximity to a Hermes store, the app can create a notification recommending the brand's bags. It can also send messages when it detects a phone is near to a restaurant or hotel where patrons are commonly seen with certain kinds of high-end bags.

Strategies like these to promote product use are effective; Laxus boasts customer retentions rates of 90%. Netflix analyzes users' previously viewed content to devise recommendations for other films and TV shows it thinks the customer will want to see. Some 80% of all content viewed on the platform is accessed via recommendations, with specific searches paling in comparison.

(Japanese original by Akane Imamura, Integrated Digital News Center)

Also in The Mainichi

The Mainichi on social media