Helping a retailer to personalize offers on 2 newly introduced digital channels

The retailer

The leading regional drugstore chain (> 170 stores) offers a variety of non-food and food products dominating the market for toiletries, cleansers and health & beauty products in a large metropolitan area. As a traditional family-owned business, the retailer has acquired a large number of loyal customers and high brand awareness in the region.

The task

For more than 15 years the retailer has run a well-established loyalty card (LC) programme. Approximately 40% of shoppers regularly showed their card at the till. In return, they were provided with exclusive, but non-targeted promotions via email and digital print, and collected loyalty points, redeemable via direct mail vouchers.

To improve the efficiency of promotions and develop the loyalty program, the retailer decided to personalize promotions using SO1’s “Optimized Discounts” solution and benefit from an automated AI-based campaign management. The retailer requested to integrate SO1 into the mobile application and a newly introduced communication channel – Check-in Coupon (CiC) kiosks (in-store terminals allowing customers to scan the loyalty card and print personalized promotional offers).

Primary goals for the first 3-6 months of roll-out were

  • to integrate smoothly and rapidly with day-one performance to establish user acceptance for both channels
  • to evaluate and compare the efficiency of the new offers on both channels

The tested solution: SO1 Optimized Discounts

The SO1 AI Engine uses state-of-the-art machine learning to generate an individual offer composition (feed) that best matches a customer’s preference, purchase likelihood and individual willingness-to-pay, called “SO1 Optimized Discounts”. The resulting offers are highly relevant and generate truly incremental sales by avoiding offers for cannibalizing products with a minimum possible discount.


Targeted SO1-offers via the app and CiC kiosks complement each other: both were able to generate sufficient reach and redemption performance.

From week one on, the number of shoppers retrieving individualized offers via the retailer app increased steadily. In comparison, the build-up and promotion of store kiosks required more lead time at first but proved to be very effective in the long run. After 30 weeks, more than 10% of cardholders were targeted by individual offers through one of the two channels. Among those, 73% retrieved offers via the CiC kiosks and 27% via the app. Only a small share of customers (5%) used both channels.

Fig. 1 & 2: Impression rate (App view or Kiosk print) as a percentage of reached cardholders. Values are depicted for the initial roll-out phase (left) and after a period of 30 weeks (right).

Across both channels, AI-offers performed well. We measured a single campaign redemption rate of 4.9% for the app and 8.3 % for the CiC kiosks, significantly outperforming standard leaflet redemption rates (2.1%). To some extent expected, the CiC kiosks performed comparably better due to targeted customers already being in the store and therefore increasing chances of successful conversions. Throughout both channels, the SO1 AI triggered sales at an average discount of 17-18%, far below the average discount on standard leaflet promotions (26.3%).

Fig. 3: Redemption performance of SO1 “Optimized Discounts” offers distributed either through the retailer’s existing app or at the Point-of-Sale kiosk.

Today, two years after the start of the roll-out, the SO1 AI has continued to learn and improve results: 10% redemption rate per offer at 15.1% discount only – and we are increasing efficiency further.

Fig. 4: Redemption Rate and average discount needed (Multichannel- Kiosk + App) at the end of the initial rollout (+30 weeks) and two years after.



Retail technology news