Personalized offers bring Tokmanni more visits and sales
Personalized offers increase customer loyalty and sales - 256% increase in mobile coupon usage with smaller average discounts.
The challenge
Tokmanni wanted to increase customer store visits and engagement, but standard coupons offered to everyone weren't working - usage rates stayed low.
What we built
We built a machine learning solution for mobile app offers that selects the products and the discount level for each customer - from the data foundation to models, discount optimization and business rules.
- 01Data
A data foundation for recommendations
We combined sales, customer, product and store data into a unified foundation from which recommendations are recalculated regularly.
- 02Models
Recommendation models learn from purchase behavior
The models learn from each customer's purchases and the order in which products are bought, and use this to estimate which products matter most to that customer.
- 03Discounts
Discount levels optimized product by product
The solution continuously tests different discount levels and learns the one that works best for each product. This way the discount encourages a purchase while staying profitable.
- 04Rollout
Business rules and a gradual rollout
Recommendations pass through the retailer's business rules before they reach the mobile app. Impact was measured with A/B tests, and use was extended step by step to all loyalty customers.
A coupon must never point to a product the store doesn't have
A personal offer for an out-of-stock product is a poor customer experience. That's why a business rule layer sits between the models and the mobile app: every recommendation is checked against store-level availability and redemption limits before it is sent to the customer. The models don't operate as a black box, but within the retailer's own rules.
Results
Impact was measured in a continuous nine-month A/B test. A permanent control group received no personalized coupons, so the results show the solution's true incremental effect. The figures refer to mobile app users.
Coupon usage jumped 256% while the average discount dropped - more sales at smaller discounts.
Mobile-app customers visited stores over 5% more often than before.
Purchases by mobile app users grew 3.5%.
Coupon usage grew 88% right after launch. After nine months of continuous learning and optimization, the increase reached 256%.
“Our loyalty customers now get offers they care about. This means more visits, more sales, and 256% more coupon use. We turned customer insights into action and real business results.”
