Rule-based targeting versus an AI solution for email recommendations

The partner:

The international company is one of the top providers of customer intelligence solutions with worldwide business activities. It serves over 1,200 organizations, among them 15 of the world’s 25 largest grocery retailers and thousands of retail brands.

individual-email-couponing

The task:

For two years, the partner has provided its client, a large US grocery chain with more than 400 stores, with an individual selection of coupons for every household that has opted to receive weekly product recommendations via email. Prior to working with SO1, the targeting relied on a rule-based system profiting from the available loyalty data on the household’s past shopping behavior. The partner wanted to improve the efficiency of his targeting. Any new solution should improve:

  • the participation rate (i.e. the redemption of at least one out of all offers per household) of a large number of non-redeeming households, which regularly receive offers but rarely visit the client’s stores
  • the (single campaign) redemption rate of email offers, in particular in household segments where little or no shopping data is available
  • the amount of manual staff work required to maintain a rule-based recommendation system

The tested solution:

The SO1 AI Engine uses state-of-the-art machine learning that deeply understands product properties and customer preferences. It calculates for each customer the individual purchase probability and the expected unit sales for every active campaign. The algorithm also favors a composition of product promotions that both enhances user engagement (redeem, drive-to-store) and increases basket size (complementary items). The resulting offers are highly relevant and satisfy the individual needs of the customer as well as his or her desire to explore new products and categories.

 

Test setup:

To minimize integration efforts and ensure continuity in the partner-client relationship SO1 designed a test pilot where the mailing process at the retailer would remain as is. For a period of six weeks, the promotional offers for 20% of randomly selected households, previously targeted with the partner’s conventional solution, were replaced by individual offers selected by the AI-Engine.

Thanks to a dedicated team on both ends, integration was achieved within eight weeks measured from the kickoff meeting to the first AI-based output.

Results:

The results show an overall increase in the redemptions of individual coupons across all segments: SO1’s AI clearly outperforms the rule based system. Customers targeted by AI redeemed at a +82% higher rate compared to the rule-based system. In addition, the participation rate also increased, on average by +60%. This increase was particularly strong within a segment of households with a large number of previously inactive users, for which little purchase data was available. Compared to the conventional system, the number of users in this segment that redeem one or more offers increased by +109%.

Overall, SO1’s AI solution demonstrated a strong performance advantage both with respect to coupon redemption rate and user engagement. Surprisingly, compared to an already successful rule-based personalization approach, the AI proved particularly successful in the reactivation of rare shoppers.

 

SO1-case-study-results

Figure: Participation rate for email promotional offers in individual user segments (little/no, medium, extensive purchase data available) and redemption rate (all segments) for SO1 AI-based targeting (blue) and the existing rule-based targeting (gray).

 

Outcome:

The overwhelmingly positive results, achieved with minimal integration and staff effort, led the partner to choose SO1’s AI as the preferred solution for its personalized marketing module and flagship product to be rolled out at its new and existing retailer clients.

At the specific retailer, it was decided to expand traffic and profit by using the SO1 engine’s feature to automatically adapt when a retailer’s business priorities change. While ensuring constant user engagement, the engine was configured to maximize the retailer’s return-on-investment by selecting high-margin products and minimizing the individual discount amounts (results coming up in a different case study).

 


PS: SO1 is one of the driving forces of retailer digitalization. We have created a very powerful AI for retail which is capable of personalizing promotions for users in real-time and across devices. The SO1 Engine sources the entire portfolio of the retailer and automatically selects the right products for each individual consumer and adjusts discounts such that revenue, profit, or consumer satisfaction are maximized. To learn more, reach out to:
US Sales: Patricia A. Cucinelli, cucinelli@so1.ai, +1 917 757 6221
EU Sales: Stephan Visarius, visarius@so1.ai, +49 160 93 59 69 95

 



We use cookies, just to track visits to our website, we store no personal details. Read more on our privacy policy.