For a new direct-to-consumer venture, Pepsi Lipton needed a quick ‘proof of concept’ to test recommendations for a personal drinks box based on data. Since this was core to the venture’s offering, the goal was to quickly learn what consumers really wanted and then improve, and thus learn how to drive sales.
SparkOptimus’ recommender allowed us to create a truly personalized offer for new consumers and quickly improve it over time, something we would not have been able to achieve using off-the-shelf systems.
Frank HaresnapeCMO Pepsi Lipton
Together with assortment nutrition experts at Pepsi Lipton, we designed, built in-house, and then tested a product recommendation system within 3 weeks. From a selection of 100s of drinks, the system creates a unique box of 12 drinks based on consumer input via a simple survey (e.g. goals, flavors, and dietary constraints). We continuously improved the survey and the recommendations, making weekly and sometimes daily updates based on consumer feedback and funnel data.
In a matter of 2 months, we improved our spot-on factor (no item switches) for recommendations with 30%, achieving a 100% increase in conversion rate. Furthermore, the average feedback rating on recommended products rose from 2.5 to 4.2 out of 5 stars. Our personalized recommendation system, while new to the market, was crucial in the new venture’s NPS score of +44.