One of Pecan’s customers is a large, multinational brewery. They develop, manufacture and distribute renowned and popular beer brands in dozens of countries around the world.
The company sources ingredients from suppliers all over the globe. The company also supplies products and services to thousands of buyers who distribute them to pubs, shops, supermarkets, vending machines, etc.
The company invests millions of dollars each month to promote its +100 beer brands, in thousands of points of sales spread across over 50 countries.
The company was searching for an AI platform to most accurately predict which “blend” of promotions will deliver maximum revenues and yield the highest ROI. The results of the AI platform would then be compared against the existing promotion plan, which was based on calculated elasticities and spend levels.
The company decided to enrich its data with external market data from Nielsen and utilize Pecan to generate a formula that predicts which combination of discount, retailer, beer brand, and date will generate maximum revenue from the promotion budget.
The company used Pecan for data preparation and removal of outliers. Next the data cleaned using Pecan, was used to train a deep learning predictive model. All this took 10 days, compared with 4-6 months required by alternative solutions, who consume huge and expensive resources, such as data scientists.
The AI models were built by the customer’s business analysts, using the Pecan platform, and not by a trained, and expensive, data scientist.
Before turning to Pecan the promotion plan was rule-based for all the products.
Using Pecan, the customer was able to predict which combination of different discounts will maximize revenues. For example: 5% discount on beer type 1, 7% on beer type 2, and 13% discount on beer type 3.
Furthermore, Pecan was able to predict which promotions of specific SKUs will be wasteful, or yield a low ROI.
Using Pecan, the company was able to increase revenues from promotions by +20%, with the same promo budget.
This means that assuming that before using Pecan, the company invested $100,000 in promotions and generated revenues of $ 100,000, using Pecan the same investment of $100,000 in promotion yielded $125,000.
In addition, the Pecan platform uncovered 8 factors which contributed most to the success of the promotion.
Some of the most common contributing factors were:
Most significant, was that the brewing company empowered its BI team with the capability of building state of the art AI predictive analytics models, without the need to hire and train data scientists for the job. The company now has a Pecan champion with data science capabilities as well as strong BI and marketing knowledge. The perfect combination to succeed.