Smart Recommendations: A Game Changer For the Beverage Industry

What are smart recommendations?

Smart recommendations come in two types: future behavior recommendations and next best action recommendations.

1. Future behavior recommendations focus on predicting the next best action, such as what you should do next.

2. Next best action recommendations provide a “picture of success” with a particular account.

Picture of success involves segmenting your customers more precisely than in the past. Today, segmentation includes thousands of multiple segments. For example, you might identify segments such as bars with medium pricing for customers aged 25-45. By analyzing detailed data from similar segments, you can determine the best revenue potential for a bar if it maintains a specific assortment and pricing strategy. For instance, if a bar currently doesn’t sell premium whiskey but could potentially do so, you use this analysis to guide recommendations.

In some cases, the next best action can be derived from patterns rather than a comprehensive picture of success—such as leveraging historical data from events like July 4th.

Who benefits most from using smart recommendations?

Typically, the biggest beneficiaries are salespeople who interact directly with customers. In the beverage industry, for example, placing a specific bottle on a particular shelf can make a significant difference. CT works with manufacturers and distributors to assist bars and restaurants by providing sales recommendations based on customer data.

What role does AI play in smart recommendation solutions?

Smart recommendations existed before the advent of AI, but AI has greatly enhanced their accuracy. Previously, multiple rule engines managed thousands of records and dependencies. AI, with its use of Large Language Models (LLMs), can analyze these dependencies much more comprehensively than a human brain could. AI can handle numerous variables—such as weather and consumption patterns—providing a wider range of recommendations.

What are some challenges around implementing smart recommendation solutions?

One major challenge is data quality. Properly structured data is often lacking; in about 90% of cases, data is inadequate or unavailable. Additionally, building the right segmentation is crucial. Customers may want 90% of recommendations to be smart, but without the necessary data, this is not always feasible. We use a maturity model to address this, similar to how you can’t run a marathon without training for shorter distances first.

How would a beverage company know if smart recommendations are right for them?

The impact of smart recommendations may not always be immediately quantifiable in economic terms. However, companies often find that their sales discipline and processes improve significantly. For example, the implementation of smart recommendations can enhance sales department practices and adherence to rules, leading to better overall performance.