Skip to main content

The risk-meter: how AI can predict share loss

Written by: Foresight Strategy

For questions about this case study contact Andrew Glor.

Unless you have been living under a rock, you have likely heard about the growing wave of artificial intelligence. Now with the rise of platforms like ChatGPT, AI is becoming increasingly prominent in both society and the business world, representing a radical change in the way we use data for decision-making.

Let’s define AI

When you imagine AI, what comes to mind? The Terminator? Amazon Alexa? Your Netflix account? In this case, it’s helpful to clarify what it means: AI is “the science of getting computers to act without being explicitly programmed.” The most used case today is machine learning.

Machine learning branches into several variations. What we can use today is Supervised Learning (humans train the machine to classify data) for brand classification and Unsupervised Learning (without human training) for clustering and neural networks. Our neural network technique (Long Short-Term Memory) learns from time-based events and millions of calculations to provide prediction outputs.

AI does not always require new methods; the real difference lies in the promise of a completely automated system that constantly learns and updates in real-time, integrating data from multiple sources efficiently. As we expand the engine to include new sources, the AI continues to improve, incorporating new predictors and testing different analytic approaches without human intervention.

AI in marketing

We’ve been closely monitoring the rise of AI, leveraging our deep expertise in analytics and decision-making to stay ahead of the curve. AI has enormous potential for transforming marketing, but currently, its impact is largely tactical. From targeted content and customer service bots to sophisticated forecasting and dynamic pricing engines, recommendation algorithms, voice recognition systems, and even automated promotional content creation, AI is already making its mark in numerous areas of marketing.

But what about strategy? Although we are nearly 100% data-driven, the role of AI in strategy work is still unclear.

In the world of consumer-packaged goods, brand managers face a challenging task: sifting through a plethora of metrics that range from controllable activities to point of sale and consumer metrics, equity studies, and macro factors. Amidst this dizzying array, it’s not always easy to determine which metrics truly drive growth. What should take priority, and what can be safely ignored? How can brand managers identify when their brand has reached a critical turning point? Will the metrics that matter today still be relevant tomorrow? We set out to test a use case for AI in the realm of strategy making, looking at a project we had worked on with the Coca-Cola global team aimed at identifying the right Key Performance Indicators (KPIs).

Can AI do a better job than humans?

AI in practice

The global team at Coca-Cola needed to track performance across over 40 markets, identify opportunities and gaps in the company’s growth model, and isolate which teams to talk to when issues arise.

We built a prototype of a dashboard using the AI engine to provide a “risk” meter, predicting how likely Coca-Cola is to gain or lose share, based on the KPIs the model has identified. When is it going to happen, what is the expected share change, and which KPIs are driving that prediction?

This system can provide value for multiple brands or markets, with warning alerts sent to the relevant team members when the risk of future share loss crosses a certain threshold. It can also provide predictions and warnings for competitors that are likely to capture that lost share.

The clustering algorithm can identify brands with similar characteristics and different KPIs, looking at which ones could become relevant for the brand of interest in the future – or if a new brand appears in the market, what KPIs are most likely to matter?

Once an opportunity is identified, marketing mix model coefficients can be integrated into the interface allowing the brand team to explore how to solve issues based on their controllable marketing activities such as media spend, price, and distribution.

This model answers a different set of questions from marketing mix models. This is about efficient predictors of share change, across multiple data sources – which may or may not be controllable. Imagine reading the weather forecast. If the main element for success is the number of products you have on display in stores, it’s important to recognize and track that – even as you model all your controllable activities for growth.

The rise of AI is transforming marketing. While its impact is currently mostly tactical, there is potential for AI to play a significant role in strategic decision-making. By leveraging the power of AI, brands can more effectively identify the key metrics that drive growth and stay ahead of the competition.