Decisions in companies are made by many people, they follow very different methods and dissimilar results are obtained. Sometimes, there are even decisions that are never questioned since, “they have always been made that way”. This is why we often help companies “put science” into decisions and leverage data and technology to ensure that each one is made in the best possible cost/benefit manner.

In an FMCG company, we optimized top-line decisions (price, discounts, promotions) by ensuring an optimal and homogeneous approach among all the people who make similar decisions in different product categories/channels/geographies, with surprising results.

Context

Many companies have decision methods that suit the taste of the decision maker responsible for the indicator they move. Thus, large companies operating in multiple categories and in many regions may have as many ways to decide as there are people involved. On the other hand, decisions are often made more based on historical experiences than on recent evidence, with intuition as the main driver of judgments and emotions as an ever-present ingredient.

A very reputable company in the market that we collaborated with found they had many different approaches to much of their top line decisions and no one could clearly argue which one was best. They called us in to be able to define an approach for these decisions that would ensure the quality of the process, fully understanding that “with better decisions, better outcomes.“

Approach

In our approach, we started by identifying the key decisions and prioritized them based on their impact, focusing on those that move the results needle. When estimating the “size of the prize” behind decisions is possible, all optimization projects are easily prioritized.

With critical decisions already prioritized, we began to analyze them in depth by understanding how they were made. We found multiple reports that were not used, confusing indicators regarding what is relevant, high discretion and lack of criteria to choose alternatives.

With a clear diagnosis, we were able to segment decisions based on their nature and thus propose different approaches to promote them in an economically convenient way. By providing the decision makers key information through “real-time decision data”, decisions that could be “increased” were distinguished from those that, due to their recurrence and structure, could be fully “automated” with advanced analytics.

Results

By optimizing the way in which recurring decisions are made, not only are opportunities to automate those decisions where algorithms do the best job found, but a single way of deciding with a greater incidence of “science” can be homogenized.

As a result of these solutions, immediate improvements are obtained in the quality and speed of decisions, and sometimes significant improvements in resource efficiency. In the medium term, decisions are strengthened by learning insights that allow opportunities to be captured earlier and more precisely. Additionally, the standardization of decision-making processes and the definition of business rules made it possible to reduce the variability of results and ensure common criteria.