Author: Duke Perrucci
Over the past 25 years, I have worked for a wide variety of companies – from Fortune 50 giants to small start-ups. The common thread that runs through all of these experiences is the immense challenge of driving consistent growth. Whether you are trying to grow the second largest soft drink brand in the world with a seemingly limitless budget or a nascent technology company that is struggling to conserve its series A funding, the business environment is tough as hell. Decision making in such a highly complex and competitive environment is fraught with peril – in fact, according to a recent McKinsey survey, only 20% of executives say their organizations excel at decision making.
One of the main reasons why good decision making is so rare is that many organizations use intuition, rather than data and analytics, as the basis for their decisions. Fortunately, there are many data-driven, AI-based technologies – such as mathematical optimization – available today that help organizations make decisions that maximize their efficiency and profitability. Unfortunately, I didn’t have access to mathematical optimization tools for most of my career – and, in many instances, this resulted in lost opportunities and lackluster growth.
Thinking back on my career, I have identified three situations where having a mathematical optimization tool would have really saved my bacon:
#1: Solving the Olive Oil Supply Problem
During my time as a brand manager at a multinational consumer products company, I was tasked with helping a struggling olive oil business. The main issue the olive oil business was facing was that access to raw material and prices simply fluctuated too much to maintain a healthy margin on the product. The raw materials for the oil were sourced from all the traditional olive-growing countries around the Mediterranean – Italy, Spain, Tunisia, Greece, France, and Turkey. If, for example, a drought hit Tunisia or a plague of locusts swept through Spain, supply would plummet and raw material costs would skyrocket.
The complex process of sourcing raw materials and blending them to a consistent flavor profile was done by hand (and mouth). The sourcing team would buy what they could, and the master taster would taste the oil to ensure it was blended to the right profile. If – at that time – I had access to a mathematical optimization tool, this complicated blending problem could have been solved to an optimal or near optimal solution and I could have extracted every penny of efficiency out of production. Our inability to find a solution to this sourcing volatility ultimately led to the divestiture of this $400-million brand due to inconsistent profit contributions.
#2: Determining the Right Marketing Mix
When I was head of client services at a data analytics provider, I had the opportunity to advise a global beverage company on how to utilize data and analytics to power their business. Each year we would run multivariate regressions to calculate and assess the impact of most of the significant marketing mix elements – TV advertising, radio, outdoor, in-store promotions, pricing, in-store real estate, etc. The regressions did a reasonably good job of identifying the contribution that each marketing lever made to a brand’s overall volume and revenue for the year.
By doing these calculations, a marketer could gain an understanding of the efficiency of each lever and decide on where to put more or less investment next year. However, there would always come a time when the brand manager would ask what the ideal ratio of mix elements and spend was to drive the highest revenue growth. These were high-stakes questions when you consider that the marketing budget for these brands ran from $100 million to $900 million. And, unfortunately, those were questions we just couldn’t answer with the tools we had. Ultimately, the brand team had to “guess” at the best allocation based on data and that could have cost them millions in revenue. A mathematical optimization tool could have told us precisely how to best allocate the marketing spend to optimize revenue growth.
#3: Creating the Perfect Media Plan
As head of the commercial business at a behavioral science consulting firm, I worked with Fortune 500 companies to help them drive greater advertising effectiveness. What made the business so unique was that it blended vast amounts of customer data with social media data and to that, our data engineers would append behavioral profiles. To put it simply, our data scientists would find lookalikes in the data – people that already showed an affinity towards a particular brand. Once identified, we would advise on ad copy, imagery, and trafficking to achieve the best response rate. Then we would use our data management platform to deliver those impressions to the right web property at the right time – and it was this part of the process that was done without much analytical rigor. Our digital team would buy the properties that got the most traffic for a given profile.
Without a doubt, this process could have been run with far greater efficiency and precision had we used mathematical optimization. With the objective of driving the lowest cost per conversion, we could have found the ideal mix of ad type, spend, and properties to deliver the optimal media plan to these customers. Digital advertising provides a brand with myriad options to build awareness and drive conversion, but finding the most efficient and effective mix without mathematical optimization is simply not possible.
Seeing the Opportunities
As I look back on my career, I can see many other missed opportunities where – because the companies I was working for didn’t have an AI tool like mathematical optimization to help us make data-driven, optimal decisions – we left money on the table or fell short of growth targets.
And now, as I work for the leading provider of mathematical optimization software and support, I can see so many opportunities for optimization in the business world and so many businesses out there today using this technology to reach new heights of efficiency and profitability.
Are you missing out on opportunities for growth by not using mathematical optimization technologies? I would encourage you to learn more about mathematical optimization – so that you can begin to identify and explore possible applications of this AI technology in your organization.