A few career highlights using optimization:
- Optimized Packaging Plan for a Major Ice Cream Producer: Crafting a packaging strategy that respected the complexities of multiple production lines and frequent changeovers. A challenging yet delicious project!
- Distribution Network for a Large Wine Company: Designing an optimal distribution network that considered multiple warehouses across the US to efficiently serve various distributors.
- Optimized Purchasing Plan for a Grocery Chain: Creating a cost-effective purchasing plan for cattle and chicken packs based on deli demand for different cuts of meat.
Transitioning into the wine business about twenty years ago, I initially focused on the “front of the house” — sales forecasting, finished goods inventory management, and packaging. Eventually, I delved deeper into winery operations and the intricate world of winemaking. As I learned more about blending wines, a thought began to percolate: Could the same optimization techniques from my earlier projects be applied to creating wine blends?
Because there are so many choices to consider when putting together a wine blend, surely there’s an optimal set of decisions that can be made. The idea was to harness an optimization engine to build out these blends. I envisioned a scenario where we could use mathematical models to create perfect blends.
But then reality hit me.
[Cue record scratch]
“Winemakers are going to be reluctant to try this if we only focus on price.”
This was the immediate reaction — and a recurring one — from wine industry colleagues when I briefed them on the idea. They were concerned that an algorithm might propose blends that make financial sense but don’t consider all the things it takes to make great wine.
The hurdle this feedback created was daunting: designing a comprehensive optimization tool (not only financially-focussed) that could navigate the complexities of the wine business, accounting for each client’s unique requirements. It was an enormous task, and I wondered if the effort would be worth it.
Is the juice worth the squeeze?
Despite the skepticism and the daunting scope of development, the idea persisted. It kept resurfacing, each time with a bit more clarity. I’d discuss it with Lars Verspohl, a data scientist friend of mine that has helped build other tools for us, and Matt Strauss, one of our developers at Claret who is deep into this nerdy kind of thing. We’d brainstorm, sketch out formulas, test small concepts, and then set it aside. It was still too big of a tool to build. Would it be worth investing both money and time into something that may not meet the winemakers needs?
December 2023
During the holiday break of 2023, I revisited my research notes, formulas, and various proofs of concept. I was also getting my head across the latest advancements in artificial intelligence (AI). I experimented with integrating these cutting-edge technologies into a new set of models. Weaving in modern AI techniques proved to be the breakthrough I had been seeking. The results were promising, reigniting my determination. Gradually, a viable solution that properly balances things like costs, quality and origin of grapes or wine lots began to emerge.
However, I realized that a model alone wasn't sufficient. The true "secret sauce" lies in combining deep expertise in both optimization and the wine industry. Simply deploying an application like this doesn’t guarantee its value. Effective use of an optimizer demands a lot of skill and intuition. Much like an F1 driver needs to understand the intricate dynamics of their car and develop an instinctive feel for its performance, working with an optimizer requires not just mathematical know-how but a nuanced understanding of why it behaves the way it does. To truly make this service thrive in the wine industry, intimate knowledge of wine production and industry intricacies is essential.
As I started to share the prototype with industry peers, I was thrilled to see immediate recognition of its value and receive constructive feedback.
Over the last few months, we’ve meticulously refined and enhanced the model. Coupling this new model and business consulting, Claret Profit Maximizer is now poised to become a crucial business resource for companies making wine and other beverage alcohol (e.g. whiskey, tequila, etc.).
Here is a video that walks through Claret Profit Maximizer.
Next Steps:
Conquer the world! Ha, just kidding. But I do need to conquer the initial winemaker resistance. I need to clearly show that Claret Profit Maximizer is a tool, akin to a refractometer or a sophisticated tank cooling system. It’s here to assist, not replace the artistry of winemaking, but to provide a solid foundation for blending decisions.
Frequently asked questions
less time forecasting
