Helping to make make-or-buy decision

Problem: The make-or-buy strategic decision is a difficult one. The Make side of the coin is often time consuming, resource intensive, difficult to maintain, but usually highly flexible and creates risk of third-party. The Buy side of the coin usually saves time and is of higher quality out-of the box. It's typically validated with other buyers. But, you have less control over the solution and simply know less about it.

Approach: Our team was tasked with building a longitudinal supply chain platform that used AI to enable optimizations like Single and Multi-Echelon Optimization models, Constrained Master Production Scheduling modeling, and various other network optimizations. We designed and prototyped this to test the Make scenario. We then pursued co-design with a vendor of new optimization platform to test and understand the Buy scenario.

Results: This process took about a year but we ended up making the platform ourselves. Our data and needs ended up being so complex that the Buy option ended up being incapable of being scaled and customized to our needs. Our platform is currently being developed and is being partly used to support decision-making within a controlled setting.

Lessons: The Make and Buy debate has a lot of ambiguity- can we build it ourselves? can it do what we need it to do? do we even have the data to do what we're trying? how do we test this? For us, the answer to the debate came to us over time. The moment of truth was when we got to a point that we could test optimizations themselves. We learned a lot about how to measure the performance and risks associated with a Make vs. Buy decision, and how to test and cut past ambiguity.