My Photo
Blog powered by Typepad
Member since 02/2006

December 2021

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31  

My Motto..

  • "If you tell the truth, you don't have to remember anything." Mark Twain (1835-1910)

My Investing Process

Translate This Page

Become a Fan

Why Real Time Data and AI is Critical to Supply Chain Optimization

Let's contemplate the world of supply chain optimization for a few minutes. Optimization tools underlie most of our decision-making in both supply chain planning and execution. We in supply chain management could not do what we do without the tools we use every day. But are we really getting the most out of our execution and planning tools? Are there other capabilities that could be incorporated into these tools to improve their performance?

Let's start with the underlying engines of these tools--optimization algorithms. There are a significant number of these exotic solution tools available, Academics spend careers fine-tuning these tools to produce better solutions. But algorithmic improvements often yield only tiny improvements in performance. Of course, some tools do provide significant improvements in performance but often only in very specific situations, not necessarily in the broad range of supply chain applications.  How can we enhance the capabilities of optimization tools with new approaches?

Perhaps one of the most important advances over the last few years has been the emergence of real-time data that can enhance the information available to decision-making tools. Examples include weather, traffic, social media, customer behavior, driver behavior, among other information. Adding these data sources to optimization calculations can yield significant improvements in efficient route structures and demand forecasts. Foxtrot Systems,(in full disclosure, an SCV portfolio company), has incorporated real-time data into their optimization tools for local delivery results, yielding improvements in route efficiency of over 30% compared to tools that just use historical data to optimize across cost and service constraints. Similarly, www.solvoyo.com uses social media data on what consumers are saying about products to enhance forecast reliability, often reducing forecast error by 10% or more.

Finally, artificial intelligence (AI) tools can glean important insights from historical data to improve execution and forecast accuracy in supply chains. By carefully examining freight movement and product demand data, AI can reveal hidden trends and tendencies not clear from the raw data. These new insights can be used to develop new constraints on the optimization tools to prevent common errors, such as 'Johnny can't deliver eight stops on rainy Thursdays, only six, so why continue to schedule him with eight'? Similarly, AI can expose trends in product consumption by region or customer segment that may not be obvious from the basic data sets.

We have a long way to go before real-time data and AI become a major part of supply chain decision-making. The principal reason is that legacy supply chain decision support tools generally cannot incorporate real-time data nor utilize AI capabilities. New tools are being developed that sit on top of the legacy systems, drawing out the data they need, then adding real-time data and AI insights to produce better solutions, and returning data back to the systems of record, such as inventory management and freight payments to make the CFO happy.

Finding startups that incorporate new data and data evaluation capabilities is what Supply Chain Ventures spends a lot of time looking for in the supply chain software ecosystem. Many startups promise they use new data sets and AI in their tools, but few do much beyond simple basics which yield only minor performance improvements. You have to kiss a lot of frogs to find a prince, but we'll keep at it.