Is Your Supply Chain Hampered by Shiny Object Syndrome?

Supply chain leaders love shiny objects. The shiny object syndrome happens when teams focus undue attention on a new and trendy idea and drop it as soon as something new takes place. Six years ago, the focus was on blockchain. Investments in the supply chain using blockchain deployments failed. Tradelens, the largest and most successful supply chain blockchain deployment, discontinued operations in the first quarter of 2023.

Today, the shiny object is Artificial Intelligence (AI). Most event agendas remind me of a kindergarten child throwing icicles at the holiday tree. Intoxication reigns. I find AI discussions everywhere, but the business use cases are few and far between. The discussions on AI lack a grounding in definition and a clear value proposition. I love outcomes. My goal is to drive value.

The Issue

Today, tension abounds. The most advanced analytics to enable AI are available from the top cloud computing companies, Amazon, Azure, and Google. However, supply chain leaders face a dilemma when adopting these more advanced technologies.

The issue? Manufacturers buy software. They are not builders. The analytics from cloud service providers are best suited for a build strategy. Supply chain leaders saddled with legacy packaged software maintenance costs are unsure what to do.

As shown in Figure 1, supply chain software innovation is not a normal distribution. In 2000, it was a bell curve with an equal number of early and late adopters. Today, the late adopters outnumber early adopters by a 3:1 factor. So, while many business leaders waft eloquently about shiny objects, they are late followers. As a result, when supply chain leaders speak on “innovation,” they give voice to the late adopter perspective.

Figure 1. Software Adoption Cycle

As a result, the packaged software market moves much slower than the evolution of tech capabilities. The gap is growing. Is this a problem? Maybe. It depends on the problem being solved.

The Promise

Adopting concepts like Large Language Models (LLM), Ontological Frameworks, Graph Databases, Vector Databases (Vector DB), NoSQL, and Schema on Read is slow. Existing technologies move structured data efficiently using relational database technologies to improve enterprise transactions, but the processing of unstructured data is an opportunity. Eighty percent of the data surrounding the supply chain is unstructured—text, images, and streaming—but is not used.

The buy strategy assumes software providers will drive innovation, but it takes time. There is also the fallacy of industry analyst help. When an industry analyst firm lacks independence (it is heavily financed by traditional technology providers) or an understanding of the Art of the Possible, the intoxication worsens, resulting in a jacked-up hype cycle.

The Tension

The answer is not easy. Most should take a deep breath. Forms of AI— machine learning, narrow AI, and pattern recognition—are evolving based on schema-on-read databases but most companies are investing in schema-on-write technologies (relational database structures) using traditional packaged software solution taxonomy definitions. This approach does not address the larger opportunity.

And I don't care how many times you try to dress up a pig with three- and four-letter fancy acronyms using forms of AI, it is still a pig. Right?

The problem? Software providers are automating traditional software definitions. Innovation is low. The current focus is to improve transactional efficiency and market insights to improve processes within a function —sales, marketing, R&D, manufacturing, procurement, and transportation. The solutions are inside-out and designed to use enterprise data better, but the market is shifting from an inside-out process focus to an outside-in business process flow.

Using unstructured data and emerging AI techniques offers great promise for better sense and intelligent response. The question for business leaders is, “How to get started?” The conundrum includes:

1. Academics Focus on Engines. I once thought that academics could close the gap. I do not see this happening. Academics tend to be backward-looking, focusing primarily on better math. I find few academics familiar enough with packaged applications to be helpful.
2. Old Tech Is Expensive to Maintain. While packaged software is cheaper for the initial install, it is more expensive over time. Today, based on studies by ASCM, 80% of software purchased is shelfware. (Purchased, but not used.) While the risk of employee turnover and failed projects is lower, maintenance costs (often 15-22% of the initial software purchase) are expensive. The cost of evolving on changing platforms is millions of dollars. (For example, the movement for SAP customers to adopt HANA is a platform change not covered by maintenance.) The question is, “How do we maintain and evolve?” The answer is complicated.
3. New Tech Is Not for Everything and Everyone. The use of newer technologies cannot be broad-brushed. Avoid generic discussions of AI. Instead, focus on clarifying definitions, improving capabilities, and aligning AI tactics with business value. The number of resources needed to understand both the supply chain potential and the promise of newer forms of technology is a constraint. Cloud-based technology providers provide tools but not process automation.
4. Faster and Hands-Free Processing of existing technology is not the Best Answer. The global supply chain was built on the assumption that demand and supply variability would be low and that government policies would be rational. Neither assumption is valid today. As a result, the importance of market data is growing. Functional efficiency throws the supply chain out of balance, reducing organizational effectiveness. The digital transformation agenda espoused by large system integrators is mainly self-serving.
5. Knowledge is Essential. Understanding the potential of new tech platforms requires unlearning. The power of the Graph and Vector DB offers the best results when we throw away conventional thinking. Our definitions of supply chain planning quickly become outdated when we map outside-in flows and apply ontological thinking.

Wrap-up

In summary, help your team side-step the shiny object syndrome. Get clear on definitions and value. Align your Information Technology strategy (IT) based on these insights and your cultural DNA.

As AI matures, delay the investment in significant platform investments. Move forward if and only when the organization is clear on outcomes.

I welcome your thoughts!

For Additional Reading, Check Out These Blog Posts:

Redefining Your Relationship With Data

Dogs That Do Not Hunt

Patterns Tell the Story

Rise or Rise-up?

Case Studies on the Use of Outside-in Data