The Use of Knowledge in Supply Chains

Economics is the study of the efficient allocation of resources. One of the reasons it is such a frustrating field is that its predictions can be so wrong, particularly when it comes to macroeconomics. If you know everyone’s preferences and you know everything there is to know about a problem and you know exactly how each of the mechanisms works, then you can prescribe perfect policy. These assumptions make the problem mathematically tractable. We can model it out with equations that we can solve. This is what they teach in school.

Friedrich Hayek, the Nobel-winning economist, wrote about the dispersion of knowledge in his essay for the American Economic Review entitled “The Use of Knowledge in Society” (1945). This hobbles the attempts at optimization.

“The peculiar character of the problem of a rational economic order is determined precisely by the fact that the knowledge of the circumstances of which we must make use never exists in concentrated or integrated form but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess. The economic problem of society is thus not merely a problem of how to allocate “given” resources—if “given” is taken to mean given to a single mind which deliberately solves the problem set by these “data.” It is rather a problem of how to secure the best use of resources known to any of the members of society, for ends whose relative importance only these individuals know. Or, to put it briefly, it is a problem of the utilization of knowledge which is not given to anyone in its totality.”

There is no one place where we can find all the knowledge we require. It is scattered. For the economy, it sits in the hearts and minds of millions of producers and consumers. For an enterprise or a government agency, it exists in myriad databases and in the minds, experiences, and relationships of the employees. In fact, some of the requisite information may exist outside of the agency in the databases and staff of other organizations, such as partners in the supply chain, or regulators, or capital markets, or clients.

Some of it you can get by asking. Some of it nobody could tell you, even if they wanted to.

Michael Polanyi was a chemist and philosopher who spoke of a second kind of knowledge: tacit knowledge, as described in this report from the London School of Economics:

“… tacit knowledge collects all those things that we know how to do but perhaps do not know how to explain (at least symbolically).”

Tacit knowledge lives with the people who keep supply chains running: planners, warehouse staff, contracting officers. It stands apart from the explicit knowledge held in ERP systems, contracts, and other tools. ERP systems bill themselves as the single source of truth, but anyone who has worked in a bureaucracy knows that is rarely true in practice.

How does each firm integrate all explicit and tacit knowledge, including intelligence that exists outside its immediate ken, to improve operating decision-making?

Hayek's answer was prices.

“Fundamentally, in a system in which the knowledge of the relevant facts is dispersed among many people, prices can act to coordinate the separate actions of different people in the same way as subjective values help the individual to coordinate the parts of his plan … The mere fact that there is one price for any commodity—or rather that local prices are connected in a manner determined by the cost of transport, etc.—brings about the solution which (it is just conceptually possible) might have been arrived at by one single mind possessing all the information which is in fact dispersed among all the people involved in the process.”

To see how this works, consider a semiconductor supply chain.

Imagine a fabless design company that sells GPUs. They hire a contract manufacturer (known in the industry as a “foundry”) to produce the chips. This foundry then ships the finished product to the design company which, in turn, sells and ships the chips to OEMs and device manufacturers for inclusion in their computers, servers, gaming consoles, and other products. The OEM sells to end buyers such as data centers or enterprise IT groups.

Each company in the supply chain has its own systems containing its firm-specific explicit knowledge and each has staff possessing tacit knowledge.

In this chain, tacit knowledge presents as the process engineer who can feel the production run drifting, or the demand planner who has an intuitive sense of the unusual seasonality associated with a customer’s product launch.

The explicit knowledge flows through systems, however imperfectly. The tacit knowledge stays with individual people.

Getting both kinds of knowledge across firm boundaries takes three things, and it works only to the extent that companies are comfortable sharing information.

A common language. In our semiconductor chain, "lead time," "capacity," and "yield" mean slightly different things at the foundry, the design firm, and the OEM. Until they mean the same thing, data can't be combined, compared, or trusted. Sextant AI builds a common ontology across firms, so explicit knowledge from disparate systems reads the same way for everyone.

A market for intelligence and risk. The foundry asks the OEM about demand and pays for the answer. Sextant AI verifies the authenticity of the response. The foundry’s tacit knowledge about OEM demand and what it means for the design company’s orders can be wrong or biased. This is why it is valuable for the foundry, which never sees the OEM directly, to supplement its view with the OEM’s intelligence. Based upon this exchange, the foundry can be comfortable pricing a capacity option for the design company. The premium the design company agrees to pay combines the foundry's capacity plan, the design company's demand forecast, the OEM's demand forecast, and the judgment each party brings to them into a single price.

People supervising agents. Agents can integrate data, validate it, and run workflows such as demand forecasting and inventory optimization. People stay in the loop to supply what the model lacks, make the calls, and override it when they know something it doesn't. The agents use the explicit knowledge. The planners bring their tacit knowledge to bear when necessary. The decision then becomes a part of the explicit knowledge base, further enriching the experience of both the agents and the planners.

Textbook models in economics assume someone who knows every preference, every constraint, every mechanism. No supply chain has one. The knowledge is split across firms, and part of it lives in people who couldn’t write it down if you asked. A common language and a marketplace let the knowledge travel. People working with agents make the call, and the decision leaves a record that improves the next one. No one holds the entire picture, but each firm can see more of it than before. That is a better, more feasible goal than the perfect plan the textbook assumes.