Let me be blunt about something most of you already suspect but paid seven figures to avoid saying out loud.
Your supply chain digital twin is probably a slideshow.
It renders beautifully. It shows every node, every lane, every DC in real time. It updates. Executives love the demo. And it cannot answer the only question that matters: if I do X tonight, what breaks tomorrow, and what is the better move?
That is not a twin. That is a very expensive rear-view mirror.
The word “twin” is doing a lot of lying
The academic literature has finally caught up to what those of us in the forward-deployed seat see every quarter. A recent study on digital twins in supply chain management investigates the implementation of Digital Twins in Supply Chain Management, highlighting the gap between their conceptual promise and practical applications.
That gap is not a rounding error. It is the whole game. The same research is direct about why: DTs are recognised for their potential to correct real-time deviations and anticipate and prevent disruptions as they emerge; however, operational deployments in SCM remain rare. Numerous studies mislabel simulation models or Digital Shadows as DTs, blurring essential distinctions.
Read that twice. Most of what your vendors call a “twin” is a digital shadow. It reflects the physical world back at you. It does not simulate a different one.
A shadow answers “what is happening.” A twin answers “what happens if.” Those are not adjacent capabilities. They are different products with different physics, different data contracts, and different compute bills. And the industry sells you the first one wearing the price tag of the second.
Here is the tell. Ask your team to close a port for 48 hours in the model and rank three response plans by margin impact. If the honest answer is “we would need to schedule a modeling run” or “let me check with the vendor,” you bought a shadow.
A twin has to run faster than the disruption
The value of a counterfactual decays by the hour. A perfect answer to “should we reroute” delivered on Thursday is worthless if the ships left Tuesday.
This is where efficiency-era supply chains get exposed. As one industry breakdown put it, supply chains have spent decades optimising for efficiency, leaner inventory, faster throughput, lower cost-per-unit. The last five years have exposed the limitation of that approach: systems optimised purely for efficiency are fragile when disrupted. The events that break you do not wait for your batch cycle. Port closures, demand spikes, supplier failures, and weather events do not follow efficiency models. They break them.
So the design spec for a real twin is not “high fidelity.” It is “fast enough to change tomorrow morning.” A twin that produces a gorgeous, defensible answer in three weeks is a research paper. A twin that produces a good-enough, ranked recommendation in twenty minutes, before the next replenishment cycle, is an operating asset.
Most programs optimize for the wrong number. They chase visual fidelity and node count because that demos well. The number that actually matters is wall-clock time from question to ranked action. Nobody puts that on the slide.
And there is a data reality underneath all of it. If your IoT coverage is patchy, your carrier data arrives in batch cycles, or your inventory systems are not connected in real time, the twin’s model of your supply chain will have blind spots that limit its predictive value. You cannot simulate tomorrow on a feed that describes last Tuesday.
The CODN nobody put in the business case
Let me reframe this the way I frame it in every steering committee: the Cost of Doing Nothing.
Not the cost of the project. The cost of the shadow you already own pretending to be a twin.
CODN is every expedited freight charge you ate because the model could describe the delay but could not test a cheaper response in time. It is every markdown on the wrong 40,000 units. It is every stockout during a demand spike your dashboard flagged beautifully and did nothing about. It is the planner who stopped trusting the tool and went back to a spreadsheet and a phone call, which means you are now paying a license fee to reproduce a rear-view mirror your team already ignores.
That number is enormous, and it hides because it never shows up as a line item. It shows up as margin that quietly leaks while everyone admires the map.
The fix is not more nodes. It is a narrower, sharper mandate. Pick three disruptions that actually hurt you: a key supplier going dark, a regional demand spike, a lane closure. Demand that the twin run those counterfactuals end to end, produce ranked actions with margin deltas, and put an answer in a planner’s hands inside one shift. Kill everything that does not serve that loop.
A twin that cannot run tomorrow is not early. It is decorative.
Stop funding the slideshow. Fund the simulation, or admit you bought art.
The retailers who win the next disruption will not be the ones with the prettiest map. They will be the ones whose twin already answered the question before the disruption finished arriving.