The problems I get called in for

Problem 01

Recurring problems

What it looks like: The same shortage, expedite or write-off returns each season. Each time it is blamed on a supplier, a planner or bad luck.
What is going on: Recurring problems come from feedback and delays, not bad luck. Orders, stock, staffing and capacity react to each other late. Those late reactions recreate the problem.
What to do: Map the loops with the people who run them. Then test changes on a model before trying them on the network.
See it working: Supply Pulse shows how quality holds, shelf life, hiring and a distant supplier turn a demand peak into empty shelves and expired stock. Lag Lab splits an overshoot into its causes.
Where it fits: Disruption Diagnostic

Problem 02

Overshoot after a demand step

What it looks like: Demand steps up and production, stock and hiring overshoot. The review blames the forecast and buys a better one.
What is going on: Overshoot after a demand step is the structure of planning, not the forecast. Part of it is stock and pipeline that must be refilled. Most of the rest comes from the decision rules.
What to do: Separate the necessary refill from what the rules add. Then change the rules, which is cheaper than a better forecast.
See it working: Lag Lab runs the same step three ways and splits the overshoot exactly. In its example, a perfect forecast removes only 7% of it.
Where it fits: Disruption Diagnostic

Problem 03

Unclear exposure

What it looks like: The risk register scores dozens of threats. Nobody can say which closure would stop shipments first, or for how long.
What is going on: Exposure is time-to-survive versus time-to-recover, not a risk-register score. A site is exposed when its stock and alternatives run out before the disruption ends.
What to do: Screen the main sites and routes on both times. Then price the responses that close the gap.
See it working: Global Disruption Atlas replays real disruptions on live chokepoint data and costs the responses. Network Stress Test finds which single site failure costs most, and what insuring against it costs.
Where it fits: Disruption Diagnostic · Network Flexibility Study

Problem 04

Capital on a single forecast

What it looks like: A new plant, warehouse or second source is approved on one demand forecast. Flexibility appears only as a note in the margin.
What is going on: Flexibility has a price you can calculate. The right to stage, defer, expand or lease has a value that can be set against committing now.
What to do: Value each option against the best fixed design. Then set the triggers for the decisions that can wait.
See it working: Real Options for Network Design compares committing now, the best fixed design, waiting, staged expansion and leasing on the same uncertain demand.
Where it fits: Network Flexibility Study

Problem 05

Decisions at different speeds

What it looks like: Commercial promises change in months, the supply chain in a year, the network in years. Each function is right on its own, and the company still misfires.
What is going on: Clockspeed mismatch is a design problem, not a coordination problem. Decisions that take effect at different speeds have to be planned together.
What to do: Make the three speeds visible in one model. Then pace the fast decisions to what the slow ones can support.
See it working: Clockspeed Balancing lets a leadership team fly a company for 20 years and see what each mismatch costs.
Where it fits: Disruption Diagnostic

The problems I get called in for.