ResilienceGrid: If Much of the Information Is Already Produced, Why Is the Loss Still Happening? Turning Supply Chain Visibility into a Priced Decision with a Named Owner

Ghana’s ports handled 1,701,246 twenty-foot equivalent units in 2024, more than double the 2014 figure. Published evidence reports that imported goods take about fourteen days to clear the border, and tracking and tracing is the weakest of the country’s six logistics components. The usual reading is that the country needs more infrastructure and more reform. This paper reads it differently. Systems in daily use produce many potentially useful signals, although their completeness, accessibility, timeliness and predictive sufficiency have not yet been demonstrated. The search in Appendix D found no publicly documented Ghanaian system that integrates those signals into a decision while there is still time to act. Publicly documented deployed systems largely answer one question, which is where the cargo is. The search did not identify a system that also answers the two that matter next: what this delay will cost the operation waiting for the goods, and who should do what about it now. The paper specifies ResilienceGrid, a decision layer that sits on top of the systems a supply chain already runs rather than replacing them. It predicts a consequential event with measured uncertainty, far enough ahead for a response to be possible. It prices what that prediction means for the operation waiting downstream. It issues a ranked instruction to a named owner with a deadline, sent to whoever can actually carry it out. And it learns, from what was recommended, what was done and what followed. The argument draws on organisational information processing theory, on the supply chain resilience literature, and on the documented record of two decades of customs digitalisation in Ghana. The specification is detailed enough to build, criticise or replicate. It gives eight design requirements, and a five-layer architecture built around a see-predict-simulate-act-learn loop. It also gives the event schema, the rule for ranking responses, the gates a forecasting model must pass before it goes into service, a data feasibility assessment, and the design options considered and rejected. ResilienceGrid is worked through as a proposed twenty-four-month pilot on industrial and mining spare parts, with a transfer case to agricultural perishables and a border clearance feature. Its evaluation protocol, indicators and identification strategy are committed to in advance, including a commitment to publish negative results. The paper is equally clear about what it does not claim. The system has not been built, and no partner has agreed to host it. None of the component technologies is new. Data availability is assessed rather than demonstrated, and the assessment is not yet verified against sample extracts. And software cannot repair the political economy that absorbed two decades of transparency reform at Ghana Customs. Four stated results would show the proposition to be wrong.