The Prediction–Response Gap
Your predictive maintenance model says failure in 14 days. Your procurement process takes 21 days to get the part on site. The model was right. The prediction was accurate. And the failure happened anyway.
This is the blind spot nobody’s measuring: the gap between what your model can see and what your organization can do about it in time.
Actionable Prognostic Horizon
Actionable Prognostic Horizon (APH) is a proposed KPI that measures usable prediction lead time: the model’s raw prediction window minus procurement time, scheduling time, and execution time. What’s left is the time you actually had to prevent the failure.
If the result is negative, the prediction was operationally useless regardless of its statistical accuracy.
The Counterintuitive Insight
A model with a shorter prediction window paired with stocked parts can deliver more operational value than a model with twice the foresight paired with long-lead-time procurement. In the worked example from the full article, a 14-day prediction on an asset with stocked parts yields 10 days of actionable time, while a 21-day prediction on an unstocked asset yields only 3 days. The bottleneck isn’t always the model. It’s often the supply chain or the scheduling process.
Why Accuracy Alone Fails
Most organizations evaluate their predictive maintenance programs on model accuracy alone. A 92% accurate model looks great in a quarterly review. But if 30% of the assets that model monitors have negative APH — meaning the prediction window is shorter than the response time — then 30% of your accurate predictions are operationally useless. APH exposes that gap in a single number.
The Data Already Exists in Maximo
If you’re running Maximo 9 and MAS Predict, every data point in this formula already exists. MAS Predict gives you the failure date prediction and alert timestamps. Your purchasing records track actual vendor delivery times at the item level. Your work order date fields — REPORTDATE, SCHEDSTART, ACTSTART, ACTFINISH — give you the scheduling and execution latency. The Inventory application’s Reorder Details tab tells you which parts are stocked and which aren’t. The data is there. The calculation just hasn’t been built.
Full Deep Dive
The full article covers the complete formula, walks through two worked examples with real numbers, maps every component to specific Maximo fields, and covers the NASA prognostic metrics research this KPI builds on.
Read the complete deep dive on Actionable Prognostic Horizon — formula, two worked examples, Maximo field mapping, and NASA research foundations.
Read Full Article on LinkedInAbout the Advanced KPI Series
This is part of the Advanced KPI Series — 45 proposed metrics for AI-era maintenance that address blind spots legacy KPIs structurally cannot see. These are proposed, not proven — grounded in research and designed for the community to test.