The risk window is the period of time a signal applies to, expressed as the next 1, 3 or 6 hours. A score without a window is not actionable, because HIGH in twenty minutes and HIGH this evening call for completely different responses. Every DwellTime score carries one.
The three windows map onto three different decisions. The next hour is what is about to land on the shift you are running now. The next three hours cover the shift you are about to hand over, which is where staffing, deliveries and driver rotations are still changeable. The next six hours reach the end of the operating day: late arrivals, final rotations, overnight cover.
Why it matters for airport businesses
Airport business decisions are short-horizon decisions. Nobody rosters a cafe six days out on a weather signal. What a manager actually does is stand in the back at 14:00 and decide whether the evening team starts at 16:00 or 17:00, and whether tomorrow's delivery needs changing before the cut-off.
Short windows also keep the product honest. Forecasts lose accuracy with distance, and a signal about dwell time a day ahead would be a guess wearing the clothes of a prediction. Matching the window to the decision keeps the score useful and keeps the claim modest.
How DwellTime uses it
Every score is issued against a window, such as the next one to three hours, so two signals are only comparable when their windows match. As a window moves closer, the level and the confidence value attached to it can both change.
Related terms
See how windows are delivered to a dashboard and an API in how DwellTime works.