After tuning transmit power and enabling band steering in a hospital's imaging suite corridor, the network team pulls a new coverage and roaming report and lays it side by side with the pre-change baseline. Roaming time between APs has dropped and 5 GHz client association has increased, but overall coverage area at the weakest points is roughly unchanged. What should the team conclude from this comparison?

Comparing against a baseline only works if the comparison is scoped to what the change was actually meant to affect. Transmit power tuning and band steering target client behavior and RF balance — roaming speed and band association — not the physical coverage footprint at the weakest measurement point, so an unchanged coverage number there doesn't indicate failure. The 'complete failure' framing conflates a metric outside the change's scope with the ones the change was designed to move, and both of those (roaming time, 5 GHz association) improved. Calling the baseline invalid because metrics moved in different directions misunderstands what a baseline is for: it's normal and expected for a tuning change to improve some measurements while leaving unrelated ones flat, and that pattern doesn't undermine the original measurement's validity. Reverting on that basis would discard a genuine improvement over a non-issue. The operationally sound move is to document that the coverage gap is a separate problem, possibly requiring an AP placement change rather than an RF-parameter change, and track it separately. A concrete check: confirm the coverage weak spot was already flagged as a known gap in the original baseline notes rather than a new regression, then open a distinct optimization item for it instead of undoing work that measurably helped.