A dashboard that shows revenue dropped 12% last Tuesday has told you something happened. It hasn't told you why, and "why" is the part that actually changes what you do next.
The manual version of this problem
Traditionally, answering "why was Tuesday down" means manually checking: was staffing lower than usual, did a regular menu item run out, was there a weather event, did a competitor open nearby, was a key staff member absent. That's four or five separate systems to check, if a manager remembers to check them at all.
What root-cause analysis actually does
NUA AI cross-references the relevant signals automatically (staffing against the roster, stock movements against the day's sales, guest visit patterns against history) and surfaces the most likely explanation, with the underlying numbers, instead of leaving that correlation work to a person after the fact.
"Tuesday's revenue was down 12% versus forecast. Most likely factor: the walk-in fridge alert at 3pm correlates with the salmon dish being 86'd for the last two hours of service, which typically accounts for 8-10% of evening covers."
Why this matters more than the forecast itself
A forecast tells you what's likely to happen. Root-cause analysis tells you what already happened and why, which is what actually informs whether Wednesday's plan should change.
Feeding forward, not just backward
Root causes identified this way don't just explain the past, they inform future recommendations, so a recurring issue (like that walk-in fridge) gets flagged as a pattern worth fixing structurally, not re-diagnosed from scratch every time.