Money & sales
Next move
An evidence-based suggestion for the next thing worth testing, built only from the money-per-band shape on that same completed night. The same evidence always produces the same suggestion — it is a prompt to test, never a prediction.
What it is
Next move is an evidence-based suggestion for the next thing worth testing at a venue, read entirely from a single completed night — the money-per-hour it earned at each level of fullness, how many hours each of those levels was actually measured for, and the Shape that summarises them. It sits on the row it describes and adds no new measurement of its own.
It is a decision layer, not a data source. The same evidence always produces the same suggestion, down to the wording — no model draws on other nights, nothing is forecast, and nothing is tuned per venue. When a valid completed-night comparison exists, that percentage appears as its own line of context; otherwise the suggestion stands on the night's own shape alone.
How to use it at your venue
Operators often read a Next move suggestion the way the revenue-management literature treats a pilot: as one specific, falsifiable idea worth trying for a single comparable stretch — a similar night or day-part — rather than a lever guaranteed to pay.[1] A grounded discipline is to note what the baseline looked like before acting, make only the one change suggested, then compare the next comparable period back against that baseline — the same "establish a baseline, act, monitor the result against a comparable control" loop the hospitality case studies ran.[1] Evidence-based management frames the venue itself as an unfinished prototype: run the small trial, keep what beats the baseline, drop what does not.[2]
The suggestion narrows to the shape of the evidence. A room still paying steeply as it fills is prompted to fill further; one whose earnings have flattened is prompted to sell more to the people already in it; a fall measured across several hours is prompted toward an operational change, while the same fall seen in only one hour is asked only for another reading. This is prescriptive analytics in the formal sense — turning what was observed into one concrete recommended action[3] — built to augment a manager's judgment about the room, never to stand in for it.[4]
How to read it
Read it as a prompt, not a verdict. The suggestion names which of three shapes the night took — still rising at the busiest level the room reached, still rising but the room never got any fuller, or peaked and then fell away below its top — and asks the one question that shape raises. A level of fullness the room never reached on that night contributes nothing and is named as missing rather than predicted, so the prompt only ever speaks to fullness you actually observed. A steeper-sounding suggestion is not a bigger opportunity; it is simply the shape that night happened to take.
What it doesn't mean
⚠ What it can't tell you
It is a suggestion to test, never a prediction, and it never claims a change in revenue would follow. One night cannot separate how full the room was from how late it got, so nothing here establishes cause. It cannot promise revenue, cannot see your costs, staffing plan or bookings, and does not use crowd headcounts or per-guest economics — those stay behind their own proof gates until calibration. It inherits the row's channel treatment exactly: classified takeaway, delivery, drive-through and shipping are out, unclassified orders are in — so where a night carries sales the POS did not say which channel they came from, the reason line says so. It supports your judgment; it does not replace it.
How it's calculated
Nothing is computed fresh here. The suggestion is derived deterministically from figures already on the row — the per-band money map, the count of hours each band was measured, and the Shape reading — plus the stored completed-night comparison when it is present and still valid. It reads no clock and does no aggregation of its own, so it updates only when a new eligible completed night changes the evidence, not because a day has passed. The same inputs produce a byte-identical suggestion every time.
The exact method
Shape is read one of three ways from the money-per-band curve: still rising at the busiest band the room reached; still rising but the top band was never reached; or peaked and fell away below the top band. Each is then narrowed by the evidence itself — a room still paying steeply as it fills is asked to fill further; a room whose curve has flattened is asked to sell more to the guests already in it. A decline is qualified by how long it was observed: a fall measured across several hours is treated as a pattern and asked for an operational change, whereas the same fall measured across a single hour is asked only for another reading, because one hour can be a quiet patch rather than a pattern. The completed-night comparison is consumed only when its verdict is compared, and only then does its percentage appear as a line of context. The unclassified-channel share is named when present but carries no threshold and triggers no branch. Inputs: the already-computed per-band money map (POS + scans), hours-observed per band, the Shape reading, the stored M26 comparison, and the night's unknown-channel net — no headcounts, no per-guest economics, no clock. Version nextMove.v1.
The research behind it
"Next move" is a suggestion, not a prophecy. It's built the same way careful operators and researchers have always improved a business: look at what actually happened (the money curve across the night, the hours you were open, how it compares to a similar completed night), form one specific and testable idea about what might work better, try it for a comparable stretch, and then check the real results against that earlier baseline before deciding whether to keep doing it. It never claims to know the future — it claims to have noticed a pattern worth testing.
-
Restaurant Revenue Management ✓ verified
Read the source ↗ -
Evidence-Based Management ✓ verified
Read the source ↗ -
From Predictive to Prescriptive Analytics ✓ verified
Read the source ↗ -
Decision Support Systems: An Organizational Perspective ✓ verified
Read the source ↗
Confidence & caveats
The logic is deterministic, so "sure" here means reproducible, not proven: given the same night, the suggestion is exactly the same every time — but it is still only a hypothesis about one room on one night. It carries no confidence score because it makes no forecast to be confident about. It is never compared against another venue, it names any fullness the room did not reach rather than filling the gap, and it withholds everything behind the calibration gates — occupancy adequacy, per-guest economics, cost, and staffing. Treat a strong-sounding prompt exactly as you would a weak one: as something to test against your own baseline, not as a promise.