[Occasion]

Definition #

The unit of the operation’s growth. An occasion is a discrete, decided event in which a household resolves where a consumable moment gets spent — and in the food domain, where the eating happens — against the full field of alternatives capable of producing that moment.

The occasion is the food-domain instance of the finite resource named in [Share Of Experience]: the household’s consumable moments. [Share Of Stomach] is the measure built on this unit — occasions won over occasions that existed. [Occasion] names the unit itself.

Three properties make it a unit and not a metric. The occasion exists whether or not the operation wins it. Its count does not depend on the operation’s capacity, hours, seats, or attendance. And it resolves in one of three states, not two: won whole, lost whole, or split between two providers.

Mechanism #

A household has roughly twenty-one eating occasions in a week. Every one of them gets decided. The operation is a candidate for some of them, is structurally absent from others, and is unaware of nearly all of them.

The occasion is decided, not scheduled. The decision happens inside a window, and the window is short and late. Most eating occasions are resolved a few hours out by whoever is available in a form that fits the rest of the day. That means an occasion is not won by being open. It is won by being present, findable, and executable inside the window in which the household is actually deciding. An operation can be open for an occasion and never be in the running for it.

The occasion exists independent of the operation. This is the property that makes it the growth unit. Every other number the operation manages to is generated by the operation’s own activity — a cover happens because somebody sat down, a ticket exists because somebody ordered. The occasion happens because a household got hungry. It is the only unit in the operator’s field whose count is set outside the four walls, which is exactly why it is the only unit that can tell the operator whether he is growing or shrinking.

The occasion has three states, and the industry’s instruments hold two. Won and not-won is the assumed binary. The third state is the split occasion — the household takes part of the event from one provider and finishes it somewhere else. A prepared entrée bought at retail and completed in the household’s own kitchen is a split occasion. The operation was a live candidate for it, was partially in the running, and lost half. Half-losses and whole-losses are different conditions requiring different moves, and no capacity-denominated metric can represent either one.

The absence is silent. A lost occasion does not enter the operation’s data as a zero. It is absent from the data entirely, and absence reads as neutral. This is where [Measurement Asymmetry] lives, and it is not a reporting flaw to be corrected with a better system — it is the shape of every ratio whose denominator is the occasions the operation already won.

The recognizable moment. The operator is looking at a sales line that rose and a Guest count that fell, and he is holding both numbers at once without a unit that reconciles them. Per person average up, covers down, and the room concludes the result is mixed. It is not mixed. It is one specific condition — a smaller field position being harvested harder — and the only reason it reads as ambiguous is that the room has no occasion count on the page. Put occasions won on the page first and the ambiguity disappears in a single line.

The refused occasion. Some occasions the operation never contests, by construction. A daypart nobody is scheduled for. A window the operation cannot execute inside. A form the operation does not produce. Those occasions are not lost in competition — they are refused by the architecture, usually without anyone having decided to refuse them. Refused occasions are the cheapest volume in this business to recover, and they are invisible to every instrument that only counts what walked in.

Load-Bearing Distinction #

Not [Share Of Stomach]. [Share Of Stomach] is the measure. [Occasion] is the thing measured. The share is a ratio — occasions won over occasions that existed. The unit is the event itself, with its own properties: a decision window, a field of alternatives, three resolution states. An operator can carry the share concept and still have no working definition of the unit underneath it, which is how the share ends up estimated from sales instead of counted from occasions.

Not [Share Of Experience]. [Share Of Experience] names the domain-crossing competitive frame — every discretionary moment across every domain competing for the household’s finite attention. The occasion is the discrete unit inside that frame, and in the food domain it is the eating occasion. The frame tells the operator who he is competing with. The unit tells him what he is competing for, one event at a time.

Not a cover. A cover is an occasion the operation already won, counted at the table. Covers cannot be lost, because a cover that does not happen was never a cover. The occasion can be lost, which is the entire difference and the entire reason the unit exists.

Not a visit. A visit is the operation’s record of an occasion it won. Visit-frequency work operates on the base that already came. Occasion work operates on the field. The distinction matters most at the exact moment operators reach for frequency programs: raising visits among Guests who already visit is harvest work on a sliver, while the field it sits inside remains uncontested.

Not a transaction or a ticket. Both are artifacts generated by a won occasion. Every metric built on them — per person average, average check, ticket average — has capacity or headcount in the denominator, and not one of them has a Guest in it. They grade the harvest. They cannot see the field.

Not [Point Of Opportunity]. [Point Of Opportunity] names the moment when gathered intelligence and the capacity to act converge, on the operator and cast side, inside an interaction already underway. The occasion is upstream of that: it is the event the household is deciding about before the operation has any contact at all. One is a moment inside a won occasion. The other is the occasion.

Not [Monetization Window]. The monetization window is a temporal ceiling the operator imposes on the Guest relationship. The occasion’s decision window is not imposed by the operator at all — it is the household’s, it is short, and the operator either is or is not present inside it. One is a frame the operator should remove. The other is a fact the operator has to build against.

The failure mode this term guards against is the one that produces no visible error at any step. Without the occasion named as a unit, growth decisions get denominated in a performance metric, the operation gets progressively better at harvesting a field it is progressively losing, and every instrument it owns confirms that things are improving.

Diagnostic Tests #

Test One — The Count Test. Ask the operator for occasions won by period for the last twelve periods. Not sales, not covers-as-a-proxy-for-sales, not an index. A count. If the number does not exist as a standing line on a standing report, the operation has no growth instrument, whatever else is on the page.

Test Two — The Order Test. Look at where occasions won sits on the period report relative to dollars per occasion. If dollars come first, or if occasions are absent, the room has been holding growth conversations off a performance metric. The order on the page is not cosmetic — it sets which number the room treats as the verdict.

Test Three — The Field Test. For one daypart, ask the operator to name every channel that could have taken the occasion. Specific, not categorical: the retail prepared-food case at the store two miles out, the breakfast program on the commute pattern, the container already in the household’s refrigerator. An operator who can name three is running the unit. An operator who names only restaurants is competing in a comp set that is losing the same occasions he is.

Test Four — The Window Test. For each daypart, ask when the household decides and whether the operation is present inside that window in a form that fits. Present means findable, reachable, and executable — not open. Most operations discover here that they are absent from decision windows they believed they were competing in.

Test Five — The Split Test. Of the occasions that left, how many left whole and how many were split. The operator will have to estimate. Have him estimate, and have him write down what the estimate rests on. An estimate with its basis named is the first working version of an instrument the operation does not currently own.

Test Six — The Denominator Audit. List every metric the managers are measured on and write the denominator beside each one. Count how many have a Guest in them rather than a seat, an hour, a square foot, or a headcount. That count is the read.

Test Seven — The Refusal Test. Name the occasions the operation does not contest at all — dayparts unstaffed, windows it cannot execute inside, forms it does not produce. Then ask, for each one, whether that refusal was decided or inherited. Inherited refusals are the cheapest recoverable volume in the operation.

Test Eight — The Attribution Test. For every period where the occasion count fell, ask for a named hypothesis: a channel, a variable, and a daypart. Not the macro. Not the consumer. A wrong named hypothesis is a working instrument. An unnamed loss is an operation that has stopped reading.

Family Position #

Unit-level IP. Sits inside Perspective — Operating Units. The food-domain instance of the finite resource named in [Share Of Experience], and the denominator [Share Of Stomach] is built on. Cross-Fundamental: the unit reorders decisions in all five.

Perspective application. The unit sets the competitor set. A denominator of capacity produces a competitor set made of other restaurants. A denominator of occasions produces a competitor set that includes a retail prepared-food case, a convenience-store breakfast program, and the household’s own kitchen. Same operator, same market, two different fields — and the difference was decided by the unit, not by the analysis run on top of it.

Product application. The Product is being chosen against alternatives at the occasion level, including alternatives that take only part of the event. A GX designed for a whole occasion competes badly for a split one, and the operation will never learn that from its own data. The unit forces the question of what form the Product takes inside the occasion the household is actually resolving.

People application. The cast is staffed to won occasions. Scheduling built off historical covers reproduces the operation’s existing field position by construction — an occasion at a daypart nobody is scheduled for is refused before it is contested. The unit turns the schedule from a cost decision into a field decision.

Performance application. Execution metrics grade how well the operation served whoever arrived. They are structurally silent on who decided not to. The unit does not replace those metrics — it puts a ceiling above them, so that flawless execution against a shrinking field reads as what it is instead of as a win.

Profit application. Margin per won occasion can rise while total occasions fall, and the P&L will report the first half of that sentence without the second. This is where [Lost Opportunity Tax] accrues without ever appearing as a line item. The unit gives the operator the count that turns an invisible tax into a number.

Fundamentals Coverage

Perspective read. The occasion originates in the household, not in the operation, and that is the whole reason it can carry the operator’s read. The industry’s default posture is to reason about growth from instruments generated inside the four walls, which means the operator’s field of view is bounded by his own activity — he can see everything he did and nothing that was decided about him. The tell is an operator who describes a soft period in terms of consumer behavior in the aggregate and cannot name a single specific alternative that took a specific occasion at a specific daypart. The response is to install the count and the field list as standing artifacts, so the read starts outside the building and works inward instead of the reverse.

Product read. The occasion is the frame in which the Product is actually judged, and the frame is more demanding than the plate. A household resolving a weeknight event is weighing effort, timing, form, and what happens to the rest of the evening, and the operation’s Product either fits that shape or does not. The expression of a Product that has never been read against the occasion is a menu that performs well among people who chose it and loses steadily to alternatives that were never tasted. Detection is the split occasion: when a household takes part of the event elsewhere and finishes the rest itself, that is direct evidence about the form of the Product, delivered in a channel the operation does not monitor. The response is to design the Product for the occasion as the household defines it, including the occasions the operation is currently only half-competing for.

People read. The cast produces hospitality inside occasions the operation won, which means the People architecture is built entirely on the operation’s existing field position and reinforces it. Origination is the schedule: labor gets built off historical demand, so the operation staffs where it already wins and is structurally absent where it does not, and the absence then confirms itself in the next period’s history. The expression is a cast that executes well and a set of dayparts nobody has ever contested. Detection is the refusal list — name the occasions the operation does not contest, then ask whether each refusal was decided or inherited. The response is to treat the schedule as a field instrument, not only a cost instrument, and to make one refused daypart contested on purpose so the operation learns what it costs to compete there.

Performance read. Performance is where the unit does its most uncomfortable work, because performance can be excellent while the field is being lost, and nothing inside the performance instrument set can report that. Origination is the denominator: throughput, turns, RevPASH, and sales per labor hour all measure the operation against its own capacity, so a smaller operation running perfectly outscores a larger one running roughly. The expression is a management team hitting every target in a contracting operation and reasonably concluding it is doing its job, because it is. Detection is the count sitting above the execution metrics on the same page, so that execution is read as performance against a field position rather than as the verdict on the operation. The response is to keep the execution metrics exactly as they are and stop letting them answer a question they were never built to answer.

Profit read. Profit is where the arithmetic of the unit is most easily mistaken for good news. Dollars per won occasion rising while occasions fall produces a margin story with no volume in it, and margin stories are the most persuasive kind because they read as discipline. Origination is the ratio: any per-occasion money metric improves when the base narrows toward its most committed members, which means contraction and quality-of-mix improvement are arithmetically indistinguishable on the P&L. The expression is an operation with better margins, a better mix, and less business than it had a year ago, defending each of the first two while the third goes unnamed. Detection is the headroom arithmetic run explicitly: at two of a household’s twenty-one occasions, an eight percent squeeze on per-occasion spend operates on two occasions, while winning a third is fifty percent more volume — and the asymmetry gets more extreme the less share the operation holds. The response is to rank capital and attention against the larger lever, which for almost every operation in this industry is the field, not the harvest.

Cross-References To Locked IP #

Parent:

• [Share Of Experience] — the domain-crossing competitive frame whose finite resource this unit is the food-domain instance of

Related:

• [Share Of Stomach] — the food-domain measure built on this unit; occasions won over occasions that existed

• [Measurement Asymmetry] — why a lost occasion is absent from the data rather than a zero in it

• [The Dashboard Trap] — the condition that results when every instrument in the room is denominated in capacity

• [The Read] — the aggregate discipline the occasion count feeds

• [Lost Opportunity Tax] — what accrues on refused and lost occasions with no line item to hold it

• [Point Of Opportunity] — the moment of convergence inside an occasion the operation already won

• [Guest] — the party doing the deciding, one occasion at a time

• [By Design Or By Default] — the choice the operator makes about which unit his growth decisions are denominated in

• [Relational Compounding] — what accumulates when occasions are won repeatedly rather than harvested once

Opposing patterns:

• [The Transactional Substitution Kit] — the five levers that manufacture volume without contesting the field

• [Transactional Affordability Lie] — the conclusion that price lost the occasion, drawn without ever reading the variable in contention

• [No Static Achievement] — the physics that makes a held field position impossible, and a coasted one a decline

• [Monetization Window] — the operator-imposed temporal ceiling the occasion’s real decision window is routinely confused with

Why This Matters #

I have been saying forever and a day that this industry does not have a demand problem it can price its way out of. It has a unit problem. Operators are making growth decisions with a number that only knows about attendance it already got, and there is no amount of rigor applied to that number that can fix what the number cannot see.

This is why the error is so durable. Every step in the reasoning is sound. The data is accurate, the analysis is competent, the conclusions follow. The only defect is upstream of all of it, in the choice of unit, and a defect in the unit is the one failure a rigorous process cannot detect in itself. That is what makes this term load-bearing rather than merely useful: it operates at the layer where the operator cannot self-correct.

It also names something the industry has recently made unavoidable. The field has developed a form of partial capture — occasions split between a retail provider and the household’s own kitchen — that the industry’s instruments cannot represent at all. When the unit is the occasion, a half-occasion is at least a visible and nameable condition. When the unit is dollars per attendance, it does not exist, and an operator can lose half of an event he was a live candidate for and never register that he was in the running.

The unit is where my framework’s whole competitive architecture lands in the operator’s hands. [Share Of Experience] tells him the field is bigger than his category. [Share Of Stomach] gives him the food-domain ratio. [Occasion] gives him the countable thing, which is the only one of the three he can put on a page tomorrow morning and manage against. A frame the operator cannot count is a frame he will nod at and never run.

Operating Consequence #

Put occasions first on the page. Every period report leads with occasions won and follows with dollars per occasion. The order is the consequence. It settles which number the room treats as growth and which one it treats as the grade, before anyone opens their mouth.

Stop reading sales and traffic as two lines. The operator reads them as one condition. Sales up on occasions down is named out loud as a smaller field position harvested harder, every time it occurs, in those terms. The naming is the discipline — the moment it gets softened to a mixed result, the deferral begins again.

Split check movement into price and quantity. Quantity growth on flat or falling occasion counts is occasion consolidation, which is field loss wearing a performance costume. The split gets reported, not the total, and it gets reported to the cast in the same terms it gets reported to the operator.

Build the field list per daypart and keep it current. For each daypart, the specific channels that could take the occasion and the variable each one wins on. Named, written, revised. An unnamed competitor is an uncontested competitor.

Refuse macro attributions. No occasion decline is ever attributed to the consumer, the economy, or the category. Every decline gets a named hypothesis — channel, variable, daypart. The operator would rather be specifically wrong than generally comfortable.

Read the decision window before any marketing decision. Presence means findable, reachable, and executable inside the window the household is actually deciding in. Open is not present. Any promotion that does not land inside the window is spend against an occasion already resolved.

Treat the schedule as a field instrument. Labor built off historical covers reproduces the operation’s current field position. At least one refused daypart gets contested on purpose per quarter, so the operation learns what competing there actually costs instead of assuming.

Rank the harvest lever second. Per-occasion improvements stay in the operation and keep getting run — they are simply no longer the growth plan. The larger lever is the field, and for an operation holding a minority of it, the arithmetic is not close.

What Changes Tomorrow #

Take the operation’s weakest daypart. Build one page with two lines on it — occasions won for the last twelve periods, and dollars per occasion for the last twelve periods, in that order — and put no other figure anywhere on the page. Then bring in the person who runs that daypart and ask one question: where did those occasions go.

The answer will be incomplete. It will also be the first time that conversation has happened in the building with the dollars removed from the room, and the dollars are what have been letting everyone agree that things are fine. Write down whatever names come back — a store, a chain, a commute pattern, a household’s own refrigerator — and mark which of the occasions the room believes were lost whole and which were split.

The indicator to read is not the count itself in the first period. It is whether the room can name a channel, a variable, and a window for the occasions that left. If it can, the operation has a working field read and the next move is to contest one named window on purpose. If it cannot, the operation has been running growth off a performance metric, and the corrective is not a new tactic — it is the page, run every period until the names start arriving.

The dollar is what the operation collects. The occasion is what it wins. Choose the denominator, or inherit one and let it choose the strategy.