Kasavana and Smith published their menu engineering framework at the Cornell School of Hotel Administration in 1982. The methodology is unchanged in its fundamentals — plot every item by contribution margin and popularity, classify the four quadrants, and act. What has changed is the interval at which the framework is useful.
In 1982, sales mix data took weeks to assemble. Competitor pricing was collected in person, on foot. A comprehensive menu engineering audit was a major internal project, and the resulting document was treated as a bound deliverable — something you commissioned once a year, presented at an ownership meeting, and returned to fourteen months later.
What has changed
Three things have shifted the interval at which the analysis is useful:
Sales mix is live. Every modern POS records item-level data in real time. The input the 1982 framework required in a two-week gathering phase now arrives as you sit down to read this. The analysis no longer has to wait for the data.
Comp sets move faster than menus. In a well-functioning dining market, your competitors adjust prices, rotate dishes, and respond to seasonal inputs continuously. An annual audit captures a snapshot of a market that was already a year old by the time you acted on it.
Guest behaviour compounds. Pricing decisions made at lunch today shape the habit patterns of guests returning in three weeks. A Plowhorse that should have been repriced in April is compounding a margin leak every week it waits for the next annual review.
Why the weekly cadence is the right one
The menu engineering framework has five inputs: contribution margin, order popularity, category balance, price anchoring, and guest sentiment. Four of those are refreshed in your POS within a day. The fifth — guest sentiment — moves in a review-cycle rhythm that correlates roughly with a week.
A weekly cadence means no more than seven days of drift accumulates before a Puzzle becomes a repositioning project, before a Plowhorse becomes a margin complaint, before a Dog becomes a procurement problem. It is the tightest interval that still leaves room for execution — which is the point.
What weekly does not mean
It does not mean weekly menu reprints. The menu is a physical artifact that your guest perceives as stable; rebuilding it every week is neither necessary nor desirable. Weekly means weekly analysis, weekly prioritisation, weekly decisions about what should or should not change — followed by a menu update when the accumulated decisions justify one.
The goal is not velocity. The goal is the elimination of drift. Weekly cadence makes the framework diagnostic again, not archaeological.