Your POS Knows Which Cocktails to Cut.

Most bars net 10 to 15% in a good year, and plenty run closer to 5. Meanwhile the cocktail list has 30 recipes and a dozen of them actually sell. Every POS can print a product mix report that shows exactly which drinks earn and which just take up prep time. Most operators never pull it.

The framework and what it is worth

Menu engineering was formalized by Kasavana and Smith in 1982 and it still works because guests have not changed: they read a menu for about 90 seconds, look closely at 7 to 9 items, and anchor on what they see first. A 2014 Cornell replication found restaurants running disciplined menu engineering gained about 10% in profitability without adding a single item. A 2026 benchmark of 1,200 independent operators puts the lift at 8 to 14% of contribution margin, annualized, for operators who run the pass quarterly and actually cut the losers.

None of that lift comes from raising prices or shrinking specs. It comes from changing what guests order by changing what the menu features. That is the cheapest margin you will ever buy.

Two numbers per drink

Pull 90 days of sales from the POS, cocktails as their own category. For each drink you need:

• Contribution margin in dollars. Menu price minus ingredient cost. Not pour cost percentage. Dollars.
• Menu mix. The drink's share of cocktail orders in the window. Compare cocktails to cocktails, never to beer or wine, or the volume categories flatten everything else.

The pour cost point matters more than any other part of this. A $16 daiquiri at 20% pour cost puts $12.80 in the drawer. A $22 martini at 30% puts $15.40. The percentage says the martini is the worse drink. The register says it is the better one by $2.60 every time it sells. Operators who plot their quadrants on contribution dollars see an annualized lift of 9 to 14%; operators who use cost percentages see 2 to 4%. Keep pour cost for purchasing conversations. Keep it out of menu decisions.

The four quadrants, bar version

Split the list at the category average on both axes. Every cocktail lands in one of four boxes:

• Stars: high margin, high volume. Leave the spec alone. Give them the best menu placement and never bury them in a redesign.
• Plowhorses: high volume, low margin. The crowd favorites that underpay. Take the price up a dollar, rework the spec cost, or batch the build so at least the labor drops.
• Puzzles: high margin, low volume. Usually a placement or naming problem, not a liquid problem. Rename it, move it up the list, have the team recommend it for two weeks. If it still does not move, it moves to the last box.
• Dogs: low margin, low volume. Cut them. Each one carries prep, dead stock on the back bar, and menu attention that belongs to a star.

What the dead drinks actually cost

A dog is never just one slow drink. It is the two bottles bought for it that pour nothing else, the syrup batched every week and dumped every week, and the 30 seconds a bartender spends re-reading a spec nobody has made since March. Industry inventory data puts beverage shrinkage at 5 to 25% of sales, and complexity is a multiplier: a 200-SKU back bar and a 30-cocktail list are harder to count, harder to train, and easier to steal from. Cutting four dogs typically removes six to ten SKUs and a prep shift's worth of work per month. That saving shows up even before the mix improves.

One honest warning: the analysis is the easy half. The hard half is cutting the founder's favorite Negroni variation that sells four a week. The report gives you the number that wins that argument.

This is the core of a menu refresh engagement: cost the list, run the quadrants, rebuild the menu around what earns. If you want a second set of eyes on your product mix, fifteen minutes is usually enough to see where the margin is hiding.

---

More notes  ·  Back to Notes →