Skip to content

No account required

Start a project

5 min · 4 sections

Analysis for kitchens, bars and anyone costing a menu.

Most menu engineering fails at the join, because a dish name in the till and a dish name on the recipe card are two different strings.

By Data Analysis App team

Which dishes on my menu actually make money?

Join the till's product mix to your recipe costs, attach modifiers to the dish they belong to, and compare within category rather than across the whole menu. A burger carrying two dollars of additions has a different margin from the base item, and ranking starters against mains produces a matrix that recommends deleting every starter.

Section 01

The join is most of the work

A product mix export knows what sold and a recipe sheet knows what it cost, and nothing connects them except a dish name written two ways, with sizes and specials making it worse.

Every uncertain match should be shown rather than assumed, because a mismatched dish silently attributes one item's cost to another's sales.

Section 02

Modifiers belong to the dish

A till that records modifiers as separate lines will hide margin from every report you run, because the additions carry different cost ratios from the base item.

Attach them before computing anything. Where the till does not track them, report modifier revenue as its own unattributed line rather than spreading it, and say so wherever it affects a figure.

Turning on modifier tracking is usually the single highest-value change available to this data.

Section 03

Engineer each category against itself

Starters and mains live in different worlds of price and popularity. Ranking them together is how a matrix concludes that the entire starter section is a dog.

Dayparts are worth splitting too, because lunch and dinner frequently have different mixes and a dish that works at one can be losing money at the other.

Section 04

What to upload

The POS product mix for a representative period, and your recipe or ingredient costs. Both are read and handed back as downloads, and the report names any relationship it can see between them, but every figure comes from one table, the largest, so the cost is not put against the dish for you.

That leaves the product mix as the file the numbers come from: what sold, in what volume, at what price, by daypart. The join to costs is the step to do yourself, and it is worth doing on a copy of the mix rather than waiting for recipe costs to be current.

See it on a real project

$128,000 of stock had not sold in nine months, and half was bought to reach a supplier discount.

Which stock is tying up cash?

Keep reading

Every check in this guide runs on every upload.

Drop the file in and the problems described here are tested before anything is reported: with the rows behind each repair kept, so you can see exactly what changed.

  • Subtotal rows detected and excluded
  • Duplicates matched on identity, not on whole rows
  • Text-typed numbers found, coerced and counted

No account needed to start. You only pay when you like what you see.