Menu Analytics: How Restaurants Turn Sales Data Into Profit

Menu Analytics: How Restaurants Turn Sales Data Into Profit

Menu Analytics: How Restaurants Turn Sales Data Into Profit

Restaurant menu analytics explained: which numbers to track, how to sort items into stars and dogs, and how to act on the results without guessing.

Restaurant manager reviewing menu performance over lunch.
Restaurant manager reviewing menu performance over lunch.

Menu analytics means using your own sales data to see which dishes earn their place: what sells, what makes money and what only fills space. The method is simple. Pull a few months of point-of-sale (POS) data, work out each item’s contribution margin and share of sales, sort items into four groups, then change prices, placement or recipes and measure again. It works best when the analysis runs inside the system that takes your orders, so the numbers are already there. That is how Shire’s POS and menu analytics are set up.

TL;DR

  • Menu analytics compares every item on two things: how often it sells (popularity) and how many dollars it leaves you (profitability).

  • Use at least three months of POS sales data, and keep meal periods apart so lunch does not distort dinner.

  • Contribution margin (price minus food cost) tells you more than food cost percentage, because it counts real dollars.

  • Sort items into Stars, Plowhorses, Puzzles and Dogs, then give each group one clear action.

  • Change a few things at a time, wait 30 to 60 days and re-measure check size, item mix and margin per guest.

Table of Contents

  • What menu analytics actually tells you

  • The numbers you need before you start

  • Sorting items into four groups

  • Where the matrix misleads you

  • Beyond the matrix: other data worth reading

  • Turning the analysis into changes

  • Checking whether it worked

  • Why the system that takes the order should run the analysis

  • Shire for your restaurant

  • A simple quarterly routine

  • FAQ

  • Sources

What menu analytics actually tells you

Most owners know their best sellers by feel. Menu analytics checks that feel against the register. A burger that sells 400 times a month may look like the backbone of the menu while leaving you a thinner margin than a pasta that sells 150 times. Without the numbers, you promote the wrong dish.

The analysis answers four practical questions:

  1. Which items do guests order most?

  2. Which items leave the most money after ingredient cost?

  3. Which items are popular but quietly losing money?

  4. Which items take up menu space and kitchen prep without paying for it?

You may also hear the term menu engineering. It is the same idea applied as a method: score each item, then redesign pricing, wording and layout around the result. Analytics is the measuring half. Engineering is what you do with the measurements.

The numbers you need before you start

You need three inputs for every item: selling price, ingredient (plate) cost and units sold in the period. Everything else is arithmetic.

Contribution margin

Contribution margin is the selling price minus the food cost of the dish. A dish that sells for $18 and costs $5 in ingredients leaves $13. Black Pearl Investments explains that this figure matters more than food cost percentage because it measures real dollars rather than a ratio.

The difference shows up fast. A $6 side with a 20% food cost leaves $4.80. A $28 steak with a 38% food cost leaves $17.36. The steak looks worse on a percentage report and is far better in the till. You cannot pay rent with a percentage.

Menu mix percentage

Menu mix is the share of all items sold that a single dish accounts for: the item’s units divided by total units sold, times 100. If you sold 5,000 items in the period and 400 were the burger, its menu mix is 8%. Divide 100% by the number of items on the menu to find the “fair share” line. Many operators use a lower bar, commonly 70% of that fair share, so a large menu does not push every item into the unpopular column. Pick one rule and keep it for every analysis, so results stay comparable from quarter to quarter.

Menu mix formula and fair share threshold

How much data is enough

A single busy weekend misleads. The National Restaurant Association’s operations textbook advises gathering at least three months of POS sales data before analyzing performance, because longer periods give more useful results, and compares each item’s sales against the entire menu over the same period. Practical rules that follow:

  • Use the same date range for every item.

  • Analyze meal periods separately. Brunch and dinner are different menus in practice.

  • Analyze bar and kitchen separately. A $14 cocktail and a $14 entree do not compete.

  • Leave out items that were on the menu for only part of the period, or flag them.

  • If a holiday or a closure fell in the range, note it before drawing conclusions.

Sorting items into four groups

Plot every item against two lines: the average popularity and the average contribution margin. Each item lands in one of four boxes.

Group

Popularity

Contribution margin

What it means

Star

High

High

Your best dishes. Protect and feature them.

Plow­horse

High

Low

Guests love it, you earn little. Fix the margin.

Puzzle

Low

High

Profitable but overlooked. Make it easier to find.

Dog

Low

Low

Neither sells nor earns. Rework or remove.

What to do with each group

Stars. Do not touch the recipe, and be careful with the price. Keep them in the most visible spots on the menu, train servers to mention them and make sure the kitchen executes them the same way every time. A small price rise on a star is often the cheapest margin gain on the menu, since guests choose it for reasons other than price.

Plowhorses. These bring traffic, so removing them is risky. Look first at cost: portion size, a pricey garnish, a supplier price that has crept up. Then consider a modest price increase or pairing the item with a high-margin side or drink. Bundling can lift the check without a visible price jump.

Puzzles. The dish earns well but guests are not choosing it. The usual causes are placement, a flat description, a confusing name or servers who never mention it. Move it to a spot where eyes land, rewrite the description around what the guest tastes, and give servers a line to say. Give it a fair trial before you decide it is a dog.

Dogs. These cost you prep time, inventory and menu space. Options are to remove them, reprice them, rework the recipe or keep them deliberately for a reason that has nothing to do with profit, such as a dietary option that keeps a table of four together. If you keep one, know why.

What to do with each group

A worked example

Say a casual dining room sells 4,000 entrees over three months across eight dishes. Fair share is 12.5% per dish (100 divided by 8). Using the 70% rule, the popularity bar is about 8.75%, or 350 units. The average contribution margin across the menu works out to $14.

Dish

Units

Price

Food cost

Margin

Group

Ribeye

520

$34

$13

$21

Star

Cheeseburger

900

$16

$6

$10

Plow­horse

Salmon

300

$27

$9

$18

Puzzle

Mushroom risotto

260

$19

$9

$10

Dog

The burger makes up the largest share of orders but leaves $4 less than the menu average per plate. Moving the price from $16 to $17 across 900 sales is $900 a quarter, a small change guests rarely notice. The salmon earns $18 a plate and sells only 300. If better placement and a server suggestion lift it to 450, that adds $2,700 in margin over the same period. The risotto is the candidate for a recipe rework or removal.

The figures are illustrative, not a benchmark. Your own numbers will differ, and that is the point.

Where the matrix misleads you

The four-box view treats every dish as independent. Guests do not. Black Pearl Investments notes that customers substitute between closely priced items, so a change to one dish can move demand to another. If you raise the burger and the chicken sandwich picks up the orders, the total result depends on the margin of both.

Three habits keep you out of trouble:

  • Change a few things at a time. If you reprice six items and rewrite ten descriptions, you cannot tell which change did what.

  • Judge by the whole check. Look at the average check and the mix of items after the change, not just the one dish.

  • Respect context. An appetizer that few people order alone may be what starts a profitable table. A drink that sells poorly may be pairing with the star entree.

Beyond the matrix: other data worth reading

Sales and cost give you the core. Other signals sharpen the picture:

  • Voids and comps. An item that is often voided or comped is costing you more than its margin shows. It may be a kitchen quality issue or a menu description that promises too much.

  • Prep and ticket time. A dish with a good margin that holds up the line slows every table. Slow tickets cost you turns, and turns are revenue.

  • Modifiers. If most guests ask for the same change, that change is part of the real recipe and part of the real cost.

  • Day and daypart patterns. A dish that is a dog on Tuesday and a star on Saturday may belong on a weekend menu.

  • Server-level selling. If one server sells twice as many puzzles as the rest, copy what that server says.

The last two are where menu data meets the floor. A menu only performs as well as the dining room sells it and the kitchen delivers it.

Turning the analysis into changes

Work through the results in this order.

  1. Fix cost errors first. Update plate costs for current supplier prices. An outdated cost makes a star look better than it is.

  2. Reprice the safe items. Plowhorses and stars with room to move take small, specific increases.

  3. Reposition puzzles. Change placement, name and description before price.

  4. Decide on the dogs. Rework, bundle or cut.

  5. Brief the team. Servers need to know which items are featured and a one-line reason a guest would like each. A change on paper that servers do not know about does nothing.

  6. Reprint or update once. Roll out all the planned changes together, so the measurement period is clean.

Checking whether it worked

After the update, hold the menu steady. Give it 30 to 60 days, then repeat the same analysis on the new period. Compare:

  • Average check size

  • Item sales mix, especially the share going to stars and puzzles

  • Contribution margin per guest (total margin divided by covers)

  • Food cost, to confirm that no supplier change is hiding the effect

Set a target before you start, such as moving the average margin per cover up by a set amount, so you know what success looks like. Then put the next review on the calendar. Many operators tie it to seasonal menu changes, which makes it a quarterly routine instead of a one-off project.

If you want a before-and-after you can trust, check at the level of the individual check. Look at what guests ordered together before and after, not just what one dish sold. That is the only way to see substitution at work.

Why the system that takes the order should run the analysis

Most restaurants start with a spreadsheet: export the sales, type in costs, build formulas. It works the first time. By the third quarter, the export is late, the costs are stale and nobody owns the file. The analysis only helps if it stays current.

That is why menu analytics fits best inside the operating system. Shire is an AI-powered restaurant operating system that includes menu analytics alongside POS integration, kitchen display sync, an owner reporting app and floor management that uses your existing CCTV cameras. Sales, tickets and table activity sit in one place, so you read item performance next to the operation that produced it instead of reconciling separate exports. Shire adapts to your workflow without replacing your existing POS or camera systems, and it gives each staff role live insight into the floor.

That matters for the signals above. Slow tickets, long table times and the pace of service are part of why a dish earns what it earns, and they are the numbers a menu spreadsheet never sees.

Shire for your restaurant

Shire Intelligence starts at $40 per month, and Shire POS and Shire Enterprise are available for operators who need a full POS suite or multi-location infrastructure. You can see the full line on the Shire Intelligence page.

Shireintelligence

For a sit-down room, the full-service restaurant software page shows how floor intelligence and staff coordination work together. If speed is your business, see the quick-service restaurant software. The Shire POS page covers orders at the item level, and the Shire Owner App puts reporting in your pocket, so you can review item performance without sitting at the back-office computer.

A simple quarterly routine

You can run this without any ceremony. Block two hours once a quarter:

  1. Export or open the last three months of item sales by meal period.

  2. Refresh plate costs for your top 20 items by sales.

  3. Calculate contribution margin and menu mix for each item.

  4. Sort into the four groups and write one action next to each.

  5. Pick no more than five changes and assign an owner to each.

  6. Brief servers before the new menu goes live.

  7. Diarize the re-check for 30 to 60 days out.

Owners who do this four times a year build a menu that gets sharper with each pass, because every cycle starts from a cleaner baseline than the one before.

FAQ

What is the difference between menu analytics and menu engineering?

Menu analytics is the measurement: sales, cost, margin and popularity for each item. Menu engineering is the action you take from it, such as repricing, repositioning or removing dishes. In practice you do both in one cycle, and the analytics part is what makes the changes defensible.

How much sales data do I need?

Use at least three months of POS data, as the National Restaurant Association’s operations textbook advises, and keep meal periods and bar versus kitchen separate. A new restaurant with less history can start with what it has and repeat the analysis as the data grows.

Should I judge dishes by food cost percentage or contribution margin?

Contribution margin is the better guide for menu decisions because it shows the dollars each dish leaves you. Food cost percentage still helps you monitor purchasing and portion control across the whole menu, so track both and rank dishes by margin.

How often should I review my menu?

A quarterly review suits most restaurants, ideally timed with seasonal menu changes. After any change, wait 30 to 60 days before re-analyzing so the results reflect the new menu and not the transition.

Can a small restaurant use menu analytics?

Yes. A single location with a modest menu has fewer items to score, so the first pass takes less time. The same steps apply: margin, mix, four groups, a few changes, a re-check. Shire Intelligence is priced from $40 per month, with tiers for smaller operations.

Sources

Recommended

Written with BabyLoveGrowth tools

Menu analytics means using your own sales data to see which dishes earn their place: what sells, what makes money and what only fills space. The method is simple. Pull a few months of point-of-sale (POS) data, work out each item’s contribution margin and share of sales, sort items into four groups, then change prices, placement or recipes and measure again. It works best when the analysis runs inside the system that takes your orders, so the numbers are already there. That is how Shire’s POS and menu analytics are set up.

TL;DR

  • Menu analytics compares every item on two things: how often it sells (popularity) and how many dollars it leaves you (profitability).

  • Use at least three months of POS sales data, and keep meal periods apart so lunch does not distort dinner.

  • Contribution margin (price minus food cost) tells you more than food cost percentage, because it counts real dollars.

  • Sort items into Stars, Plowhorses, Puzzles and Dogs, then give each group one clear action.

  • Change a few things at a time, wait 30 to 60 days and re-measure check size, item mix and margin per guest.

Table of Contents

  • What menu analytics actually tells you

  • The numbers you need before you start

  • Sorting items into four groups

  • Where the matrix misleads you

  • Beyond the matrix: other data worth reading

  • Turning the analysis into changes

  • Checking whether it worked

  • Why the system that takes the order should run the analysis

  • Shire for your restaurant

  • A simple quarterly routine

  • FAQ

  • Sources

What menu analytics actually tells you

Most owners know their best sellers by feel. Menu analytics checks that feel against the register. A burger that sells 400 times a month may look like the backbone of the menu while leaving you a thinner margin than a pasta that sells 150 times. Without the numbers, you promote the wrong dish.

The analysis answers four practical questions:

  1. Which items do guests order most?

  2. Which items leave the most money after ingredient cost?

  3. Which items are popular but quietly losing money?

  4. Which items take up menu space and kitchen prep without paying for it?

You may also hear the term menu engineering. It is the same idea applied as a method: score each item, then redesign pricing, wording and layout around the result. Analytics is the measuring half. Engineering is what you do with the measurements.

The numbers you need before you start

You need three inputs for every item: selling price, ingredient (plate) cost and units sold in the period. Everything else is arithmetic.

Contribution margin

Contribution margin is the selling price minus the food cost of the dish. A dish that sells for $18 and costs $5 in ingredients leaves $13. Black Pearl Investments explains that this figure matters more than food cost percentage because it measures real dollars rather than a ratio.

The difference shows up fast. A $6 side with a 20% food cost leaves $4.80. A $28 steak with a 38% food cost leaves $17.36. The steak looks worse on a percentage report and is far better in the till. You cannot pay rent with a percentage.

Menu mix percentage

Menu mix is the share of all items sold that a single dish accounts for: the item’s units divided by total units sold, times 100. If you sold 5,000 items in the period and 400 were the burger, its menu mix is 8%. Divide 100% by the number of items on the menu to find the “fair share” line. Many operators use a lower bar, commonly 70% of that fair share, so a large menu does not push every item into the unpopular column. Pick one rule and keep it for every analysis, so results stay comparable from quarter to quarter.

Menu mix formula and fair share threshold

How much data is enough

A single busy weekend misleads. The National Restaurant Association’s operations textbook advises gathering at least three months of POS sales data before analyzing performance, because longer periods give more useful results, and compares each item’s sales against the entire menu over the same period. Practical rules that follow:

  • Use the same date range for every item.

  • Analyze meal periods separately. Brunch and dinner are different menus in practice.

  • Analyze bar and kitchen separately. A $14 cocktail and a $14 entree do not compete.

  • Leave out items that were on the menu for only part of the period, or flag them.

  • If a holiday or a closure fell in the range, note it before drawing conclusions.

Sorting items into four groups

Plot every item against two lines: the average popularity and the average contribution margin. Each item lands in one of four boxes.

Star

  • Popularity: High

  • Contribution margin: High

  • What it means: Your best dishes. Protect and feature them.

Plowhorse

  • Popularity: High

  • Contribution margin: Low

  • What it means: Guests love it, you earn little. Fix the margin.

Puzzle

  • Popularity: Low

  • Contribution margin: High

  • What it means: Profitable but overlooked. Make it easier to find.

Dog

  • Popularity: Low

  • Contribution margin: Low

  • What it means: Neither sells nor earns. Rework or remove.

What to do with each group

Stars. Do not touch the recipe, and be careful with the price. Keep them in the most visible spots on the menu, train servers to mention them and make sure the kitchen executes them the same way every time. A small price rise on a star is often the cheapest margin gain on the menu, since guests choose it for reasons other than price.

Plowhorses. These bring traffic, so removing them is risky. Look first at cost: portion size, a pricey garnish, a supplier price that has crept up. Then consider a modest price increase or pairing the item with a high-margin side or drink. Bundling can lift the check without a visible price jump.

Puzzles. The dish earns well but guests are not choosing it. The usual causes are placement, a flat description, a confusing name or servers who never mention it. Move it to a spot where eyes land, rewrite the description around what the guest tastes, and give servers a line to say. Give it a fair trial before you decide it is a dog.

Dogs. These cost you prep time, inventory and menu space. Options are to remove them, reprice them, rework the recipe or keep them deliberately for a reason that has nothing to do with profit, such as a dietary option that keeps a table of four together. If you keep one, know why.

What to do with each group

A worked example

Say a casual dining room sells 4,000 entrees over three months across eight dishes. Fair share is 12.5% per dish (100 divided by 8). Using the 70% rule, the popularity bar is about 8.75%, or 350 units. The average contribution margin across the menu works out to $14.

Ribeye

  • Units: 520

  • Price: $34

  • Food cost: $13

  • Margin: $21

  • Group: Star

Cheeseburger

  • Units: 900

  • Price: $16

  • Food cost: $6

  • Margin: $10

  • Group: Plowhorse

Salmon

  • Units: 300

  • Price: $27

  • Food cost: $9

  • Margin: $18

  • Group: Puzzle

Mushroom risotto

  • Units: 260

  • Price: $19

  • Food cost: $9

  • Margin: $10

  • Group: Dog

The burger makes up the largest share of orders but leaves $4 less than the menu average per plate. Moving the price from $16 to $17 across 900 sales is $900 a quarter, a small change guests rarely notice. The salmon earns $18 a plate and sells only 300. If better placement and a server suggestion lift it to 450, that adds $2,700 in margin over the same period. The risotto is the candidate for a recipe rework or removal.

The figures are illustrative, not a benchmark. Your own numbers will differ, and that is the point.

Where the matrix misleads you

The four-box view treats every dish as independent. Guests do not. Black Pearl Investments notes that customers substitute between closely priced items, so a change to one dish can move demand to another. If you raise the burger and the chicken sandwich picks up the orders, the total result depends on the margin of both.

Three habits keep you out of trouble:

  • Change a few things at a time. If you reprice six items and rewrite ten descriptions, you cannot tell which change did what.

  • Judge by the whole check. Look at the average check and the mix of items after the change, not just the one dish.

  • Respect context. An appetizer that few people order alone may be what starts a profitable table. A drink that sells poorly may be pairing with the star entree.

Beyond the matrix: other data worth reading

Sales and cost give you the core. Other signals sharpen the picture:

  • Voids and comps. An item that is often voided or comped is costing you more than its margin shows. It may be a kitchen quality issue or a menu description that promises too much.

  • Prep and ticket time. A dish with a good margin that holds up the line slows every table. Slow tickets cost you turns, and turns are revenue.

  • Modifiers. If most guests ask for the same change, that change is part of the real recipe and part of the real cost.

  • Day and daypart patterns. A dish that is a dog on Tuesday and a star on Saturday may belong on a weekend menu.

  • Server-level selling. If one server sells twice as many puzzles as the rest, copy what that server says.

The last two are where menu data meets the floor. A menu only performs as well as the dining room sells it and the kitchen delivers it.

Turning the analysis into changes

Work through the results in this order.

  1. Fix cost errors first. Update plate costs for current supplier prices. An outdated cost makes a star look better than it is.

  2. Reprice the safe items. Plowhorses and stars with room to move take small, specific increases.

  3. Reposition puzzles. Change placement, name and description before price.

  4. Decide on the dogs. Rework, bundle or cut.

  5. Brief the team. Servers need to know which items are featured and a one-line reason a guest would like each. A change on paper that servers do not know about does nothing.

  6. Reprint or update once. Roll out all the planned changes together, so the measurement period is clean.

Checking whether it worked

After the update, hold the menu steady. Give it 30 to 60 days, then repeat the same analysis on the new period. Compare:

  • Average check size

  • Item sales mix, especially the share going to stars and puzzles

  • Contribution margin per guest (total margin divided by covers)

  • Food cost, to confirm that no supplier change is hiding the effect

Set a target before you start, such as moving the average margin per cover up by a set amount, so you know what success looks like. Then put the next review on the calendar. Many operators tie it to seasonal menu changes, which makes it a quarterly routine instead of a one-off project.

If you want a before-and-after you can trust, check at the level of the individual check. Look at what guests ordered together before and after, not just what one dish sold. That is the only way to see substitution at work.

Why the system that takes the order should run the analysis

Most restaurants start with a spreadsheet: export the sales, type in costs, build formulas. It works the first time. By the third quarter, the export is late, the costs are stale and nobody owns the file. The analysis only helps if it stays current.

That is why menu analytics fits best inside the operating system. Shire is an AI-powered restaurant operating system that includes menu analytics alongside POS integration, kitchen display sync, an owner reporting app and floor management that uses your existing CCTV cameras. Sales, tickets and table activity sit in one place, so you read item performance next to the operation that produced it instead of reconciling separate exports. Shire adapts to your workflow without replacing your existing POS or camera systems, and it gives each staff role live insight into the floor.

That matters for the signals above. Slow tickets, long table times and the pace of service are part of why a dish earns what it earns, and they are the numbers a menu spreadsheet never sees.

Shire for your restaurant

Shire Intelligence starts at $40 per month, and Shire POS and Shire Enterprise are available for operators who need a full POS suite or multi-location infrastructure. You can see the full line on the Shire Intelligence page.

Shireintelligence

For a sit-down room, the full-service restaurant software page shows how floor intelligence and staff coordination work together. If speed is your business, see the quick-service restaurant software. The Shire POS page covers orders at the item level, and the Shire Owner App puts reporting in your pocket, so you can review item performance without sitting at the back-office computer.

A simple quarterly routine

You can run this without any ceremony. Block two hours once a quarter:

  1. Export or open the last three months of item sales by meal period.

  2. Refresh plate costs for your top 20 items by sales.

  3. Calculate contribution margin and menu mix for each item.

  4. Sort into the four groups and write one action next to each.

  5. Pick no more than five changes and assign an owner to each.

  6. Brief servers before the new menu goes live.

  7. Diarize the re-check for 30 to 60 days out.

Owners who do this four times a year build a menu that gets sharper with each pass, because every cycle starts from a cleaner baseline than the one before.

FAQ

What is the difference between menu analytics and menu engineering?

Menu analytics is the measurement: sales, cost, margin and popularity for each item. Menu engineering is the action you take from it, such as repricing, repositioning or removing dishes. In practice you do both in one cycle, and the analytics part is what makes the changes defensible.

How much sales data do I need?

Use at least three months of POS data, as the National Restaurant Association’s operations textbook advises, and keep meal periods and bar versus kitchen separate. A new restaurant with less history can start with what it has and repeat the analysis as the data grows.

Should I judge dishes by food cost percentage or contribution margin?

Contribution margin is the better guide for menu decisions because it shows the dollars each dish leaves you. Food cost percentage still helps you monitor purchasing and portion control across the whole menu, so track both and rank dishes by margin.

How often should I review my menu?

A quarterly review suits most restaurants, ideally timed with seasonal menu changes. After any change, wait 30 to 60 days before re-analyzing so the results reflect the new menu and not the transition.

Can a small restaurant use menu analytics?

Yes. A single location with a modest menu has fewer items to score, so the first pass takes less time. The same steps apply: margin, mix, four groups, a few changes, a re-check. Shire Intelligence is priced from $40 per month, with tiers for smaller operations.

Sources

Recommended

Written with BabyLoveGrowth tools

PUT IT INTO PRACTICE

Turn restaurant data into better decisions.

Explore how Shire connects restaurant operations, service, and reporting in one system.