Your Best-Seller May Be Your Worst Menu Decision

Let me start with the decision that usually creates the problem.

You run your menu-engineering report and an item appears as a Star.

It is popular. Its contribution is relatively strong. Everybody in the room can understand the conclusion quickly. “Good product. Sell more.”

That may be sensible. But I would not conclude that yet. Popularity tells me that guests are buying the item. Contribution tells me something about the economics of each sale. Neither one, on its own, tells me whether increasing sales of that item will improve the restaurant. That is the first limitation I want you to keep in mind.

A Star can still create the wrong operating result

Suppose your strongest main course sells in large volume and gives you good contribution per plate. Now suppose that during your busiest two hours it occupies the station that is already constraining the kitchen. Every additional sale uses scarce grill time. Service begins to slow, another item queues behind it and the operating team starts working around the product rather than through it. Would the item still classify as a Star?

Yes. Nothing in that description automatically makes the popularity or contribution calculation wrong.

What changes is the management question.

Instead of asking:

“Can we sell more?”

I would ask:

“What happens to the operating system if we sell more?”

That is a much better starting point.

A Star is a classification. It is not an instruction.

The matrix is useful because it simplifies the menu

I do not want to make traditional menu engineering sound weak. Its simplicity is one of its strengths. A restaurant might have twenty, thirty or fifty items. Looking at every dish individually can quickly become an unfocused discussion. Popularity and contribution give management a fast way to organise the menu into different commercial patterns.

A relatively popular, relatively high-contribution item behaves differently from a low-volume, low-contribution item. That is useful.

The problem begins when management moves directly from classification to prescription.

The four-box matrix is very good at saying:

“Look here.”

It is much weaker at saying:

“Do this.”

That is why this guide preserves the classical screen instead of replacing it. SCREEN is the first stage of the TRAIL method, but it is deliberately treated as orientation rather than instruction.

Take the Plowhorse

A Plowhorse is popular but generates relatively lower contribution.

The obvious management reaction is often:

“Raise the price.” Maybe. But why is the contribution low? Perhaps the item really is underpriced.

Perhaps the recipe cost has moved and nobody updated the selling price. Perhaps the portion has become larger than the intended specification. Perhaps discounts are pulling the net selling value down. Perhaps the product is deliberately positioned as an accessible entry item.

Perhaps the guest who buys it also buys a drink or side with very attractive economics. Or perhaps the item is already doing an important commercial job and the better decision is to improve recipe architecture rather than put demand at risk. Those situations all produce the same broad classification. They do not justify the same action.

Now take the Puzzle

A Puzzle produces relatively good contribution but lower popularity.

A common response is:

“Promote it.” Again, maybe. But what if the low demand exists because the price-value relationship is wrong? The item may have attractive contribution per plate precisely because the selling price is high. If that price is suppressing demand, spending more money to promote the product may simply push more marketing behind the wrong proposition.

The problem could also sit somewhere else. Perhaps the item is frequently unavailable. Perhaps the description is unclear. Perhaps the preparation is difficult and the service team does not recommend it enthusiastically because they know what will happen in the kitchen.

Perhaps the item is behaving exactly as management intended because it is a premium niche choice rather than a volume product. One classification. Several possible explanations. Several very different decisions.

And then there is the Dog

The Dog is where automatic menu engineering can become particularly dangerous. Low popularity. Lower contribution. Remove it.

That conclusion sounds logical until you ask what disappears with the item. Suppose it is the only credible vegan main course. Suppose it provides an important dietary option. Suppose it is a local signature item.

Suppose it gives the menu an important lower or higher price point. Suppose it is required for a package or a specific guest occasion.

Its standalone sales may be weak while its line-up role remains important.

That does not mean the item deserves permanent protection. A weak vegan curry can still be replaced with a better vegan main course. A local signature dish can still be redesigned. A product can keep the same role while the SKU changes.

But the decision has changed from:

“Remove the Dog.”

to:

“If I remove this item, what do I need to replace so the menu still performs its intended role?”

That is a much stronger management conversation. Later in the guide we will compare two Dogs from the worked example: the Vegan Curry, which protects a required menu role, and the Imported Chocolate Tart, which does not protect anything equally unique. The matrix classifies both as Dogs, but the wider evidence points management in different directions.

Same quadrant, different role, different management decision.

A monthly result can also be temporary

There is another weakness in using the quadrant as the answer. The quadrant itself can move. Suppose an item has been a Puzzle for several months and then suddenly becomes a Star. What changed?

A promotion may have been running. A food festival may have shifted demand. A competing item may have been unavailable for part of the period. The menu position may have changed.

The recipe may have changed. The price may have changed. The hotel may simply have had a different guest mix. The current classification can therefore be mathematically correct while still being a poor basis for a permanent menu decision.

The opposite can also happen. A seasonal item can look weak when averaged over the full year but perform exactly as intended during the season when the item matters. I do not want you to stop reviewing the menu monthly. A monthly movement is useful because it tells you that something changed.

I want you to separate the review frequency from the decision horizon.

Review regularly. Make structural decisions from representative evidence. The Seafood Plate worked example later in the guide will show this clearly: the item moves from Puzzle to Star during a specifically identified promotional period and then returns to Puzzle. The Star classification is real for that month, but it does not establish that the item permanently became a Star.

A promoted month can be real and still be non-representative of normal demand.

The item may be profitable and still be a poor use of capacity

There is also a cost question that the classic matrix does not fully resolve. Traditional contribution is normally built from the item’s selling economics less product cost. That is useful and we will calculate it properly in the next chapters. But suppose two dishes both generate strong contribution.

One occupies the constrained grill for ten minutes. The other occupies it for fourteen. If the grill has plenty of available capacity, the difference may not matter very much. If the grill is already the bottleneck at peak dinner, the difference becomes commercially important.

The second item may give more contribution per plate and less contribution from each minute of the scarce resource. That does not automatically make it a bad product. It simply gives management another piece of evidence that the four-box model does not contain.

This is the reason TRAIL separates Retained Economics from Activity & Capacity rather than trying to force every operational issue into one contribution number. The framework limits bottleneck-minute economics to situations where a genuine constrained resource has actually been identified.

The item may also affect other items

The same issue applies to what the guest buys around the item. A Club Sandwich may frequently appear on the same check as a Local Beer. A premium Seafood Plate may have a strong association with a glass of wine. If you reprice, remove or promote the focal item, part of the commercial effect may appear somewhere else on the check.

There is another side to this. If Seafood becomes less attractive, some guests may choose Grilled Chicken instead. If you remove one dessert, some demand may shift to another dessert rather than disappear. The item is therefore not always the complete unit of economic analysis.

Sometimes I need to look at the basket, the category or the system effect.

That does not mean every association is causal. The framework is explicit about that boundary. Items appearing together does not prove that one caused the other to be purchased, and associated contribution must not be added again as if it were new restaurant revenue. The relationship is evidence to investigate and, where worthwhile, test.

So what is the matrix actually telling me?

By now, I hope the distinction is becoming clear. When an item appears as a Star, Plowhorse, Puzzle or Dog, I do not want the discussion to end. I want it to begin.

A Star asks:

Why is this working, and what should I protect before I try to increase it?

A Plowhorse asks:

Why do guests want this item while its contribution remains relatively weak?

A Puzzle asks:

Why does an economically attractive item attract relatively less demand?

A Dog asks:

Is this genuinely a weak item, or is it performing another required job on the menu?

Those questions preserve the value of traditional menu engineering without allowing the label to make the decision for us.

What I recommend instead

The solution is not to replace one simple model with a more complicated automatic model. I do not want TRAIL to become another spreadsheet that tells you what to do. The companion workbook deliberately calculates and flags evidence but leaves the management action open.

The sequence is:

SCREEN the menu using the classical model.

Then follow the priority item through:

Trend & Target

Retained Economics

Activity & Capacity

Interactions

Line-up Role

Then TEST the proposed intervention.

The test matters because even a very good diagnosis is still a hypothesis until you change something and observe what happens.

SCREEN tells you where to look. TRAIL builds the decision. TEST verifies it.

What I want you to stop doing

Before we move into the calculations, I want you to make one behavioural change. Stop translating the quadrant directly into an instruction.

Do not say:

“Star — promote.” “Plowhorse — increase the price.” “Puzzle — promote.” “Dog — remove.”

Instead say:

“This classification tells me where I need to investigate.”

That single change protects you from many of the weak decisions we will address throughout the rest of the guide.

Before you move on

Take one menu item that management has recently discussed changing. Write down its current classification if you already have one. Then, without making a decision yet, ask five simple questions: Is the current result representative?

What does the item really retain? What operational load does it create? What does it interact with? What role would disappear if I removed it?

If you cannot answer those questions yet, that is fine.

You have already learned something useful:

the quadrant is not enough evidence to act.

In the next chapter, I will go back to the classical menu-engineering model and show you exactly what it still does very well, how we calculate it, and where I want you to stop before the label becomes the decision.

In the first chapter, I deliberately challenged the way the four-box model is sometimes used. Now I want to do the opposite. I want to defend it. Traditional menu engineering remains useful because it forces us to look at two basic commercial questions that every restaurant should understand:

What are guests choosing?

And:

What does each sale contribute?

Those two questions give us a fast way to screen a menu that may contain dozens of different products. The problem is not the matrix. The problem is asking the matrix to make a decision it was never designed to make. So before we follow an item through TRAIL, I want to make sure we understand the classical screen properly.

Start by comparing the right items

Imagine that I give you one list containing:

• a Club Sandwich
• a glass of wine
• a cheesecake
• a cocktail
• a grilled chicken main course

and a coffee.

Then I ask:

“Which one is popular?” You can calculate which item sold the most, but the comparison may not tell you very much. A guest ordering a main course has a different purchase opportunity from a guest deciding whether to order dessert. A beverage may be purchased with the meal rather than instead of it.

Lunch and dinner may behave differently. Room service may behave differently from the restaurant. For that reason, the training workbook keeps the classical screen within controlled categories. Mains are compared with Mains.

Desserts with Desserts. Beverages with Beverages. We will go deeper into the analysis population in the next chapter.

For now, remember one simple rule:

Compare items that had a reasonably similar opportunity to be purchased.

Otherwise, your calculation may be mathematically correct while the comparison itself is weak.

Build the screen from a meaningful analysis population.

The first question: what are guests choosing?

The first axis of classical menu engineering is popularity.

In practical terms, I want to know how much of the relevant menu mix each item represents. Let us use the four Main Course items from the TRAIL training example.

During the current period, the sales are:

These are synthetic training figures from the companion workbook, not property results. Now I can calculate the menu mix.

For the Club Sandwich:

260 ÷ 780 = 33.3%

For Grilled Chicken:

300 ÷ 780 = 38.5%

For the Local Seafood Plate:

100 ÷ 780 = 12.8%

And for the Vegan Curry:

120 ÷ 780 = 15.4%

Immediately, I can see something useful. The Club Sandwich and Grilled Chicken account for much more of the Main Course mix than Seafood or Vegan Curry. That does not yet tell me whether any of these items are good or bad. It simply tells me what guests are choosing relatively often within this population.

Popularity is a screen within the chosen category, not a universal judgement.

We still need a popularity benchmark

If I only know that an item represents 15% of category sales, I still need some basis for deciding whether that is relatively high or low.

The workbook uses a conventional popularity factor of 70% of equal-share popularity as an editable screening default. It is deliberately labelled as a conventional default rather than a universal rule.

Let me show you the calculation. There are four Main Course items.

If every item sold equally, each would represent:

100% ÷ 4 = 25%

The workbook then applies the 70% popularity factor:

25% × 70% = 17.5%

So for this training example, the popularity threshold is:

17.5%

That means:

Club Sandwich at 33.3% is above the threshold. Grilled Chicken at 38.5% is above it. Local Seafood Plate at 12.8% is below it. Vegan Curry at 15.4% is below it.

We have now divided the category into relatively higher-popularity and lower-popularity items. But I want to put a warning beside that calculation immediately.

Do not turn 70% into a law

Please do not take the 70% factor from this example and treat it as a universal hotel or restaurant standard. The workbook makes the factor editable for a reason. A category with four permanent Main Courses is different from a long cocktail list. A seasonal menu is different from an all-day menu.

A buffet decision environment is different from an à la carte menu. And even within the same outlet, the correct analysis population may change depending on the management question.

So the threshold is useful as a screening device.

It is not a substitute for judgement. Later, when we reach Trend & Target, we will ask whether the resulting classification is even representative enough to deserve confidence. For now, the popularity calculation gives me one axis of the matrix. I still need the second.

The second question: what does each sale contribute?

The second axis is contribution.

For the classical SCREEN used in this guide, I calculate what the workbook calls:

Classic Contribution Margin per unit.

The calculation is deliberately straightforward:

Net revenue per item − recipe or product cost = Classic Contribution Margin

Notice that I said net revenue, not automatically the printed menu price.

That distinction matters.

Suppose the Club Sandwich is listed at $19.00.

But the average discount or allowance attached to the item is $0.30.

The economic selling value used in the training workbook is therefore:

$19.00 − $0.30 = $18.70

The current recipe or product cost is:

$6.80

So the Classic Contribution Margin is:

$18.70 − $6.80 = $11.90

The workbook uses the same approach for the other Main Courses.

Now we have something different from popularity. The Local Seafood Plate sells relatively few units, but it produces the highest Classic Contribution Margin per sale in this group. The Club Sandwich sells much more frequently, but its contribution per sale is lower. Already, the menu is starting to tell us a more interesting story.

Classic Contribution Margin keeps the first screen simple and recognisable.

Now we need a contribution benchmark

Just as popularity needs a benchmark, contribution needs one too.

In the TRAIL training workbook, the four Main Courses are compared with the category’s weighted-average Classic Contribution Margin.

For this example, that benchmark is approximately:

$14.18 per item sold

So:

Club Sandwich at $11.90 is below the benchmark. Grilled Chicken at $15.30 is above it. Local Seafood Plate at $19.00 is above it. Vegan Curry at $12.30 is below it.

Now we have both axes.

Popularity:

above or below 17.5%

Contribution:

above or below $14.18

And now the four-box model becomes very useful.

One category. Four different patterns.

Look at what happens when we place the four Main Courses into the matrix.

Grilled Chicken — STAR

Grilled Chicken has:

38.5% menu mix

and

$15.30 Classic Contribution Margin

It is above both benchmarks.

So it becomes a:

STAR

In classical terms, this is a relatively popular item with relatively strong contribution. That is useful information. But remember Chapter 1.

It does not yet mean:

“Promote it harder.”

Club Sandwich — PLOWHORSE

Club Sandwich has:

33.3% menu mix

and

$11.90 Classic Contribution Margin

Popularity is above the benchmark. Contribution is below it.

So it becomes a:

PLOWHORSE

Guests clearly want the product. The economics per sale are relatively weaker than the category benchmark. That deserves management attention. It does not yet tell us whether the answer is price, portion, recipe, discounting or something else.

Local Seafood Plate — PUZZLE

The Local Seafood Plate has:

12.8% menu mix

and

$19.00 Classic Contribution Margin

Popularity is below the threshold. Contribution is well above the benchmark.

So it becomes a:

PUZZLE

Each sale looks economically attractive under the classical calculation. But relatively fewer guests are choosing it. That gives me a very useful question. Why?

Vegan Curry — DOG

The Vegan Curry has:

15.4% menu mix

and

$12.30 Classic Contribution Margin

It is below both benchmarks.

So it becomes a:

DOG

Again, the classification is useful. It tells me that the item is relatively weak on both classical dimensions.

What it does not yet tell me is whether the right answer is removal.

We will eventually discover that this particular item has an important dietary and assortment role in the synthetic menu example. But the classical matrix does not know that yet. And that is exactly the point.

One category can contain four very different commercial patterns.

This is what the matrix does very well

Look at what we have achieved with a fairly simple calculation. We started with four Main Courses. Within a few minutes, we can see four different commercial patterns:

Grilled Chicken: strong popularity, strong contribution.

Club Sandwich: strong popularity, weaker contribution.

Local Seafood Plate: weaker popularity, strong contribution.

Vegan Curry: weaker popularity, weaker contribution.

That is a meaningful management screen. If the restaurant had thirty items, this kind of classification could help us quickly identify where management attention is likely to be most useful. This is why I do not want to throw away classic menu engineering. It reduces complexity.

It makes patterns visible. And it gives different departments a common language. A Chef, Restaurant Manager, Finance Manager and GM can all look at the same matrix and quickly see where the commercial tension sits.

The workbook therefore describes the matrix explicitly as an orientation screen. It is designed to tell management where to investigate rather than automatically decide the item.

Translate the classification into a question

This is the point where I want you to develop a new habit.

Every time you see a quadrant, convert it into a question.

Not an instruction.

For a Star, ask:

Why is this item working, and what must I protect before I try to improve or increase it?

For a Plowhorse, ask:

Why is an item that guests clearly want producing relatively weaker contribution?

For a Puzzle, ask:

Why is an economically attractive item not being chosen more often?

For a Dog, ask:

Is this item genuinely weak, or does it perform another role that the two-axis matrix cannot see?

That is also how the companion workbook frames the classifications: every item ends with the same essential instruction—

Investigate. Do not act from the label alone.

Convert each quadrant into an investigation question, not an automatic action.

Why I keep Classic Contribution separate

There is one more design choice I want to explain before we leave this chapter.

You may notice that the workbook contains both:

Classic Contribution Margin

and

Retained Contribution Margin.

We are deliberately not using Retained Contribution Margin to build the classical four-box screen.

Why? Because I want the SCREEN to remain recognisable and clean. First, we calculate the conventional menu-engineering view properly. Then we challenge it with additional evidence.

Later, under R — Retained Economics, we will bring in costs such as supported packaging, channel or order costs, and other directly decision-caused costs.

For example, the Club Sandwich has a Classic Contribution Margin of $11.90, but the synthetic workbook later shows Retained Contribution of $10.90 after supported direct decision costs are included.

That later view is important. But I do not want to quietly mix the two calculations and then pretend the traditional quadrant means something different.

The sequence matters:

Calculate the classical screen cleanly.

Then:

challenge it with better evidence.

SCREEN is a filter, not the final decision

So where are we now?

We know:

Grilled Chicken — Star

Club Sandwich — Plowhorse

Local Seafood Plate — Puzzle

Vegan Curry — Dog

If I were sitting in the menu review meeting, would I now make four decisions? No.

I would say:

“Good. Now I know where I want to look.”

The next step is not:

• raise the Club Sandwich price
• promote Seafood
• protect Chicken

and remove Vegan Curry. The next step is to ask whether these comparisons were built on the right population and whether the current classifications deserve to be interpreted the way we think they do. That is why the first part of our overall framework is called:

SCREEN

It tells me:

Where should I investigate further?

It does not tell me:

What should I do?

If you only remember one thing

Traditional menu engineering is useful because popularity and contribution quickly reveal different commercial patterns across the menu.

Its value comes from screening the menu, not from automatically prescribing the action.

Ask this question

When you look at a Star, Plowhorse, Puzzle or Dog, ask:

“What question is this classification asking me to investigate?”

Do not begin with:

“What action normally belongs to this box?”

Do this next

Take one category from your own menu. Do not analyse the full restaurant yet. Choose one reasonably comparable group—perhaps Main Courses, Desserts or Cocktails.

For each item, gather:

• units sold
• actual net selling value where available

and current trusted recipe or product cost. Calculate the menu mix and Classic Contribution Margin. Then classify the items. Stop there.

Do not reprice anything. Do not remove anything. Do not start promoting anything.

In the next chapter, we will build the SCREEN more carefully and deal with one of the most important questions in the entire process:

Are we comparing the right items in the first place?

By now, you know how the four-box matrix works. You know how popularity and Classic Contribution Margin can place an item into one of four broad patterns. Before we move into TRAIL, I want to spend a little more time on the SCREEN itself.

The mathematics is not the difficult part. The quality of the screen depends heavily on what you decide to compare. I can calculate a perfect percentage from the wrong population and still end up with a weak management signal.

SCREEN has one job

I use SCREEN to answer one question: where should I investigate further? I am not trying to set the final price, decide whether an item stays on the menu, allocate kitchen labour or explain the whole basket. Those questions come later.

At this stage I want a quick commercial orientation. What are guests choosing? What is contributing? Which items deserve a closer look?

Start with the decision population

Suppose the POS gives you hundreds of item lines. You could put mains, desserts, beer, wine, coffee, cocktails and room-service items into one matrix. I would not do that, because those items did not all have the same opportunity to be chosen.

A guest choosing a main course is not normally choosing between Grilled Chicken and a glass of wine. The wine may be bought with the meal. Dessert happens at another point in the occasion. Room service may behave differently from the restaurant.

So I start with a controlled population: Mains with Mains, Desserts with Desserts, Beverages with Beverages. If daypart or channel materially changes the purchase opportunity, I may split further.

Build the screen from items that had a reasonably comparable opportunity to be purchased.

Sometimes category is enough

If eight dinner mains are available on the same menu during the same period, analysing them together may be perfectly sensible. I would not create extra segments simply because the POS allows me to.

Sometimes category is too broad

An All-Day menu can be different. The Club Sandwich may sell from lunch through late night, while some dinner mains appear only after 6 p.m. A broad category can hide those differences. When the decision changes by daypart, channel or concept, split the population where that separation makes the comparison more meaningful.

Do not segment until the data become meaningless. The purpose is not to create the most granular report possible. It is to create a comparison management can interpret.

Write the population on the report

I recommend recording the outlet, population and period at the top of the SCREEN. A menu mix of 20 percent means very little unless everybody knows: 20 percent of what? The denominator is part of the decision.

What evidence do I need?

For the classical SCREEN, I need only a small set of controlled fields: item, category or population, units sold, net revenue per unit and current recipe or product cost. From those fields I can calculate menu mix and Classic Contribution Margin.

Workbook reference — 02 ITEM DATA: the first SCREEN inputs sit together before the deeper TRAIL evidence.

Make sure the revenue and cost inputs are credible

If discounts or allowances are material, I prefer a reliable net selling value rather than a printed menu price the operation did not actually earn. I also want the recipe or product cost to be current enough to use. A correct formula cannot rescue stale source evidence.

Build the screen in two steps

First calculate popularity: item units divided by total units in the analysis population. Then compare that result with the chosen popularity threshold. Second calculate Classic Contribution Margin: net revenue per item less recipe or product cost, and compare it with the category contribution benchmark.

The four-item Main Course example turns the calculations into a simple orientation screen.

Let the workbook calculate, but do not let it think for you

I do not want management manually rebuilding the same calculations every month if the tool can do them reliably. Let the workbook calculate and flag. But management still has to understand how the population was defined, whether the product cost is credible and whether the period is representative.

The tool can give you a clean answer to the calculation. It cannot decide whether the question was well designed.

When should I stop the SCREEN?

Stop when you know what population you analysed, what period you analysed, what guests chose, what Classic Contribution Margin shows and which quadrant each item currently occupies. Do not continue automatically into repricing, promotion, deletion or recipe redesign.

The next question is whether that clean-looking classification is a stable pattern, a temporary movement or a distorted month. That is where we begin the TRAIL with T — Trend & Target.

A correct month can still lead to a wrong decision

In the previous chapter, we built the traditional menu-engineering screen. Now I want to show you one of the easiest ways to misuse it. You run the analysis at month-end. An item becomes a Star.

So you promote it. Another becomes a Dog. So you start discussing whether to remove it. The calculations may be completely correct.

But the decisions may still be wrong. Because one month is a snapshot. And most menu decisions are not one-month decisions. If I change a recipe, remove a dish, increase its price, redesign the menu or change the kitchen around it, I'm making a decision that can affect several future periods.

So before I make that decision, I want to know: Is what I am seeing now actually representative? That is the first letter in TRAIL. T — Trend and Target.

Review monthly. Decide from representative evidence.

Workbook reference — 04 TREND & TARGET: current class, representative periods and trend status are visible together.

I am not saying you should stop reviewing menu engineering every month. I recommend exactly the opposite. Review it regularly. A monthly movement is useful because it tells you:

“Something changed. Go and look.”

But there is an important difference between the review frequency and the decision horizon. I can monitor an item monthly. But before making a structural decision, I want to see whether the result persists across representative periods. And I use the word representative deliberately.

I don't simply mean:

“Take the last three months and average them.”

A three-month average that includes a food festival, two weeks of stockout and a major promotion may be worse than one clean comparable month. So the question is not: How much history do I have? It is:

How much comparable, usable history do I have?

The Seafood Plate: mathematically a Star, operationally still a Puzzle

Trend separates a representative pattern from a promoted or otherwise distorted month.

Let's use the Local Seafood Plate in the workbook. In January, February, March and April, it sits in the Puzzle quadrant. Good contribution. Lower popularity.

Then look at May. Its sales jump dramatically. It becomes a Star. If I reviewed May in isolation, I could easily say:

“Excellent. The seafood plate has finally worked. Let's push it harder.”

But the workbook contains another piece of evidence. May was affected by a seafood festival and promotion that materially lifted demand. So I don't delete May. The sales happened.

The revenue is real. The guests bought the product. But I mark that period as distorted for the purpose of deciding the item's normal position. Then June comes.

The promotion is over. The Seafood Plate returns to Puzzle. Now look at the story. It's not:

Puzzle → Star → management success. It's: Puzzle → Puzzle → Puzzle → Puzzle → promotion-distorted Star → Puzzle. That tells me something completely different.

The May result may prove that the product can respond to an intervention. That's valuable evidence. But it does not prove that the item's underlying position permanently changed. This is why I keep saying:

The data can be right and the interpretation can still be wrong.

Distortion is not bad data

This distinction is important. A promotion is not bad data. A banquet group is not bad data. A festival is not bad data.

A price change is not bad data. Those are real operating events. What matters is whether I treat that period as representative of normal trading. So when I review the trend, I want you to record the reasons that can change the classification.

For example: a promotion; a menu-price change; a recipe change;

a food festival; a package inclusion; a temporary guest segment; a server incentive;

a menu-placement change; a delivery-platform campaign; or a competing item being unavailable. Once I know that something changed, the movement becomes much more useful.

Instead of asking:

“Why did this item become a Star?”

I can ask:

“What intervention caused demand to move, and do I want to repeat it?”

That's a management question.

Availability: don't punish an item for not being available

Check availability before interpreting weak demand.

There is another distortion I want you to watch carefully. Availability. Suppose an item was available for only eighteen days in a thirty-day month because a supplier could not deliver one of the ingredients. And then your monthly report says:

“Low sales. Dog.”

That is not a fair test of guest demand. You didn't give the guest the same opportunity to buy it. The workbook deliberately tracks available days against possible days. In our training settings, a period below the minimum availability threshold is treated as non-representative by default.

That threshold is editable. It is an operating control, not a universal industry rule. The principle matters more than the percentage: Don't interpret low demand until you know the product was actually available to be demanded.

This sounds obvious. But if sales data and availability data sit in different systems, this mistake happens very easily.

Stable, moving, distorted or evidence required

Use simple management flags to describe the quality of the trend evidence.

Once I've cleaned the timeline, I want a simple management view. I don't need a complicated statistical model just to start the conversation. In the TRAIL workbook, an item's trend can help me distinguish between situations such as:

STABLE.

The current classification keeps appearing across representative periods.

MOVING.

The classification is changing enough that I should be cautious about treating today's box as permanent.

DISTORTED.

A known event materially affected the period I am looking at. Or:

EVIDENCE REQUIRED.

I simply don't have enough representative information to make the decision confidently. These labels are not a statistical confidence interval. They're management flags. They're there to stop me from pretending that every monthly classification deserves the same level of confidence.

And the default settings in the workbook are editable because different hotels and outlets will have different volumes, seasons and operating rhythms.

Trend tells me what is happening. Target tells me whether I should care.

Trend shows what is happening; Target tells you whether the position is actually off plan.

Now let's move to the second half of T. Target. By Target, I don't mean a budget number that somebody typed into Excel. I mean:

What role was this item supposed to play on the menu? Look at the Club Sandwich. In the current period, it is a Plowhorse. Popular.

But below the category contribution benchmark. Now look across the history. It isn't just a Plowhorse this month. It has remained a Plowhorse across the representative periods in the workbook.

So I would call that a persistent pattern. But management's intended position for the Club Sandwich is:

Star.

Now I have a useful management gap. Actual position: Plowhorse. Intended position: Star. That does not mean:

“Force it into the Star box.”

It means: Why is an item that management intended to be both popular and economically strong consistently falling short on the economics? Now I know where to investigate. Maybe the price is wrong.

Maybe the recipe economics have deteriorated. Maybe the portion has become too generous. Maybe discounting is pulling down the retained value. Maybe we deliberately need it as an accessible all-day anchor and the original target was unrealistic.

The target gives me context. It does not dictate the answer.

The reverse case: a Puzzle may be exactly where we wanted it

This also works in the opposite direction. The Local Seafood Plate is a Puzzle. Good contribution. Lower popularity.

Should I automatically try to turn it into a Star? Not necessarily. In the workbook, its intended position is also Puzzle. Why?

Because it has a role as a local-identity and premium-anchor item. Perhaps management never expected it to become the highest-volume main course. Its role may be to: represent the destination;

offer a premium choice; support the price ladder; create a signature dining story; and appeal strongly to a smaller segment.

If it is fulfilling that role and its economics are sound, then being a Puzzle may not be a failure at all. This is why classification without intended positioning can be misleading. Two items can sit in exactly the same quadrant and require completely different management conversations.

Where this takes us next

So before you touch price, promotion, recipe or menu placement, I want you to complete the T in TRAIL. Ask: What is the item doing now? What has it done across representative periods?

What distorted those periods? Was it actually available? And: What job was the item intended to do?

When I have that, I can separate: a temporary movement; a persistent problem; an intervention that worked;

a classification that is behaving exactly as designed; and a situation where I simply need more evidence. Now we're ready for the next question. Because even if the trend is stable, the traditional contribution number may still be incomplete.

So in the next chapter, we're going to look at: R — Retained Economics. Not just:

“What is the food cost?”

But:

“When I sell this item, what does the hotel actually retain?”

The contribution number may not be the contribution you actually keep

We've now screened the menu. And we've checked whether the classifications are representative. The next letter in TRAIL is: R — Retained Economics.

This is where I want you to challenge one of the most common assumptions in traditional menu engineering. We normally calculate contribution as: selling value minus food or product cost. That is useful.

But it doesn't necessarily tell me what the hotel actually retains from that sale. Because the sale may also cause other costs. A discount. Packaging.

An ordering or channel charge. A commission. A specific consumable. Or another direct cost that exists because I sold that item in that particular way.

So now my question changes from:

“What is the food contribution?”

to:

“What does this sale actually leave behind before I consider the wider operating load?”

That is Retained Economics.

Start with the revenue you actually retain

The first adjustment happens before we even talk about cost. I want you to use the net revenue from the item, not automatically the printed menu price. Look at the Club Sandwich in our workbook. The list price is:

19 dollars. But the average discount or allowance is 30 cents. So the net revenue per sandwich is: 18 dollars 70.

That's the economic starting point I want. Not 19. Now, in your property, the difference could come from: a discount;

a loyalty benefit; a promotional offer; a package treatment; an employee discount;

or another controlled commercial adjustment. I'm not telling you that every theoretical discount must be allocated to every item. I'm saying: Use the selling value the operation actually earns from the transaction, where you have reliable evidence.

Otherwise, we can begin the menu analysis with revenue that the restaurant never actually retained.

Classic contribution is still useful

Retained Economics extends the classical item contribution without pretending to calculate “true profit”.

Now let's calculate the classical contribution. For the Club Sandwich: Net revenue is 18.70. Approved recipe or product cost is 6.80.

So: 18.70 minus 6.80 gives us 11.90. That's the Classic Contribution Margin per unit. And I still want that number.

Remember, we used it in our four-box SCREEN. The problem is not the calculation. The problem is assuming that 11.90 is the end of the economic story. Because now I ask:

What other costs occur directly because this sandwich was sold? In our synthetic training example, the Club Sandwich has: 60 cents of packaging or item consumables and

40 cents of other direct decision cost. That's another dollar attached to each sale. So the Retained Contribution becomes: 18.70 minus 6.80 minus 0.60 minus 0.40

which gives us: 10 dollars 90. Classic contribution: 11.90.

Retained contribution: 10.90. Same sandwich. Two different views.

One dollar per item becomes material at volume

Now one dollar may not sound dramatic. But this is a popular item. We sold 260 Club Sandwiches in the current period. Using Classic Contribution:

260 multiplied by 11.90 gives us approximately: 3,094 dollars. Using Retained Contribution: 260 multiplied by 10.90 gives us approximately:

2,834 dollars. That's a difference of: 260 dollars in one period. And the workbook flags those direct costs as material for review.

Now imagine repeating that across: hundreds of transactions; multiple items; room service;

takeaway; different selling channels; and twelve months. This is why I don't want management looking only at food-cost percentage and assuming the remaining contribution is fully available to cover the operation.

Some of that value may already have been consumed by the way we sold the item.

What belongs in Retained Economics?

Include only supported directly attributable or decision-caused costs.

Workbook reference — 06 RETAINED ECONOMICS: direct costs and retained contribution sit in one focused view.

So what should you include? My rule is quite strict. Include costs that are: directly attributable

or decision-caused and supported by evidence. For example:

the current trusted recipe or product cost; a channel or order charge that arises from the sale; item-specific packaging; specific consumables;

or another identifiable cost that exists because that transaction happened. The workbook has separate fields for exactly those items: Channel or order cost. Packaging or consumables. Other direct decision cost. But I do not want you inventing numbers simply because the spreadsheet has a column.

If you cannot support the cost: don't manufacture it. Mark the evidence gap. The objective is a better economic decision.

Not a more complicated spreadsheet.

What does NOT belong here

Now this boundary is extremely important. Retained Contribution is not what I am going to call the “true profit” of the dish. I am not going to take: the Executive Chef's full salary;

kitchen rent; depreciation; electricity; administration;

general stewarding cost; and every other hotel overhead, then allocate them across menu items using an arbitrary percentage and claim:

“This dish earns 3 dollars 42 of true profit.”

That can create false precision very quickly. Some of those costs may matter to a particular decision. But I need evidence that they genuinely change because of that decision. The C03 workbook therefore specifically tells you not to force rent, full chef salary or arbitrary overhead allocations into an item and call the result true profit.

That's deliberate. I want to trace what I can support. And keep the rest in the correct decision layer.

Don't mix Retained Economics with Activity & Capacity

There is another reason I keep those costs out. Some of the most important economics of a menu item are not best understood through another cost allocation. They're better understood through capacity. Suppose one dish occupies your grill for twelve minutes at the exact time when the grill is already the bottleneck.

I could try to allocate more chef salary to that dish. But that may miss the real decision issue. The more useful question might be: What contribution am I generating from that scarce grill capacity?

That's the next letter in TRAIL: A — Activity and Capacity. So in R, I want to establish the economics we can trace directly to the sale. Then in A, we'll look at the operating resources the item consumes.

Keeping those two layers separate prevents double-counting and keeps the logic understandable. The workbook follows exactly that boundary.

Channel economics can change the item

The same recipe can have different economics by selling channel.

Now let's add one more dimension. The same menu item does not necessarily have the same retained economics in every selling channel. Imagine the same dish being sold: in the restaurant;

through room service; as takeaway; or through an external ordering channel. The recipe might be identical.

But the transaction economics may not be. One channel may require: special packaging; a direct order charge;

different discounts; or another transaction-specific cost. So if that difference is material and you have reliable data, don't average everything together and pretend the item has one universal contribution. Analyse the economics at the level where the decision actually changes.

The C03 data-readiness check therefore asks specifically whether the property has controlled channel/order cost and packaging/item-consumable evidence before using those fields. And if you don't have it? That's fine. Start at the analytical level your evidence supports.

Don't invent Level 3 economics from Level 1 records.

A Plowhorse can look weaker — but the action still isn't obvious

Now come back to our Club Sandwich. The classical screen already classified it as a Plowhorse. Popular. But contribution below the category benchmark.

Then Trend & Target showed us that this isn't just a one-month accident. The classification is stable. And management had actually intended this item to perform as a Star. Now Retained Economics tells us something else.

The Classic Contribution is: 11.90. But after directly attributable costs, the retained contribution is: 10.90.

So the economic gap has not disappeared. It has become more visible. Does that mean I raise the price? Still no.

Maybe I will. But first I want to know what is causing those economics. Could I redesign a cost that the guest doesn't value? Could I improve the recipe?

Could I change packaging? Could I change how the product is sold by channel? Could I protect its strong demand and improve the economics underneath it? That's exactly why we're following the TRAIL instead of reacting to the quadrant.

Where this takes us next

So complete the R in TRAIL with four questions. What net revenue do I actually earn from the item? What is the current trusted product cost? What other costs are directly caused by this sale and supported by evidence?

And: How much contribution is actually retained after those costs? But remember the boundary. Retained Contribution is not full restaurant profit.

And it doesn't yet tell me how difficult the item is to produce. A dish can still show excellent retained contribution and be a terrible use of the resource that is constraining your kitchen. So in the next chapter, we're going to look at: A — Activity and Capacity.

And this is where a high-contribution item can suddenly become much more interesting. Because the next question isn't:

“How much does the plate contribute?”

It's:

“What does the plate consume in order to generate that contribution?”

By now, we have a better economic view of the item. We know what the guest paid. We know what the recipe or product cost was. We know which directly supported costs sit behind the sale.

And we know what contribution is retained.

Now I want to ask a different question:

What does the item consume in order to generate that contribution?

That is the A in TRAIL:

Activity and Capacity

This is where a menu item that looks attractive on a per-plate basis can start to look very different operationally.

Because the restaurant does not have unlimited:

• grill space
• prep capacity
• pass capacity
• service time
• equipment

or labour attention. At some point, one of those resources may become the thing that limits what the operation can sell. And when that happens, contribution per plate is no longer enough.

Do not turn this into arbitrary labour costing

Let me start with an important boundary.

Activity and Capacity does not mean that I now allocate the Chef’s full salary across every dish.

I am not trying to create a theoretical labour cost that says:

• this dish carries $1.73 of labour
• another carries $2.41

and therefore one is more profitable. That can quickly become another exercise in false precision. What I want to know is much more practical.

What activity does this item create?

Where does that activity happen?

And:

Is any of that activity consuming a resource that is genuinely constrained?

That is a different question. Sometimes time is simply time. Sometimes time is the resource that is stopping the restaurant from producing or serving more. I need to know which situation I am dealing with. The C03 logic explicitly limits contribution-per-bottleneck-minute analysis to cases where management has identified a genuine constrained resource.

Start with an operating profile

For each priority item, I would look at a small number of operating questions. Where is the item prepared? How much mise en place does it require? Which station does it use?

How much time does it occupy that station? How many service touches does it create? Does it have high waste exposure? How often is it remade or reworked?

And, most importantly:

Does any of that activity actually limit the operation?

I do not want one universal number pretending to explain all of those things. A dish can have low service time and high preparation burden. Another can have simple preparation but create a queue at one piece of equipment. Another can move quickly through the kitchen but create significant waste.

Another can repeatedly come back for rework because execution is difficult. Those are different operating problems.

So I want an operating profile, not one artificial cost-per-minute number.

Activity profile — look at the work the item creates before assigning an action.

MISE EN PLACE → STATION → BOTTLENECK → SERVICE → WASTE → REWORK

First identify the genuine constraint

This is the most important step. Do not begin by calculating contribution per minute. First identify whether there is a real constraint. In the synthetic training example, only one real-time resource is identified as constrained:

The Grill

The reason is clearly stated.

During peak dinner service, approximately 7 p.m. to 9 p.m., there is queue pressure at that station.

The Cold Line is not currently flagged as constrained. The Hot Line is not. Pastry is not. The Bar is not.

That distinction matters. Suppose Pastry has plenty of unused capacity. Reducing a dessert from four minutes to three minutes might sound efficient, but it may not release anything commercially valuable. Now suppose the Grill is already the point where orders are queueing and service time is deteriorating.

One Grill minute has a very different operational value.

So I do not begin with:

“How long does this dish take?”

I begin with:

“Where are we genuinely constrained?”

Contribution per constrained minute matters only when the resource is genuinely constrained.

More time is not automatically worse

This is worth making explicit. A dish that takes longer is not automatically a bad item. If the station has spare capacity, the extra time may not be the limiting issue.

Likewise, a very fast item can still create:

• high mise en place
• waste
• rework

or service complexity elsewhere. That is why I would be careful with statements such as: “Every item over ten minutes is inefficient.” The workbook does use operating thresholds in the training example, but they are editable flags, not industry standards.

The useful question is not:

“Is this item slow?”

It is:

“Is the activity attached to this item creating a meaningful operating constraint?”

Now compare the two Grill items

Once I have identified a genuine constraint, I can use a more useful economic measure. Let us compare two Main Courses.

Grilled Chicken

Retained contribution per plate:

$14.80

Constrained Grill time:

10 minutes

So:

$14.80 ÷ 10 = $1.48 per constrained Grill minute

Now look at the Local Seafood Plate.

Local Seafood Plate

Retained contribution per plate:

$18.40

Constrained Grill time:

14 minutes

So:

$18.40 ÷ 14 ≈ $1.31 per constrained Grill minute

This is where the economics become interesting. If I only look at retained contribution per plate:

Seafood looks stronger.

$18.40 is higher than $14.80.

But if the Grill is genuinely the resource limiting dinner output:

Grilled Chicken generates more contribution from each minute of that scarce resource.

$1.48 versus approximately $1.31.

Do not jump from this calculation to “remove Seafood”

This is where the framework can easily be misused again. The Seafood Plate gives less contribution per constrained Grill minute than the Grilled Chicken.

Does that mean:

“Remove Seafood”?

No. That would simply replace one automatic rule with another. We already know from the earlier chapters that the Seafood Plate may have:

• strong retained contribution
• a premium position
• local identity

and later we will see strong beverage interaction and a required line-up role. So the calculation is not the decision.

It tells me about the trade-off.

The better question becomes:

“Can I improve or protect the value of this item while reducing the pressure it creates on the constrained Grill?”

That may lead to a very different intervention from removal.

Do not create a league table of “profit per kitchen minute”

I would strongly avoid turning the whole menu into a ranking called:

Contribution per Kitchen Minute

and then sorting from highest to lowest. Why?

Because the metric only becomes meaningful when:

• there is a real constraint
• the item materially consumes that constraint

and management is considering a decision that affects that capacity. If the Grill is unconstrained at 3 p.m., the same calculation may have very little decision value. At 8 p.m., when the station is queueing, it may become highly relevant.

So even this metric can be time-dependent.

The same dish can have different capacity economics at different points in the day. That is why I want Finance and Operations working together on this part of the review. Finance can calculate the contribution. The POS can tell me what sold.

But the Chef and operating team know where the service actually starts to break.