Case Study 2: The Occasion Was Real. The Position Wasn't.

Boston Market, home meal replacement, and the limits of getting the guest right

Background

Chapter 2 argues that concepts should be built on occasions rather than on cuisines. This case is the argument's most important qualification, because it describes a company that identified a real occasion with unusual clarity, built a national business on it, and then lost the position anyway.

Boston Chicken was founded in Newton, Massachusetts, in 1985, selling rotisserie chicken and home-style sides. In 1992 a group of investors led by executives with backgrounds in high-growth retail took control and began expanding the chain aggressively. The company went public in November 1993 in one of the most widely reported initial public offerings of that year — the shares rose sharply on their first day of trading — and expansion accelerated.

In 1995 the company renamed itself Boston Market, reflecting a menu that had broadened beyond chicken to turkey, meatloaf, ham, and an expanded range of sides. The renaming was a positioning decision, and an explicit one. The concept was no longer "rotisserie chicken." It was home meal replacement: dinner for a household that wanted a home-style meal on a night when nobody was going to cook it.

That is a well-specified occasion. It names the guest (a household, usually with more than one person to feed), the trigger (nobody wants to cook tonight), the time budget (short), the party (family), and — critically — the alternative (cooking, or ordering pizza). It is exactly the field on the occasion sheet that §2.2 says everybody skips, and Boston Market did not skip it.

By the late 1990s the company operated on the order of 1,200 locations. Much of that growth was financed through a structure in which the parent company lent money to area developers who built and operated stores. The structure, and the accounting treatment applied to it, attracted substantial criticism in the financial press. In October 1998 the company filed for Chapter 11 bankruptcy protection.

McDonald's acquired the brand in 2000 and sold it to a private-equity buyer in 2007; it changed hands again in 2020. The following years brought widely reported store closures and legal disputes with vendors and employees. The brand still exists, at a small fraction of its late-1990s footprint.

(Public record only. No financial figures are asserted here beyond what is publicly documented, and no characterization is offered of any individual's conduct beyond the reported facts.)

The operating issue

Run the concept through this chapter's tools and the problem becomes visible before the balance sheet does.

The competitive set was wider than the concept assumed

§2.4 insists that the largest competitor for a weeknight meal occasion is usually not a restaurant. Home meal replacement collided with that fact more directly than almost any concept in modern American foodservice, because home meal replacement is the not-a-restaurant occasion. Its competitive set was:

• Supermarkets, which sell prepared foods at a food-retail margin and can price a rotisserie chicken as a loss leader to drive basket size. This is not a fair fight. A grocery chain does not need the chicken to make money.
• Pizza delivery, which owned the same trigger — nobody is cooking — with a lower price, an established ordering habit, and delivery to the door.
• Casual-dining chains offering takeout on the same occasion, with a broader menu.
• Cooking, which remained free.

The differentiator was copyable by Tuesday

Apply the three-part test from §2.6.

Perceptible? Yes. A rotisserie line is visible, smells like dinner, and reads as home-style.

Hard to copy? No. A rotisserie oven is an ordinary piece of commercial equipment. Through the 1990s American supermarkets installed rotisserie programs at scale, put them near the entrance, and priced them aggressively. The product that distinguished the concept became, within a few years, an item available at a lower price in a store the guest was already visiting.

Worth its cost? Increasingly not, once the item it was built on was available cheaper elsewhere.

This is the differentiation test failing in real time, and it is worth sitting with. Nothing about the occasion analysis was wrong. The concept correctly identified a large, real, recurring need. What it did not have was a defensible answer to sentence three of the three-sentence test — why us instead of the specific alternative — once the alternative acquired the same product.

Supply was added before the position was proven

§2.4 ends with a number founders rarely compute: how much supply of a given occasion a market can absorb. Boston Market's expansion added that supply at national scale and at high speed, and it did so using capital that was raised on the promise of growth rather than earned from proven unit performance.

That sequencing is the deeper issue, and it is not a Chapter 2 problem — it is Chapter 35's, arriving about thirty-three chapters early. A concept is proven by repeat guests at a unit, not by the number of units. When growth capital is available before concept-market fit is demonstrated, the growth becomes the evidence, and the evidence is circular.

A shorter companion: differentiation nobody asked for

A second, smaller case makes the same point from the other end of the map.

Eatsa opened in San Francisco in 2015 selling quinoa bowls through an almost fully automated front of house: order on a screen or an app, collect your food from a numbered glass cubby, and interact with no one. The technology worked, it was widely covered, and it expanded to a handful of locations in several cities. Most of them closed in 2019, and the company pivoted to licensing its technology to other operators, rebranding as Brightloom.

Run the same test. Perceptible? Certainly — it was the most visible thing about the restaurant. Did it change the guest's choice? This is where it fails. The automation solved an operator's problem (labor cost, throughput) and presented it to the guest as a novelty rather than a benefit. Guests choosing lunch weigh food, price, speed, and convenience. Removing the human being was not on that list for enough of them, often enough, and once the novelty was spent there was a bowl of quinoa competing on the merits of the bowl of quinoa.

The lesson is not that restaurant technology fails. It is §2.6's second failure mode stated plainly: a differentiator that solves your problem rather than the guest's is a cost, not a position.

What it shows

Getting the occasion right is necessary and not sufficient. This is the single most important qualification to Chapter 2. The chapter's method will get you a real guest, a real occasion, and a real trade area. It will not, on its own, get you a defensible answer to why the guest picks you — and that third sentence is the one that has to survive contact with competitors who can read a market too.

"Defensible" means: what happens when they copy it? Ask the question at concept stage, in writing, about your one real difference. If the honest answer is "a competitor could match this within a season," your position is not the difference. It has to be something else — the room, the location, the accumulated relationship with a neighborhood, the trained team, or a cost structure they cannot reach — and you should know which, before you build.

The competitor you dismiss is the one that gets you. Boston Market's most dangerous competitor was not another restaurant, and a competitive-set exercise limited to restaurants would never have found it. Grocery prepared-foods counters appear in Figure 2.6 as establishment H for exactly this reason.

Scale is not proof. A thousand locations is evidence that a company can raise money and build stores. It is not evidence that the concept works, and confusing the two is a specific, recurring, and expensive error in this industry.

Lesson

Sentence three of the three-sentence test — "they choose us over [the specific alternative] because [a reason they would recognize]" — is the sentence that fails, and it fails later than the others.

Sentences one and two are wrong immediately if they are wrong at all: nobody comes, or they will not pay. Sentence three can be true on opening day and false in year four, because it is the only one of the three whose truth depends on what other people decide to do.

So write it with the erosion in mind. State the difference, state how long you believe it would take a competent competitor to copy it, and state what you would do on the day they did. A concept that has an answer to that last question is a business. A concept that does not is a good idea with an expiration date nobody has written down.

Discussion questions

1. Boston Market's occasion analysis was, on the evidence, better than most restaurant concepts ever attempt. Reconstruct the three-sentence test as the company might have written it in 1995, then identify precisely where it fails and in what year you would expect the failure to become visible.

2. Supermarket rotisserie chicken is priced to drive basket size rather than to earn a margin on the chicken. Name two other situations in which a restaurant finds itself competing with a business whose economics it cannot match. What should an operator do in that position?

3. The case argues that "scale is not proof." What would constitute proof that a concept works well enough to replicate? Write the three-item checklist you would insist on before opening a second location, and compare it to Chapter 35's treatment when you get there.

4. Apply the three-part differentiation test to Eatsa's automation. Then argue the other side: under what concept, occasion, and trade area might that same automation have passed all three?

5. The chapter's method produced a defensible-looking concept for Bellwether. Using this case as a lens, identify the single element of Bellwether's positioning most vulnerable to being copied, say how long you think it would take, and write the paragraph the plan should contain about it.

6. Both cases here involve concepts that read as correct at launch and were undermined by other people's decisions. Does that make concept development less valuable than this chapter claims, or differently valuable? Defend your answer with reference to what a concept document is actually for.