Company, customer, and competitor is a lens, not a speech. In a live case you almost never say the brand name. You pick the one of the three that actually decides the prompt, split it into numbers, and ignore the other two until they matter.
The tree is useful when the question is “can we win?” rather than “why did profit fall?” Winning is a fit problem: who the buyer is, what you can do that they will pay for, and who else can do it cheaper or better. Reciting three buckets and touring all of them is how candidates burn eight minutes and still have no hypothesis.
The tree, not the slogan
State the decision in one sentence, then hang three branches that do not overlap:
| Branch | What you actually test | A number that belongs here |
|---|---|---|
| Customer | Who buys, what they pay for, how they switch | Willingness to pay, frequency, mix of segments |
| Company | What you can deliver at a cost that still leaves margin | Unit cost, capacity, capability you already have |
| Competitor | Who else the customer will compare you to | Relative price, relative quality, likely response |
If a fact could sit in two branches, you have not defined the branches. “Brand” is not a branch. Brand is evidence under customer (do they care) or company (can we sustain it).
A clean opening sounds like: “I would start with whether the business customer actually values on-time shuttles enough to pay a premium, then check whether we can deliver that reliably, then whether the low-cost carrier can copy it in a year.” That is the same tree. You did not name a framework.
Worked example: regional airline, not a grocery promo
Prompt. Skyline Air flies 48 turboprops on city pairs under 400 miles. Load factor is 72%. A low-cost carrier just opened six of the same routes at fares 18% below Skyline’s. The CEO asks whether to match fares or double down on weekday business shuttles.
Clarify. Success is next-year contribution, not load factor. The LCC will not exit. We can cut frequencies.
Do not tour all three. The customer split is the first test. Weekday 6am–8pm traffic is 58% of seats but 71% of contribution, because those passengers pay $148 average fare versus $79 on weekend leisure. Matching the LCC on all fares would drag the business fare toward $121 (18% off $148) and the leisure fare toward $65.
Do the company math before you “compete.” Variable cost is $62 per seat flown (fuel, crew, airport). Contribution today: business $86 per seat, leisure $17. After a full match: business $59, leisure $3. Leisure almost vanishes. Business still contributes, but you gave away $27 per business seat to chase a segment that was never going to love you.
Competitor branch, last, and only as a response test: the LCC’s cost per seat is about $48. They can sit at $79 forever. Skyline cannot win a price war on leisure. They can win if business passengers will pay for last-seat availability and a 6:10am bank that the LCC does not operate.
Recommendation. Do not match fares network-wide. Hold business fares, cut two weekend frequencies that fill below 60%, and add a 6:10am bank on the three densest city pairs. Risk: corporates put Skyline on a dual-source contract and still book the LCC. Next step: pull the last 12 weeks of corporate contract mix by route.
The so-what is a number: protecting $86 contribution on the seats that matter beats buying load factor at $3 contribution.
When this lens is the wrong tool
Skip it when the prompt is a P&L diagnosis. If profit fell $14m and you do not yet know whether volume, price, mix, or cost moved, you need a profitability bridge, not a tour of “the company.” Company/customer/competitor will not isolate the driver.
It is also the wrong first tree for a pure operations prompt (turn time, yield, utilization) and for a portfolio prompt (which of four businesses to fund). Those belong to an ops cost tree or a growth-share allocation, not a 3C speech.
Use it when the decision is competitive position: launch, reposition, respond to an entrant, or choose a segment to serve.
The mistake that is unique to this lens
The unique fail is symmetric touring: three minutes on company, three on customer, three on competitor, no hypothesis. Interviewers hear a textbook. The fix is to bet. In the Skyline case the bet is “this is a customer-mix problem, not a cost problem.” You prove or kill that bet with one fare-by-segment table. If the bet dies, you rotate. If it lives, you never needed a full competitor biography.
A close second is stuffing “market trends” into company. Trends are either customer (demand) or competitor (capacity additions). Company is what this airline can do on Monday.
How to say it in the room
- Restate the decision and the metric.
- Name the first branch you will test and why the other two can wait.
- Ask for the one exhibit that would kill the bet.
- Convert the winner into a recommendation with a risk.
Practice that sequence on a live prompt, not by reciting labels.
Build a 3-branch tree under time
Run a structure drill where the prompt is hidden from this article. If you say the framework name, you probably over-indexed.