What this pack runs.
Nothing in any reporting requirement asks a company to say which of its customers is also building what it sells them, so the relationship appears where every large customer appears: a revenue line, a share of the total, a name in the concentration note. Samsung Display commands the largest share of Apple's OLED orders while Samsung Electronics sells phones against the iPhone. Amazon's parcel volume passed UPS and FedEx by 2023, and in May 2026 it opened the network to outside firms and both their shares fell close to ten percent. Neither relationship was concealed; neither was ever stated as a period with an end. This pack turns it into one multiplication. Revenue times the margin on their business gives what the relationship earns a year; multiplied by the years they would need to replace you, it gives the whole of what it is worth, because after that they no longer need you. Set against the same annual profit at the multiple you trade on — which prices a stream as though it had no end — the replacement time over the multiple says what share of that price the years actually support. It stops on five conditions: no revenue from the customer, no total revenue, no margin on their business, no replacement time, and nothing said about what you are as a share of their own costs.
The casebooks these cases come from.
The pack reads nine sourced decisions through the framework. They are drawn from these casebooks, each one a set of verified records on a single decision type.
| Casebook | Verified records |
|---|---|
| Vertical-Integration Bet | 33 |
A verified record has passed a mechanical claim-versus-source check, a second model's read, and human approval. The count is what exists today, not a target.
Loading the rest of the sample…
16 files for this one decision.
The extract holds one case of the nine, and the number without the model that produces it. $499, one-time.
See the full pack →