The Enterprise Discount

What this pack runs.

A price cut to sign one large customer is a normal commercial act and frequently the only way to sign the account: a reference, a volume, a standard, a foothold, bought with a price no other customer gets. What makes it a decision later is that the two halves age differently. The gap against list is in the contract and in every conversation about the account, so it is visible every quarter, while the customers the logo brought in are cited by every rep and counted by none, so they stay a claim. By the time the account reaches an agenda it has usually been reclassified in the meeting from a trophy into a mistake, and the conversation becomes how to reprice. That is not the comparison. The list was never available, so the gap measures the distance to a deal that did not exist. This pack turns it into a subtraction, a multiplication, and a division. What the customer pays a year, less what serving them costs, over the term and discounted, plus what was spent once, is what the concession costs, and it is usually a fraction of the discount. Against it: the customers that would not have signed without the reference, at what one of them is worth, which is what such a customer pays a year, at the margin you earn on it, for the years it stays. The first divided by the second is the break-even pull-through, how many customers the logo must win before the concession is repaid, and its distance from a count read off the win notes is what a board can watch. The model sweeps the count from none to fifteen, because it is the one figure no meeting will settle. It stops on six conditions: no deal price, no list price for the same volume, no cost of serving this customer on its own, no term, no value for a customer won on the logo, and no count of customers the logo has brought in.

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.

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.