Every product manager building a cheaper tier assumes the job is subtraction: take the full product, remove what costs money, sell the remainder for less. The economics run the other direction more often than anyone expects. In 1990 IBM shipped the LaserPrinter E, a lower-priced sibling of its standard LaserPrinter that printed noticeably fewer pages per minute. The two machines were essentially the same printer. The slower one was slower because IBM had added chips whose job was to make it wait — the company spent money to degrade its own product.

Economists call this damaged goods, and it is second-degree price discrimination in work clothes. A firm facing customers with very different willingness to pay has three options: price high and forfeit the budget segment, price low and hand a discount to everyone who would have paid full freight, or build a fence. The fence has to sit on something the high-value buyer genuinely will not tolerate. Intel ran the same play with the 486SX, a 486DX sold cheaper with its floating-point unit out of action — engineers who needed the math coprocessor paid up, and everyone else got a real machine at a real discount. The cost of building the crippled version is the price of that fence, and it is usually far below the margin it protects.

The failure mode is not building a cheap tier — it is drawing the fence on a dimension your best customers do not care about. Strip a feature enterprise buyers were indifferent to and you have segmented nothing; you have published a lower price and invited your own base to walk down to it. The diagnostic question is not what this tier costs to make, but which specific buyer would refuse this version and why. If you cannot name that buyer and that reason, the tier is a discount pretending to be a strategy.