At Intuit I managed over $200M a year in discount allocation across bundling and offer strategies. At that size a discount budget isn't a line item, it's a system, and small changes in how you allocate it move very large numbers.
The first principle is that not all discount dollars do equal work. A dollar that converts a customer who would have left is worth far more than a dollar handed to a customer who'd have paid full price anyway. Most of the work was telling those two cases apart and shifting dollars toward the first.
That meant segmenting hard and targeting offers by behavior, not spraying a blanket discount across everyone. Blanket discounts feel fair and they quietly burn the budget on people who never needed the nudge.
The second principle is that you measure incrementality, not redemption. A high redemption rate just means people took free money. The question is whether the offer changed what they did. Those are different numbers and only one of them matters.
The work delivered $50M-plus in annualized revenue across these strategies. The lever wasn't spending more. It was spending the same budget where it actually changed an outcome.