The short answer
Pricing leakage is the gap between the price your business meant to charge and the price it actually collected. You find it by comparing invoice lines against list or target price and asking what explains the gap: order size, customer, product, or the rep who wrote the quote. When discount depth tracks the rep rather than the customer, you have leakage, and it can be sized from twelve to twenty-four months of records you already keep.
Why it hides
Most sales teams are paid on revenue, not margin. Under that plan a discount is a rational move for each rep: it closes the order faster and costs them almost nothing. Added up across thousands of invoice lines, it becomes one of the most expensive habits a business has, and because no single discount looks unreasonable, nobody flags it.
The arithmetic is why it matters. At a 4 percent operating margin, a 1 percent price increase adds about 25 percent to operating profit, provided volume holds. McKinsey has put the average effect across large companies at roughly 8.7 percent, an estimate it restated in 20265. Few other levers move profit that much for so little capital.
Where to look first
You need four fields you almost certainly already have: invoice line, list or target price, customer, and the rep or estimator who owned the quote. With those, three questions do most of the work.
- Does discount depth follow order size? If small repeat orders are discounted as heavily as large new ones, the discount is not buying volume.
- Does it follow the customer? Some accounts genuinely earn better pricing. If the same customer pays different prices depending on who quoted them, something else is going on.
- Does it follow the rep? This is the strongest signal. When discounting tracks the person rather than the deal, it is a habit, not a strategy.
Write the hunch down before you open the data
Before anyone runs a query, ask your sales lead to write down what they believe: for example, that repeat orders average within two points of list. Put a number and a range on it. This sounds like ceremony, but it protects you from the most common analytical mistake, which is deciding after the fact that the result was obvious all along9.
Size it in three layers, not one
A single leakage number invites an argument. Three numbers invite a decision.
Observed
Addressable
Capturable
Specimen · illustrative figures
On $4.1M of repeat-order revenue, orders under 50 units ran 9 to 13 percent below list, and discount depth tracked the rep rather than order size. Observed $410K–$520K, addressable $300K–$380K, capturable $180K–$260K a year.
Prove it before you build anything
A pattern in historical data is a candidate, not a result. Check that the gap you found in one stretch of your history also appears in a stretch you did not search. Then test it live: randomize by customer, give half the reps a recommended price beside the quote tool, and agree in advance what result would end the test. If margin does not move by the amount you agreed, you stop, and it cost you six weeks rather than a software rollout.
What usually fixes it
Rarely a new policy, and rarely a new commission plan on day one. The mechanism that works most often is a pricing engine that sits beside the existing quote tool and shows the rep a second number drawn from your own win history. The rep still decides. Most of the time, the second number is enough.
Sources
- 5McKinsey & Company, “B2B pricing: Navigating the next phase of the AI revolution,” April 2026; originally Marn, Roegner & Zawada, “The Power of Pricing,” McKinsey Quarterly, 2003.
- 9Fischhoff, B. (1975). Hindsight ≠ foresight. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299.
Market figures describe the landscape and are not Modven results.