A salesperson knocks ten percent off to close a deal before quarter-end. It feels like a small concession, a nudge to get the customer over the line. But on a product carrying a thirty-percent margin, that ten points off the price doesn't cost you ten percent of the profit. It costs you a third of it. The discount is the cheapest thing in the room to give and the most expensive thing to give back.

The quick version

  • Discounting governance is the set of rules, visibility and approvals that decide who can give away price, how much, and on what evidence, so margin isn't leaked one "reasonable" exception at a time.
  • The list price is rarely the price you keep. Layer after layer of discounts, rebates, terms and freight pull it down to the pocket price, what actually lands in your account.
  • Because margin is a thin slice of the price, small discounts hit profit hard: for the average large company, a 1% move in realised price swings operating profit by roughly 8%.
  • The fix isn't "stop discounting." It's making the cost of each discount visible at the moment of the decision, and matching approval authority to the size of the giveaway.

The idea in depth

Start with a fact most pricing problems trace back to: the number on the price list is fiction. Between that list price and the cash you finally bank sits a long staircase of deductions, negotiated discounts, volume rebates, early-payment terms, promotional allowances, freight, the cost of carrying the receivable. McKinsey consultants Michael Marn and Robert Rosiello named the staircase the pocket-price waterfall in their 1992 Harvard Business Review article "Managing Price, Gaining Profit". Their point was uncomfortable and durable: leaders manage the list price closely and the bottom of the waterfall, the pocket price, barely at all, even though the pocket price is the only one that pays the bills.

The same authors found that pocket prices for a single product, sold to different customers, routinely spread across a band of more than 60%. Identical widget, wildly different realised price, and usually not because the cheap end of the band represents the strategic accounts. It's noise: who negotiated hardest, who got a distracted rep, which deal landed in a soft quarter. Before you touch anyone's discount authority, build your own waterfall. Take one real product, lay out every deduction from list to pocket, and rank your customers by pocket price. The leaks announce themselves, and so does the band.

flowchart TD
  A(["List price £100"]) --> B(["Negotiated discount −12"])
  B --> C(["Volume rebate −6"])
  C --> D(["Invoice price £82"])
  D --> E(["Payment terms −3"])
  E --> F(["Promo / co-op −4"])
  F --> G(["Freight & handling −3"])
  G --> H(["Pocket price £72"])
  H --> I(["Pocket margin: minus COGS & cost-to-serve"])
The pocket-price waterfall: list price erodes, step by step, to the cash you actually keep. Illustrative figures. Leaders Loop

Why a small discount is never small

The reason discounting deserves governance and not just goodwill is arithmetic. A discount comes entirely out of margin, and margin is a thin slice of price. In the same body of pricing research, Marn and two colleagues showed in the McKinsey Quarterly article "The Power of Pricing" (2003) that for a typical S&P 1500 company, a 1% improvement in price, holding volume steady, lifts operating profit by about 8%. The lever runs both ways. Every point you give away at the negotiating table pulls roughly eight times its weight out of the bottom line.

A discount comes entirely out of margin, and margin is a thin slice of price. That is why one point of price can be eight points of profit.

The practical answer is to put that ratio in front of the person who signs the discount, in their own product's numbers, at the moment they decide. Most discount approval flows show the deal size and the discount percentage. Almost none show the line that matters: this concession costs us £X of margin, equivalent to selling Y more units at full price to break even. When the giveaway is denominated in profit rather than in a percentage off a list nobody believes, the easy "sure, knock off ten percent" gets harder, which is the point.

An honest limitation. The 1%-equals-8% figure is a headline, not a law, and it has been fairly criticised, pricing academic John Dawes points out that it assumes you can raise price with no loss of volume, which real demand rarely allows. Raise prices and some customers walk; the true profit effect depends on how price-sensitive your market is. Read the 8% as the size of the prize when a discount was genuinely unnecessary, the deal would have closed anyway, not as a promise that any price rise prints money. Governance is about cutting the unnecessary giveaways, not refusing to ever trade price for volume.

Governing the giveaway: rules, not heroics

Good deal-economics governance rests on a plain observation: people give away more when the cost is invisible and nobody is checking. A discount feels free to the rep, it isn't their margin, and the pressure only builds as the quarter closes. So the remedy has to be structural rather than motivational. Three controls do most of the work.

First, a discount approval ladder: the deeper the discount, the higher the sign-off. A rep can approve to a threshold; beyond it the deal escalates to a manager, then to a deal desk or finance. This is the same logic as any spending-authority matrix, you wouldn't let anyone wire any sum without a second signature, and a discount is a wire transfer out of margin. Second, a floor price below which no one sells without an executive exception, so the cheap end of the pocket-price band has a hard stop. Third, a deal desk for large or non-standard deals, a small cross-functional review (sales, finance, product) that prices the concession, checks it against the floor, and records why it was granted. The work, then, is to write the ladder down, set the thresholds off your real margin maths, and route approvals through it automatically rather than through a hallway conversation.

flowchart TD
  A(["Rep proposes a discount"]) --> B{"Within rep
authority?"}
  B -->|Yes| C(["Auto-approve · log it"])
  B -->|No| D{"Above the
floor price?"}
  D -->|Yes| E(["Manager / deal-desk review"])
  D -->|No| F(["Executive exception only"])
  E --> G(["Decision recorded with rationale"])
  F --> G
  C --> H(["Feeds the pocket-price waterfall"])
  G --> H
A discount approval ladder: authority scales with the size of the giveaway, and every exception leaves a record. Leaders Loop

The deeper fix is to design the product and packaging so the rep has less to negotiate away, what Madhavan Ramanujam and Georg Tacke call building the price into the product rather than bolting it on after. Governance catches leaks; good pricing strategy and pricing mechanisms (tiers, good-better-best, fenced concessions) stop the boat taking on water at all. And a caution: clamp authority too hard and you slow good deals and breed workarounds, reps quote freight as "free," reclassify discounts as "rebates," sandbag forecasts. Aim for a ladder loose enough that standard deals fly through and tight enough that the unusual ones get a second pair of eyes.

A worked example

Illustrative figures, the numbers below are made up to show the mechanics, not drawn from a real company.

A mid-market software firm sells a £100,000 annual contract at a 30% gross margin, £30,000 of margin per deal. A rep, racing a quarter-end clock, offers 15% off to land it. On the spreadsheet it reads as a £15,000 discount on a £100,000 deal: painful but survivable. Run it through the pocket-price lens and it looks different. That £15,000 comes straight out of the £30,000 of margin, half the profit on the deal is gone. To earn that £15,000 back at the new discounted margin, the team would need to close roughly another full deal of the same size.

Now add governance. The firm sets rep authority at 10%, with anything deeper routed to a deal desk and a hard floor at 20%. The 15% ask now triggers a thirty-minute review. The desk asks the one question the hallway never does: what evidence says we'd lose this deal at 10%? Often there is none, the customer never countered, the rep was buying their own comfort. The deal closes at 10%: £10,000 given away instead of £15,000, £5,000 of margin saved on a single contract, and a logged rationale that teaches the next negotiation. Multiply that across a year of deals and you have recovered a point or two of realised price, which, per the 8% multiplier, is a disproportionate amount of operating profit, with no new customers and no new product.

Frequently asked questions

Isn't discounting just part of selling?

Yes, the goal isn't zero discounts. A well-placed discount wins a strategic account, clears ageing stock, or rewards genuine volume. Governance doesn't ban the tool; it makes sure the discount buys something. The test for any concession is simple: what did we get in return, a longer term, a bigger commitment, a reference, a faster close, and would the deal have happened without it? Discounts that buy nothing are the ones the ladder is built to catch.

What's the difference between list price, invoice price and pocket price?

List price is the published number. Invoice price is what survives the on-invoice discounts and appears on the bill. Pocket price is what's left after the off-invoice deductions too, rebates, payment terms, promotional allowances, freight, the cash that actually reaches your pocket. Most companies manage the top of that staircase and ignore the bottom, which is exactly where Marn and Rosiello found the margin hiding.

Won't an approval ladder just slow our deals down?

Only if you set the thresholds wrong. Tune the ladder so the large majority of standard deals fall inside rep authority and clear instantly; only the deep, unusual discounts escalate. If most deals are tripping the ladder, your list price or your standard discount is the real problem, fix the price, not the approvals. Speed and control aren't opposites here; a clear rule is faster than an ad-hoc plea to a manager.

How do I start if we have no governance at all?

Build one pocket-price waterfall for a single important product and rank customers by pocket price. That one chart usually ends the debate, leaders can see the band and the leaks for themselves. Then set a floor price and a single approval threshold off your real margin. You don't need a deal-desk function on day one; you need visibility and one rule, then iterate.

How does this connect to our unit economics?

Directly. The discount you grant is the gap between your list price and your realised price, and realised price is the top line of unit economics, it flows straight into contribution margin, payback and lifetime value. Sloppy discounting doesn't just dent one deal; it quietly degrades every model built on the assumption that you sell at list.

Related in the Toolkit

Where to go next