In short: Tier margins compound, so twenty, eight and twenty-five points stacked on a hundred produce a shelf price of 181 rather than 153, and the business case has to be built backwards from the shelf. The percentage pays for five different functions, being financing, physical handling, coverage, risk and market access, of which only financing and risk scale with your price. So a general price increase pays the distributor more for work whose cost did not move, and a volume drop underpays them against a fixed cost base. The waterfall breaks because each deduction is signed off by a different function and nobody has printed the whole thing on one page.
You are pricing a market. The distributor wants twenty-two points. Below them a wholesale layer takes eight. The retailer expects twenty-five on the shelf price. Local VAT is twenty. Someone adds twenty-two, eight and twenty-five, gets fifty-five, takes fifty-five percent off the target shelf price, and the ex-factory number that comes out looks like a business.
Eighteen months later the shelf price sits fourteen percent above where the plan put it, volume is behind, and the distributor is asking for two more points because their costs have gone up. Nobody in the room can say what the current points are buying, so the conversation runs on assertion.
Margins compound, and adding the points understates the shelf
Take a hundred ex-factory and work upward, with each tier taking its margin on its own selling price, which is how they all quote.
The distributor at twenty percent sells at 100 divided by 0.80, or 125. The wholesaler at eight percent sells at 125 divided by 0.92, or 135.87. The retailer at twenty-five percent on retail sells at 135.87 divided by 0.75, or 181.16. Add twenty percent VAT and the shelf price is 217.39.
Fifty-three points of stated margin produced an eighty-one percent uplift before tax. The gap between fifty-three and eighty-one is compounding, and it is twenty-eight points of ex-factory price that never appeared in anybody's calculation.
The related trap is the word itself. A distributor who says twenty-five percent may mean twenty-five on cost, which is a sale at 125, or twenty-five on selling price, which is a sale at 133.33. Carry both versions up through the two tiers above and the shelf price differs by 6.7 percent on the same agreement. That gets discovered when a regional manager photographs a shelf and the price does not match the model.
Work backwards from the shelf, because that is the only fixed number
The shelf price is set by the market, meaning by the adjacent pack and the nearest competitor. Everything between the shelf and your bank account is a deduction, so your ex-factory price is a residual and needs calculating as one.
Target retail of 199 including twenty percent VAT is 165.83 net of tax. The retailer at twenty-five percent leaves 124.37. The wholesaler at eight percent leaves 114.42. The distributor at twenty percent leaves 91.54, and that is the number the business case has to survive on.
A case built on a hundred, because someone subtracted the stacked percentages from the retail price, is carrying a gross margin assumption 8.5 percent of revenue too high before a single case has moved.
The same arithmetic gives you the number to bring to a margin negotiation. Moving the distributor from twenty to twenty-two points takes 114.42 times 0.78, or 89.25, against 91.54. Two points of theirs costs 2.5 percent of your revenue, because the concession comes out of the smallest number in the chain. Saying that out loud changes how the request gets framed.
Five functions inside one percentage
The margin is a payment for work, and the percentage indexes that payment to value. Five distinct pieces of work sit underneath it, each with its own cost driver.
Financing. They buy your stock, hold it, and extend credit to the trade below them. On the numbers above, with forty-five days of stock, sixty days credit given and thirty days payment to you, the cash cycle is seventy-five days. At a local cost of debt of fourteen percent on goods costing 91.54, that is 2.63 per case, or 2.3 points of their 114.42 selling price. At thirty percent, which is ordinary in several markets, it is 5.64 per case and 4.9 points. W1 covers the cash conversion cycle itself and DD5 covers what payment terms cost the party granting them.
Physical handling. Warehousing, secondary transport, refrigeration where relevant. Driven by cases, drops and distance.
Coverage. Field sales, order capture, merchandising, and collecting money from the trade. Driven by outlets served and visit frequency.
Risk. Bad debt below them, breakage, expiry, and obsolescence on whatever you launched and then discontinued. Driven by the trade's payment behaviour and your shelf life.
Market access. Import licence, product registration, importer of record status, regulatory holding, and in some markets a local ownership requirement. This is the least discussed of the five and usually the reason the distributor has an outside option in the negotiation.
A percentage pays for value while the work is priced per case
Only financing and risk in that list scale with your price. Hold the distributor at twenty percent and compare two items. On a hundred ex-factory they earn twenty-five per case at a 125 selling price. On a premium item at three hundred ex-factory they earn seventy-five per case at a 375 selling price. The pallet space, the pick, the drop, the invoice line and the collection call are identical.
Financing and risk are genuinely three times larger on the premium item, so part of that gap is earned. Physical handling and coverage are not, and on most portfolios they are the larger share of the real cost. Kaplan and Anderson made the general version of this argument in Harvard Business Review in 2004 with time-driven activity based costing, which prices an activity by the time it consumes rather than by the value passing through it. W4 works the customer-level version of that costing.
Two consequences follow, and they point in opposite directions.
A general price increase quietly transfers money down the chain. Twenty points on a hundred is twenty-five per case. Twenty points on 115 is 28.75. The extra 3.75 buys no additional work, and on a portfolio taking two consecutive high single digit increases the cumulative transfer is material enough to show up in your own margin bridge as an unexplained mix effect.
A volume decline underpays them. The same percentage on twenty percent fewer cases delivers twenty percent less cash against a warehouse lease and a sales force that are fixed inside a year. They come back for points, and the request is structurally justified even when the negotiation makes it sound opportunistic.
Where the structure breaks under pressure
Nobody owns the whole waterfall. List price sits with pricing, invoice discounts with the sales lead, off-invoice support and listing fees with the account team, logistics allowances with supply chain, settlement discount with finance, returns with quality, and the rebate accrual with accounting. Every line is defensible on its own and signed by someone with the authority to sign it. The sum has often never been printed on one page. Marn and Rosiello set out the price waterfall and the pocket price in Harvard Business Review in 1992, and their sensitivity figure is still the reason to do the work: for the average company in the S&P 1500 sample they used, a one percent improvement in realised price with volume held raised operating profit by 11.1 percent, a larger response than to an equivalent move in volume or unit cost.
Volume brackets get traded rather than earned. Any tier whose buying price steps down at a threshold will buy to the threshold. What arrives in your history is a period end spike and a period start hole, which the forecasting system learns as seasonality that does not exist in consumption. PP2 covers the rebate mechanics and the accrual problem underneath that behaviour.
Announced price increases arrive as stock gain. Publish a list price rise with a date and every tier below buys ahead of it. Those cases ship and get invoiced without anyone consuming them, and the two months that follow look like a collapse that gets investigated as a commercial failure.
Price differentials pull product sideways. When the net landed cost into two adjacent markets differs by more than the per-case cost of moving a container between them, product moves, and the test is arithmetic rather than a matter of judgement. What that flow does to a market signal is U5's subject.
Credit runs out before demand does. In markets where the distributor finances the trade below them, order size stops tracking consumption and starts tracking their receivables position, which U9 and Z10 both work through.
Rebuilding the payment so it survives an argument
Pay each function on its own driver.
A per-case or per-drop logistics allowance for the physical work, indexed to a published fuel or wage measure rather than to your price list. A coverage payment tied to something countable, being active outlets served or visits completed, verified from their order data rather than from their claim. A financing element written as days multiplied by an agreed rate, so that when rates move the conversation already has an answer. A performance element paid against things you can check, being sell-out reporting delivered on time, stock accuracy, and availability measured in a sample of outlets. Then a residual percentage carrying risk and market access.
Most of the benefit is available without reopening the contract. Keep the headline percentage and write down what each point is buying. The next request for two more points then has to name which function got more expensive, and every one of the five has an answer that can be tested. If it is financing, a rate moved and you can look it up. If it is coverage, the outlet count moved and they can show it. If it is physical, the drop size fell and their own delivery data says so.
The one structural change worth making even inside a percentage deal is to cap the value scaling on the physical element, because a premium item does not cost more to move.
Where this stops
You can only price these functions if you know the costs, and the costs belong to the distributor. Most principals negotiate against numbers they cannot audit and will never be shown. What you can build is your own bottom-up estimate from things you can observe: cases shipped, drop counts, outlets on their reported customer list, days of stock implied by their orders against whatever sell-out they report. It will be wrong in the details and roughly right on the order of magnitude, which is enough to move the discussion from assertion to arithmetic.
Where the distributor holds the import licence, the product registrations, or customer relationships that would take three years to rebuild, margin is set by bargaining power and the functional breakdown becomes an argument rather than a lever. Know which position you are in before opening the conversation, because the same analysis presented from the weak side is an invitation to renegotiate everything at once.
A five-component structure also decays. Somebody has to reconcile the coverage claim against the outlet data every quarter and check that the financing element still matches the rate it was written against. Where nobody owns that reconciliation, the structure collapses back into a single percentage inside two years, and it is usually a higher percentage than the one you started with.
Take one item in your largest distributor market, build the waterfall from list price down to what actually reaches your bank counting every deduction wherever it is booked, and set the total next to your own per-case estimate of the physical work being done for it.