In short: A price pack ladder is a set of deliberate gaps, and the number that decides whether a shopper climbs it is the incremental price per unit between consecutive rungs rather than the price per unit of each pack read on its own. A middle rung priced too close to the one below it pulls volume from both directions at once, and the margin arithmetic on that trade is usually negative even though the unit share looks healthy. Packaging cost falls with size much more slowly than price per unit does, so most of the gradient in a working ladder is a segmentation decision rather than a cost recovery. The ladder you publish is a list price ladder, and a fortnight of promotion can invert it at the shelf.
The category has nine active pack sizes and four of them matter. The 500ml and the 750ml sit forty cents apart on shelf, the 750 takes about seventy percent of the volume, and nobody in the room can say what the 500 is for any more. Somebody proposes deleting it. Somebody else points out it is the only pack the convenience channel will list. The meeting ends without a decision, which is the decision.
That is the ordinary state of a portfolio that grew by addition. Each pack arrived to answer a question that was live at the time: a customer asked for it, a competitor launched one, a promotion needed a unique code. What nobody did was design the set as a set.
The ladder is a set of gaps
The useful object is the gap between rungs rather than the rungs themselves. A shopper standing at the shelf is not evaluating a 500ml bottle against an abstract sense of what water should cost. They are comparing it with the pack beside it, and the comparison they run is what the extra volume costs.
So the design variables are the spacing of the sizes, the price gradient across them, which channel sees which rung, and the format of each. Number of rungs comes last, after you know what each one is separating.
Each rung should have a buyer and an occasion attached to it before it has a price. A single serve pack is bought on impulse, consumed immediately, and competes with everything else in the chiller rather than with your own larger packs. A mid size pack is a household purchase with a weekly cadence. A bulk pack is a stocking decision with storage implications, and the shopper making it has done arithmetic. Those are genuinely different buyers with different sensitivity, and the ladder exists to let you charge them differently without having to identify them.
The number that decides trade-up
Take a three rung ladder. A 330ml pack at 1.20, a 500ml at 1.55, a one litre at 2.60.
Price per 100ml runs 36.4 cents, 31.0 cents, 26.0 cents. Monotone decreasing, which is the condition every ladder is checked against and the one that tells you least.
Now compute the incremental price per unit, which is what the step actually costs. Going from 330ml to 500ml buys 170ml for 35 cents, so the increment is 20.6 cents per 100ml. Going from 500ml to one litre buys 500ml for 1.05, so the increment is 21.0 cents per 100ml. The two steps are priced almost identically, and a shopper who trades up once faces the same proposition trading up again.
Change one number. Drop the 500ml to 1.40, which looks like a modest competitive response and passes every price per unit check, since 28 cents still sits between 36.4 and 26.0. The increments now read very differently. The step from 330ml costs 20 cents for 170ml, or 11.8 cents per 100ml, which is half the previous rate. The step to one litre costs 1.20 for 500ml, or 24 cents per 100ml, which is double it.
You have built a ladder with a cheap first step and an expensive second one. The 500 will take volume from the 330 because trading up is nearly free, and it will hold volume that would otherwise have gone to the litre because trading up again looks expensive. Both movements show up as growth on the 500, and the pack looks like a success.
What the middle rung costs when it absorbs both directions
Put margins on it. Say the contribution per pack is 45 cents on the 330ml, 52 cents on the 500ml at the original 1.55, and 78 cents on the litre.
At 1.40 the 500ml contributes 37 cents. Suppose the move pulls 100 units a week up from the 330 and 60 units a week down from the litre. The trade up is worth 100 times minus 8 cents, since you moved buyers from a 45 cent pack to a 37 cent one. The trade down is worth 60 times minus 41 cents. Total effect is minus 33.20 a week per store on a move that raised the 500ml unit share by about a third.
The direction of the arithmetic survives quite a lot of variation in the assumptions. What it depends on is the shape: a middle rung with a lower absolute contribution than the rung above it, priced to be easy to reach from below, sitting in a portfolio where a meaningful share of volume was already at the top. That configuration is common, and it is created by exactly the sort of tactical price move that never gets reviewed as an architecture change.
The behavioural literature says the middle rung is doing more work than its price alone would suggest. Simonson and Tversky (1992), in the Journal of Marketing Research, showed that adding an extreme option shifts share toward the middle of a set, an effect they called extremeness aversion. Huber, Payne and Puto (1982), in the Journal of Consumer Research, showed that an option which is dominated by one alternative and not by another lifts the share of the option that dominates it. Both results say the same thing for a portfolio manager: the position of a pack in the set changes its share independently of its own attributes, so a rung can earn its place by shaping choice even when it sells modestly.
That cuts against the instinct to delete slow packs. A pack doing four percent of volume that anchors the top of the ladder may be paying for itself in the mix it produces below. The way to find out is to remove it from a subset of stores and watch what happens to the packs either side, which is a test worth running before a delisting decision rather than after.
Estimating how much volume each rung actually loses to a price change is a separate exercise with its own method, covered in C1.
Cost does not explain the gradient
The usual defence of a declining price per unit is that larger packs cost less to make per unit. Check the size of that effect before it carries an argument.
For a rigid plastic bottle, the components are resin, closure, label, filling time and secondary packaging. Resin scales with surface area rather than volume, which is the reason the effect is weaker than intuition suggests: a container of double the volume has roughly 1.6 times the surface area, so it uses about 1.6 times the material rather than twice. Closure and label are nearly fixed per unit. Filling time per container is close to fixed for a given line speed.
Put numbers on it. Suppose packaging plus filling comes to 11 cents for the 330ml and 22 cents for the litre. Per 100ml that is 3.3 cents against 2.2 cents, a gap of 1.1 cents. The price gap per 100ml between those two packs is 10.4 cents. Cost explains about a tenth of it.
The other nine tenths is the segmentation decision, and it is the part worth defending explicitly. A ladder that recovers cost differences and nothing else has given away the entire willingness to pay difference between an impulse buyer and a stock up buyer, which is the reason for having a ladder.
Rungs have to be earned in the plant and on the shelf
Every rung is a change part, a run length, a case configuration, a pallet pattern and a slot in the item master. Those costs are real and they fall on people who were not in the pricing meeting.
The plant constraint is minimum economic run length. A pack that sells 40,000 units a year on a line with a four hour changeover and a minimum run of 60,000 will be made once every eighteen months, which means it carries a year and a half of coverage or it goes out of stock. Changeover economics have their own treatment in EE1, but the pricing implication is direct: a rung whose annual volume falls below the minimum run is a rung you are financing with working capital.
The shelf constraint is facings. A retailer allocating four facings to your brand will not stock five packs, so a ladder with more rungs than the shelf can hold becomes a different ladder in every account depending on which packs the category manager chose. At that point you have lost control of the architecture, and the gaps that the shopper actually sees were designed by somebody else.
Three or four rungs per channel is usually the practical ceiling, with different rungs in different channels rather than more rungs everywhere.
Choosing what each channel gets
Channel exclusivity is the cheapest instrument in the design, since it lets the same brand carry different price points without either one undercutting the other in front of the same shopper.
The standard pattern gives the impulse channel a single serve pack it holds alone, the grocery channel the mid range, and the club or wholesale channel a multipack or bulk format that nobody else lists. Each rung then competes with the alternatives available in that location rather than with your own portfolio.
Two conditions have to hold for it to work. The formats need to be visibly different, since a shopper who can see that the club multipack contains the same bottle as the grocery single will read the grocery price as a penalty. And the packs need to stay in their channels, which is a terms and enforcement question rather than a pricing one. Where the same buyer can reach two channels, the structure needs different protection, and that is the subject of PP3.
Where this stops
A ladder is a static structure and the shelf is not. The prices you designed are list prices, and what the shopper compares is the shelf price after whatever promotion is running that week. Two rungs promoting on different cycles can invert for a fortnight at a time, and a shopper who learns that the litre goes to 2.10 every third week has been taught to wait rather than to trade up. Marn and Rosiello (1992), in the Harvard Business Review, made the general version of this point with the pocket price waterfall: the gap between list and realised price is wide, it varies by customer, and it is where the pricing decision actually resolves. Checking the ladder on realised prices by pack over a rolling quarter, rather than on the price list, is the only version of the check that means anything.
The architecture also assumes a portfolio large enough to support separation. In a category doing modest volume, three rungs may leave every rung below the minimum economic run, and the honest answer is two packs designed well. The technique scales down by removing rungs rather than by narrowing gaps.
And a ladder cannot fix a value proposition. If the entry pack is priced above the private label mid pack, no amount of gradient design moves a shopper who has already decided the brand is not worth the premium.
Take your current portfolio, list the packs in size order with their shelf prices from the last four weeks of scan data, and compute the incremental price per 100ml between each consecutive pair. Any step that comes in below half the step beside it is a rung absorbing its neighbour, and that is where the next price decision belongs.