Inflation and CPG margins: how rising costs squeeze a consumer brand, and what to measure

Meet Dana. She runs a made-up hand soap brand.

Monday morning, she opens the sales report. Sales are up 9%. Orders are up too.

At 10:15, her accountant calls.

“Cash is down again. Third month in a row.”

Dana pulls up the bills. Nothing broke. No one made a mistake she can point to.

The factory price crept up. So did freight. So did the warehouse fees. So did her ads.

Every cost went up a little, in different places, at different times.

Inflation squeezes a consumer brand’s margin from many small directions at once. Most of it hides until you measure cost per unit and margin per order by channel, every month.

CPG stands for consumer packaged goods: things people buy often and use up, like soap, snacks or toothpaste. These brands live on thin margins per unit and high volume. A few cents per unit, times a lot of units, is a lot of money.

This post walks through where the squeeze comes from, why prices lag behind it, how mix can hide it, and what to measure. It follows Dana’s brand the whole way.

Where the squeeze comes from

Margin is what is left from a sale after costs. When costs rise and price stays put, margin shrinks. The trouble is that the costs rise in different places, so no single line looks alarming.

The wider economy shows the same pattern. U.S. consumer prices rose 3.7% in the 12 months to September 2026 (BLS Consumer Price Index, September 2026). The Federal Reserve aims for 2% over the longer run (Federal Reserve FAQ). So prices are rising faster than the goal.

Producer prices rose faster still. The prices U.S. producers get for their goods and services rose 6.3% over the same 12 months (BLS Producer Price Index, September 2026). Energy drove most of September’s jump in goods prices. Energy feeds into freight, plastic and packaging.

Here is the catch. Your shelf price lives in the consumer world. Many of your costs come from the producer side.

Some of these are in product cost. Some come later, when you sell. Landed cost is the full cost to get one unit into your warehouse, ready to sell. It includes the product, freight in, duties and handling. Our guide to COGS and landed cost shows how to build it.

The costs after landing, like fees, shipping and ads, are where channels differ. That is why one blended margin misleads.

A worked example: one soap, two channels

Dana’s brand sells a three-pack of foam hand soap for $24. It sells on Amazon and on its own store. Here is one unit last year and this year. All brand numbers below are illustrative. The Amazon fee rate is Amazon’s published rate.

Landed cost per unit

Each line rose by a different amount. Freight rose the most in percent, but the factory price added the most in dollars.

Contribution margin per unit, by channel

Contribution margin is what a sale leaves after every cost that exists only because that sale happened. For the full method, see contribution margin for Amazon and Shopify.

The referral fee is the cut Amazon takes from each sale. For Beauty, Health and Personal Care, Amazon charges 15% on items priced over $10, and 8% at $10 or less, as of October 2026 (Amazon selling fees). The fee is a percent of price, so it did not change here. It will change if the price does.

No line rose by more than half. Yet Amazon margin per unit fell 43%. The store went from thin to losing money on every order.

That is the shape of an inflation squeeze. Many small increases, stacked. Each owner of each cost line can say “mine only went up a little.” All of them are right.

Why price increases lag

So why not just raise the price? Here is where it gets slow.

The obvious fix is to raise the price. Most brands do, eventually. Here is why it usually takes months.

Old stock hides the problem. You are selling soap you bought months ago, at last year’s cost. The new cost only reaches your books when that stock runs out. By the time margin drops on paper, the higher cost has been locked in for a while.

Retailers want notice. Stores and distributors often ask for advance notice and new cost sheets before a price change. That process can take a season.

Channels need to line up. If you raise the price on your store but not on Amazon, shoppers find the cheaper one. Raising everywhere at once takes planning.

Subscribers expect a steady price. Changing it can trigger cancellations, so brands delay.

Fear of losing rank. On a marketplace, a price rise can cut sales, which can cut search rank, which cuts sales again. That fear is real. It is also why you test instead of guessing.

So the squeeze shows up late, and the fix comes later still. Monthly measurement shortens both gaps.

Mix effects: when no product changes but margin does

Now the sneaky part. Margin can fall even when no cost moves at all.

A mix effect is a change in total margin caused by selling a different blend. Each product’s margin stays the same. The weights change.

Say Dana’s brand sells 5,000 units a month. Use this year’s per-unit margins from the table above. All numbers are illustrative.

Same units. Same per-unit margins. The brand earned $2,850 less a month because more sales moved to the losing channel. A store revenue chart would show that month as a win.

Mix also works across products. A brand that sells more of its cheap trial size and less of its large refill will see margin fall with no cost change at all. The same holds for growing a bundle that carries a deeper discount.

During inflation, mix and cost changes arrive together. If you only watch blended margin, you cannot tell which one moved it. That matters, because the fixes are different. If cost moved it, look at price or suppliers. If mix moved it, look at where the ad money and promos are pushing sales.

What to measure every month

Fees per unit are their own topic. See Amazon fees per unit for how to read them.

Two rules make these numbers trustworthy. Date every cost, so an old order is not costed at a new price. And read fees from settled data, meaning the marketplace’s statement of what it actually paid you, not an estimate.

How to test a price rise

Price elasticity is how much demand changes when price changes. You do not need a formula. You need a test.

Back to the soap on Amazon. At $24 with this year’s costs, it makes $2.50 a unit. Say it sells 4,000 units a month. That is $10,000 a month in contribution margin.

Now try $26. The referral fee rises to $3.90. Margin per unit becomes $4.20. All numbers are illustrative.

The price rise only loses money if units fall by about 40%. That is the break-even volume. Knowing it before you test tells you how scared to be.

A few rules for the test itself:

1. Change one product on one channel. Keep the rest steady, so you know what caused the result.
2. Hold ads and promos steady. Otherwise you are testing the ads, not the price.
3. Run long enough to cover normal weekly swings. Two to four weeks is a common starting point.
4. Judge on margin dollars, not units. Units will usually fall. The question is whether margin rises.
5. Watch for things that spoil the read. A stock-out, a rival’s sale or a holiday can each swamp the effect.

One caveat: if conversion drops, ad cost per order may rise too. Watch both.

Where Synthesis fits, and where it does not

Synthesis is ontology software and an AI harness in one plan. It connects your stores, ads, email, books and files, and checks numbers against their source. For example, Amazon sales tie to settlements and ad spend ties to each account. So contribution margin by channel can be read the same way every month, in Claude or ChatGPT, or in the Monday report.

It will not negotiate your freight or pick your price. A price test still needs your judgment about what shoppers will accept. And margin by channel is only as good as the landed costs you give it. If a product has no cost on file, that gap needs filling first.

A checklist for this month

1. Rebuild landed cost per unit for your top products, with this year’s freight and duties. Date each cost.
2. Work out when the new cost hits. How many weeks of old stock are left?
3. Calculate contribution margin per order by channel, at the new cost.
4. Find any channel or product now below zero. Those need action first.
5. Split last quarter’s margin change into cost effects and mix effects.
6. Pick one product for a price test. Work out the break-even volume first.
7. Set a monthly date to repeat steps 1, 3 and 5.

The short version

Inflation rarely arrives as one big cost. It arrives as many small ones, on different bills, months apart. Old stock hides it, and mix can make it look worse or better than it is.

The defense is dull and it works. Know your landed cost per unit. Know your margin per order on each channel. Watch the mix. Test prices with a break-even number in hand.