In September, U.S. producer prices rose 1.3% in a single month. That is a large move for one month, and it is the kind of move that quietly rewrites a lot of savings reports.
Here is the situation it creates. Say your team ran a sourcing project last spring. You negotiated hard, the new pricing landed, and the invoices came in lower than the year before. You reported the difference as savings, because that is what the difference looked like.
Then January arrives, and someone in finance asks a question that is entirely reasonable: how much of that was us?
It is the sharpest question in cost work, and in our experience at Alliance Cost Containment (ACC) it is the one that comes up most often once a program has run long enough to be measured. Not “did you find savings.” Can you prove the savings were yours?
The index moved. Your categories may not have.
It is worth looking underneath that 1.3%, because the detail says more than the headline does. It was the largest monthly advance in producer prices since March 2022, and it was nowhere near evenly spread. Final demand energy jumped 10.8%. Transportation and warehousing services rose 3.4%.
Inside the very same report, other things fell. Commercial electric power declined. Thermoplastic resins declined. Several food inputs dropped sharply.
So a supplier citing “the PPI went up” is not, by itself, making an argument. The index is a blend, and a blend can conceal two categories moving in opposite directions. Cost pressure can be entirely real and a requested increase can still be wrong.
Which brings us back around, because that is the same problem savings measurement has, pointed the other way. If a single blended number cannot justify a price increase, a single blended year-over-year comparison cannot prove a saving either.
Dollar-to-dollar is a comparison, not a measurement
Comparing last year’s invoice to this year’s invoice is easy, which is most of its appeal. It is also the method most likely to mislead you, and it misleads in both directions.
When a category’s market price falls, a flat year-over-year comparison hands you credit you did not earn. Take a hypothetical to make the shape of it clear. Suppose the market drops six percent and your negotiated price drops four — your invoice is lower and your report says “savings” — but you actually lost two points of ground against what was available. That is a report that feels good in Q2 and does not survive a hard look in Q4.
When a market price rises, the same method punishes real work. Reverse the hypothetical: a team holds pricing flat through a year when the category moves up eight percent. That is genuinely difficult work. Dollar to dollar, that shows as zero. Or worse — if part of the increase was absorbed, it shows as a cost increase, and the people who fought it off look like they lost.
Both failures come from the same place. A year-over-year comparison measures what happened. Savings is a claim about what would have happened otherwise.
What we do instead
When savings in a category are market-driven, we calculate the variance from market pricing — not a flat dollar-to-dollar difference.
That means the baseline is not what you paid last year. The baseline is what that same item, at that same volume, on those same terms, would cost at current market. The savings is the gap between that number and what you actually pay.
Where a category’s pricing is tied to an index or to other public pricing, we agree the benchmark and its source with the client before the work starts. That matters, because it settles the yardstick in advance rather than leaving it to be discussed later — and it is the honest answer to a reasonable CFO question, which is whether measuring against market lets us claim a saving in a year when the invoice went up.
It is a less flattering method on the way down and a more honest one on the way up. Some months it makes our number smaller than an invoice comparison would. We would rather carry that than defend a figure that falls apart the first time somebody pulls an index.
This is not a special rule for volatile categories. It is the standard we hold across all of our gain-share work: if we cannot measure it, we do not count it.
The same instinct applies to scope: addressable spend is usually smaller than total spend, and we would rather show you both numbers and explain the gap than quote a percentage against a total we know is inflated.
Why this matters more now than it did three years ago
In a flat market, dollar-to-dollar and market variance land in roughly the same place and nobody notices the difference. That has not been the environment for a while. Categories have been moving — sometimes sharply, sometimes in opposite directions inside the same AP file — and the gap between the two methods has grown wide enough to change conclusions, not just decimal points.
It matters for a second reason too. Savings reports travel now. They go into board packages, lender updates, diligence files. A number casually built in a spreadsheet in March becomes a number somebody else is relying on in October. The methodology note matters as much as the figure.
The part that makes people defensive, and shouldn’t
There is a version of this conversation that lands badly, and we work hard not to have it. It is the version that implies the prior work was sloppy.
Often it wasn’t. Many of the organizations we talk to have already run a serious internal cost push, and the people who ran it describe it as genuinely hard work. Teams go category by category, pull real money out, and are rightly proud of it.
What tends to survive those efforts is not the easy money. It is the categories nobody had the depth or the time to chase — where the pricing is complicated, the market data is scattered, and the person who owns the relationship has eleven other things on their desk. Those categories are also, not coincidentally, the ones where dollar-to-dollar measurement breaks down worst.
What to ask for
If you are evaluating a cost program — ours or anyone else’s — a few questions are worth asking before the work starts rather than after:
- What is the baseline, and is it last year’s price or current market?
- How is a market move handled when it goes against us? Does that show as a loss, or does it quietly get excluded?
- Who verifies the number, and against what source?
- What is in the addressable base, and what got scoped out, and why?
The answers tell you a lot. A program that has thought carefully about measurement will have those answers ready and will usually raise the uncomfortable parts before you have to ask.
It is complicated. It is also the whole job. Finding the money is the fun part. Proving it was yours is the part that lets anybody act on it.
Source: U.S. Bureau of Labor Statistics, Producer Price Index news release, September 2026
