The July CPI report came out last week and the food-at-home number was 2.7% higher than a year earlier, which is the same reading as June. By recent standards that is a calm number. It is roughly in line with general inflation and nowhere near the 2022 stretch when groceries were running double digits.
And every time one of these lands I see the same reaction, which is some version of: that is not what my receipt says.
I have been staring at our own price data for long enough now to have an opinion about why. The 2.7% is correct. The reaction is also correct. And the thing that reconciles them is not really about inflation at all.
What the number is measuring
Food at home is a weighted average across a very large basket. Every category is in there, weighted by how much Americans collectively spend on it. Some things went up, some went down, and 2.7% is the middle of all of it.
You do not buy the middle. Nobody buys the middle. You buy roughly forty items, over and over, and the inflation rate of those forty items is your inflation rate. It can be a long way from the average in either direction.
Even inside the official data the spread is obvious. In one recent month beef and veal rose while fresh vegetables fell, in the same month and the same report. A household that eats a lot of beef and few vegetables had a completely different month from one that does the reverse, and both of them are inside that single national figure.
Eggs are the clearest recent case. They fell hard this year because they had risen absurdly the year before, and I wrote about that whole cycle in March. A big decline in one category pulls the national average down for everybody, including people who barely buy the thing. If you do not eat many eggs, you got none of that relief and you are still counted inside a number that includes it.
That is not a flaw in the statistic. It is a flaw in using the statistic as a personal forecast.
The comparison nobody makes
Here is what I actually want to put in front of you.
Earlier this year I compared the three Pittsburgh chains we track on price per unit, using our own data rather than anybody's index. Across the staple categories where the data was clean enough to call, the gap between the cheapest chain and the next one up ran between about thirty and sixty-six percent, depending on the category.
Thirty to sixty-six percent. Against annual grocery inflation of 2.7%.
SMOPPER INSIGHTS
A year of grocery inflation, next to the cost of picking a store
Inflation: BLS food-at-home CPI, July 2026, year over year. Store gap: trimmed median price per canonical unit between two chains in the same market, Smopper data.
The store you walk into is worth something like a decade of average grocery inflation, decided once, and then re-decided every week by default because you did not think about it.
I find that genuinely clarifying, and I would like to be careful about what I am and am not claiming with it.
Three honest caveats on that chart
They are not the same kind of number. Inflation is a change over time; the store gap is a difference at one moment. Putting them on one axis is a rhetorical move, not a statistical one. I think it is a fair rhetorical move, because both of them land on your bill in dollars and you can only do something about one of them.
The store gap is a category-wide median, not a receipt. The two chains do not stock the same mix, and a store with a deeper premium range lands higher on a median. If you buy the cheapest qualifying item at each store the real gap is narrower than this. If you buy from the middle of what is in front of you, which is what most people do, it looks about like this.
Shelf price only. No digital coupons or loyalty discounts, which systematically favours the chain with less promotional machinery. A committed coupon-clipper at the more expensive store closes some of that gap. I cannot tell you how much, because those offers are account-specific and we do not collect them.
Even discounting all three of those generously, the comparison does not get close. Nothing plausible turns a forty percent structural gap into 2.7%.
Why the conversation is the wrong shape
Grocery inflation gets talked about constantly. It is in the news every month, it has a number, it is nobody's individual fault, and it happens to everybody at once. That last part is what makes it good to talk about. It is a shared experience with no accusation attached.
Store choice has none of those properties. It is not news, there is no monthly release, it varies by person, and there is an unpleasant implication buried in it that if you are paying more you could have paid less. So it does not get discussed, and the far larger number sits there unexamined while everybody follows the small one closely.
I am not going to pretend the choice is free. It is not.
Distance is real. If the cheap store is a twenty-minute drive and the expensive one is on your way home, the gap has to be big before it is worth it, and for a lot of people it is not. Time is real too, especially if you are shopping with children in tow. And there is a version of "just shop at the cheaper store" that is offered to people who do not have a car, or who live somewhere the cheaper store simply does not operate, and that advice is worse than useless.
What I would say is narrower. If both stores are already plausible for you, the difference between them is the largest single lever on your grocery bill that you control, and it is larger than inflation by an order of magnitude. That is worth knowing even if you decide, having known it, to keep shopping where you shop.
The other lever, which is smaller and more fun
Substitution. When a category runs hot for twelve months straight, that is the moment to check whether the thing you are loyal to still deserves it.
I ran that experiment on twelve products and kept nine of the swaps, which is a better hit rate than I expected and a worse one than the internet would tell you. The pattern that came out of it has held up since: the swaps that work are the ones where the product has a specification, and the ones that fail are the ones where the product has a recipe. Flour has a specification. Coffee is all recipe.
That is worth real money and it is worth less than the store question. Do them in that order.
What I would actually do with all this
Very little about the inflation, honestly. You cannot opt out of it, the categories that run hot are not the ones you would choose to give up, and the correct response to a 2.7% year is to notice it and carry on.
The two things that are actually under your control are where you shop and what you are loyal to, and both of them are worth multiples of the number in the news.
A national average is a description of a country. It is not a forecast for your kitchen and it never was. Price your own list and you will find out in about four minutes which half of that 2.7% you actually live in, and more usefully what the store on the other side of town would have charged you for exactly the same cart.