Dollar General pulled back on self-checkout. Shrink outran the savings.
In early 2024, Dollar General announced it was removing self-checkout from 300 stores and limiting the format in thousands more. The chain, with nearly 20,000 stores in the United States, had bet on self-checkout to cut labor. The math did not close.
The case became an example of “self-checkout failed”. That reading is shallow. Self-checkout did what it promised, reduce registers. What was missing was measuring what it opened on the other side.
What was announced
In prior years, Dollar General expanded self-checkout as a central piece of efficiency. Fewer cashiers per store, shorter lines, lower labor cost. The logic was solid on paper.
Trade press (Retail Dive, CNBC) covered the expansion as a productivity move. The number that mattered, at the time, was labor hours saved.
What changed in the public record
In March 2024, on the earnings call, the company communicated the turn. It removed self-checkout from the 300 highest-loss stores. It converted thousands of stores from full self-checkout to assisted, limiting the item count. The CEO, back in the role, named the cause: shrink.
The labor savings were real. The loss that came with it was larger. Unwatched self-checkout had moved the cost from the register to the inventory, and the inventory charged more.
The decision that explains the gap
It was one. Self-checkout was scaled as a productivity decision, measured in hours saved, with no loss criterion per store measured before the rollout.
When shrinkage rose, there was no trigger to fire early. The correction came at earnings-call scale, not pilot-store scale. Whoever measures loss per store before scaling catches the problem in week three; whoever measures only labor catches it in quarter nine.
The criterion that would have caught it earlier
One number, defined before the first rollout: the loss delta per store with self-checkout, compared against the labor savings in the same store, with a review trigger if loss exceeded savings. That is the indicator that goes on the operation’s wall.
Take the efficiency project on your roadmap right now. Can you say, in one sentence, which loss indicator it could move in the wrong direction? If the answer is no, you are in the same position Dollar General was in before 2024.
This piece rests only on the public record. We have no access to the chain’s internal operation; the reading is built on what was disclosed.
Defining that criterion alone is hard, because it is a fight between operations, loss prevention, and finance, and each area defends the number it already measures. Send me the three indicators your operation already tracks on self-checkout and loss. In one hour I will send back which would have predicted Dollar General’s reversal and which are noise. If we see a project shape, the Diagnóstico is two weeks and ends with a one-page document: three indicators, three target ranges, three review triggers.