Payoff Reads · Self-checkout

Did self-checkout pay for itself, and where would it pay to add more self-checkout units?

Find out what self-checkout did to labor, rush-hour sales, basket size and shrink at every store that installed it, vs. similar stores that didn't install it, and which stores should get it next.

7 Paid for itself
2 No significant effect
1 Cost money

Heartland Markets is a fictional chain. Every number on this page is made up to show the format.

What your numbers can't show

The labor savings show up in payroll. Sales from customers who stayed because the line was shorter don't show up in any report.

What you have: payroll, which shows fewer hours at the stores with self-checkout.
What it can't show: 3 things that land in other reports, months apart, or in none, so usually nobody lines them up:
- whether shorter lines during the morning and lunch rush kept customers who used to give up and leave. No report records a sale that never happened.
- whether customers bought less without a cashier prompting an add-on.
- whether more customers walked out without paying.
What the read shows: labor hours, rush-hour transactions, basket size and shrink, read together at each store that installed it, and the net in dollars.

Operations, loss prevention and IT often disagree about self-checkout because each sees one part of it. The read puts all of it on one page.

What lands on your desk

Self-checkout at ten stores, as Seurat reads it

Heartland Markets put self-checkout in 10 of its 42 stores, in 3 waves from February 2025. Each store is read for the 26 weeks after its install, vs. a twin built from stores without it.

The answer, in four lines:

Labor fell 34 hours per store per week, statistically significant: about +$620 per store per week.
Shorter lines kept customers: transactions during the morning and lunch rush rose 3.0%, statistically significant, while other hours didn't change. That's about +$210 per store per week in margin, most of it at the 4 stores with the longest lunch lines.
Baskets got 1.8% smaller, statistically significant: about −$300 per store per week in margin.
Net: about +$530 per store per week before shrink, and it paid for itself at 7 of the 10 stores. Shrink rose 0.2 points of sales, which isn't statistically significant after 2 inventory counts.

all 10 stores vs. their twins

What changed	Result	Verdict
Labor hours per store per week	−34	Significant drop
Rush-hour transactions	+3.0%	Significant lift
Basket size	−1.8%	Significant drop
Shrink, in points of sales	+0.2	No significant effect
Net per store per week, before shrink	+$530	Paid for itself

store by store (6 of 10)

Store	Labor hours per week	Rush-hour transactions	Basket size	Net per week, before shrink	Verdict
Store 7, Center Point Rd	−41	+6.4%	−0.9%	+$1,050	Paid for itself
Store 12, Ely Rd	−38	+5.8%	−1.2%	+$900	Paid for itself
Store 36, Westgate Dr	−31	+5.9%	−2.4%	+$580	Paid for itself
Store 29, Market St	−35	+0.8%	−1.5%	+$440	Paid for itself
Store 19, Park Ave	−29	+0.6%	−2.9%	+$90	No significant effect
Store 41, Bridge St	−27	+0.4%	−4.9%	−$300	Cost money
All 10 stores	−34	+3.0%	−1.8%	+$530	Paid for itself

Store 41: its cashiers used to sell the hot food add-on at the counter. Without them, baskets fell most.

Rush hours: the 4 stores with the longest lunch lines before the install gained the most, about 6% more transactions during the rush. At the other 6, rush-hour transactions barely moved.

Shrink, explained: shrink is measured as a share of sales, so 0.2 points means going from, say, 1.5% of sales to 1.7%.

Where it would pay next: of the 32 stores without self-checkout, 9 have a morning and lunch rush as busy as those 4 stores had before their installs, so those 9 should get it next. The 4 busiest stores that already have it are the first place to try a 2nd self-checkout unit.

KeepWhat to do: keep it, install it next at those 9 stores, and put the add-on prompt on the self-checkout screen at the stores where baskets fell most.

The confidence band shows the 95% confidence interval. When the band spans zero, the result isn't statistically significant. Heartland Markets is a fictional chain. Every number on this page is made up to show the format.

How the read works

Each store that installed self-checkout gets a twin from your stores that didn't install it, weighted so the twin matches the store's weekly sales and transactions for a year or more before the install.
Labor hours, transactions by hour and basket size are read week by week vs. the twin, and shrink at each inventory count.
The read checks itself. The twin has to have tracked the store closely before the install. Placebo tests then rerun the read on a made-up install date, when nothing changed, and on stores that never installed it. Both should show no effect. If either shows an effect, the read isn't working properly, and we tell you.
Each store gets a net in dollars: the labor saved, plus the margin on the extra rush-hour transactions, minus the margin lost on smaller baskets, minus any extra shrink, with a confidence interval and a verdict.
To show where it would pay next, the read finds the store traits that went with the biggest gains, such as a busy rush, and ranks the stores without self-checkout on those traits.

Side note: Installs usually go store by store, the setup this type of read handles best: each wave is compared to stores still waiting for theirs.

Can your chain be read?

Measure and setup	Smallest change it can show
Labor hours	Steady week to week, so the easiest of the 5 to read
Transactions or basket size, 10 of 40 stores, 26 weeks after the install	About 1.4%
Rush-hour transactions, 10 of 40 stores, 26 weeks after the install	About 2.5%
Basket size at 1 store on its own, 26 weeks after the install	About 4%
Shrink at inventory counts, 10 of 40 stores, 4 counts before the install and 2 after it	About a quarter of your shrink: if shrink is 1.6% of sales, a change of about 0.4 points

Rough figures from a standard formula, before we see your data.

Scope this read

What we need, and what stays with you

We need:

Install dates at each store
Weekly labor hours by store, totals only
Weekly transactions and sales by store, with item counts if you have them
Transactions by hour, so we can read the morning and lunch rush
Shrink at each inventory count, by store
1 to 2 years of history

We never ask for:

Cashier names, IDs or schedules
Customer or card data
Your chain's totals or your P&L

How we handle your data →

How long it takes

About 2 weeks from the day your data lands, then a 60-minute readout with whoever needs to hear it.
We scope it to your chain on a short call: which stores, which measures, which counts.

Scope this read Book a call

Questions buyers ask us

We installed it everywhere within a few weeks. Can it still be read? Usually, if the installs were spread over a month or more: the stores still waiting are the comparison. If every store went live the same week, we'll say so on the first call, before you commit.
Can it tell us about shrink? Only if shrink changes by about a quarter, because it's counted every few months rather than every week. Labor, transactions and basket size are weekly and read well.
Can it tell us where to put more self-checkout units? Yes. The read shows which kinds of stores gained most, usually the ones with the busiest rush, and ranks the stores without self-checkout by how closely they match those stores.
We tried self-checkout before and took it out. Can you read that? Yes. Taking it out is a change too, read the same way.
What if the read shows it didn't pay? A read with no significant lift is a private finding. You see every result first, and nothing goes anywhere without your approval.
Do we have to connect any systems? No. One export from your back office and payroll, sent however your IT team prefers.
Who does the work? We do. Your team runs the export, shares it with us, and comes back to hear our readout.

After this read

Read several changes in one request, or keep your stores measured week by week.

Scope several changes →How monitoring works →