Suppose new receiving software removes three minutes of manual data entry every time a store records an incoming pallet. At 20 pallet records a day, that saves an hour of staff time per store. Across 50 stores, the business case shows 50 hours saved every day.
The payroll report shows no change.
Those hours are spread across locations, employees and the working day. Nobody can collect them into a spare team, and no individual necessarily has enough time released in the right place to finish a shift earlier.
The arithmetic can be correct while the promised financial saving remains unproven. That leaves a practical question: what can the stores do with those minutes that makes the software worth paying for?
First, check the time saving itself. Watch comparable receipts move through the old and new processes. Measure the work people actually perform, including exceptions and corrections. A shorter interval between system timestamps might include less waiting without reducing anyone’s active work. A faster receiving task might also leave a supervisor investigating incomplete records later.
Check how often staff use the software successfully and which receipts still need manual handling. The three-minute saving from a straightforward demonstration cannot automatically be applied to every pallet across every store.
Once the saving is established, DataDocks CEO Nick Rakovsky wants the discussion to extend beyond shorter shifts:
“You might not be able to save a shift. That doesn’t mean the improvement has no return. You could have more time to check loads properly, fewer errors or less overtime. Those are the things I’d be looking at.”
Take the first possibility. Our stores receive pallets of household goods, and some quantity discrepancies are only discovered later, when staff try to replenish shelves. Someone then has to investigate whether the delivery was short, the receipt was wrong or the goods are somewhere else in the back room.
Suppose the receiving team uses part of the released data-entry time to check selected pallets more thoroughly before putting them away. Define which deliveries warrant the check, what staff should verify and how they should record a discrepancy. Three minutes beside the pallet might fit that work even though they cannot be turned into a shorter shift.
We now have something specific to test: does this change reduce the discrepancies that require investigation later?
A supermarket field experiment by Jie Gong and I. P. L. Png illustrates why the remaining work matters. Cashiers rotated between conventional checkouts and a format where customers paid at machines. When relieved of payment collection, cashiers scanned purchases more than 10% faster. The measured improvement concerned scanning speed, rather than payroll or whole-store profit. It shows that evaluating automation can involve examining how the rest of the job changes. The supermarket field study.
Our proposed receiving checks need their own evidence. Start with similar stores and collect baseline data before introducing the change. A staged rollout gives us a comparison with stores still using the existing process; where practical, randomly choosing the rollout order strengthens that comparison.
Track receipt volumes, supplier and product mix, staffing, and the checks actually performed. A quiet week with straightforward deliveries tells us little about performance during a promotion. If software and new checking procedures arrive together, the evaluation concerns that combined change. It cannot attribute every improvement to the software alone.
There is an easy measurement trap here. Better receiving checks might increase the number of discrepancies reported at the door. On a simple error dashboard, the trial could look worse precisely because people are catching problems earlier.
Separate those early discoveries from discrepancies that escape receiving. After routine receiving checks and before putaway, use the same sample-audit method in both groups to compare recorded quantities with the goods actually received. Record investigation and correction work both at receiving and afterward. Keep an eye on receiving completion times, so that more thorough checking does not quietly create a backlog.
The old data-entry task may be shorter while the complete receiving process takes just as long. If the difference is being spent on useful checks, that is a deliberate use of the released time. The evidence needs to show whether those checks help.
The UK Government Digital Service’s Digital and Data Benefits framework, published in April 2026, offers a useful way to assess this kind of outcome: connect a reduction in errors with the cost of reprocessing them. It also recommends testing uncertain assumptions about adoption and efficiency gains. This is appraisal guidance; applying it to our stores means measuring which later investigations disappear and what resources they would have consumed. The 2026 benefits framework.
That last step determines the kind of benefit we can claim.
If employees spend less of their existing shifts investigating discrepancies, the store has released more capacity. If paid overtime falls because those investigations no longer spill beyond normal hours, there may be an expenditure saving. The same operational improvement can have different financial consequences in different stores.
The Government Efficiency Framework makes a useful distinction between direct spending reductions and improved output with unchanged spending. Although written for government, that distinction helps structure a retailer’s proposal: explain what improved, then show separately whether expenditure changed. The efficiency framework.
Different claims call for different evidence:
| Proposed benefit | What would support it? |
|---|---|
| Fewer costly corrections | Fewer discrepancies escaping receiving, with lower investigation or correction effort |
| More work completed with existing staff | More comparable receipts completed within the same paid hours, with quality maintained |
| Less overtime | Lower paid overtime against a credible comparison at equivalent workload and service |
| An additional shift avoided | Evidence that the expected workload would otherwise require that shift |
These are possible outcomes to investigate, rather than benefits to add automatically to every business case. More receiving capacity has limited immediate value if there is no additional work to process. An unused hiring budget does not prove that the software prevented a necessary hire.
A wage-rate calculation can help describe the resources involved, provided its assumptions are clear. It does not establish that the wage bill will fall. And if the released minutes are spent checking pallets, they cannot also be counted as time available to shorten shifts. Avoid claiming the same benefit twice through different calculations.
Now compare the demonstrated result with the full cost: licences, equipment, integration, training, implementation and ongoing support. Include the effort required to handle exceptions. Test whether the case still holds with lower adoption or a smaller reduction in discrepancies.
Nick’s wider view of value still comes with a demand for evidence:
“If the only case is that it feels a little better, I’d start questioning whether it’s worth it. You need to show what improves and how that helps you get where the business wants to go.”
A retailer can decide that a more reliable receiving process is worth funding while payroll stays unchanged. The proposal should explain the improvement being purchased and its cost. If later phases are expected to reduce overtime, make that a conditional benefit with a separate test.
Before expanding, agree which outcome would justify the next group of stores, who owns it and when it will be reviewed. Sites with frequent discrepancies may have a different case from sites where receiving already works well.
The next proposal should be able to explain what staff will do differently with the released minutes and how the business will know it helped. If it can only show a larger total of hours saved, an important part of the investment case is still unfinished.
