Multi-Store Hardware Refresh Cycles: When and How

Every multi-store IT rollout eventually answers to the refresh question: replace the fleet on a schedule, or run it until it hurts? Both answers cost money; only one of them costs it predictably. Here's how retail estates decide, and how the resulting rollout should run.

The case against "run to fail"

Sweating assets feels frugal and bills the difference through the back door. Ageing fleets fail more, and every failure in a store is a lane down during trade, a call-out, and sometimes an emergency freight leg. Support costs climb; spares for discontinued models get scarce and expensive; and the estate slides out of OS support, a compliance cliff with a date on it. Worst, unplanned failure clusters arrive on their schedule, not yours, and never during a quiet week.

What actually sets the cycle

The textbook says four to six years for front-of-house retail hardware. The better answer comes from your own fleet's data:

• Fault trend by device age and model. Break/fix reporting shows exactly where the failure curve bends upward. Refresh just before the bend.
• OS and platform support horizons. A terminal that can't run the next supported build has a hard retirement date regardless of its health.
• Business initiatives. New payment types, loyalty platforms and self-checkout programs often need capability the old fleet lacks. Riding a refresh on an initiative's coattails shares the freight, the night windows and the store disruption.
• Residual value. Recent-generation hardware retires into meaningful remarketing returns; obsolete hardware retires into recycling fees. Earlier refresh recovers more.

Running the refresh as a program

A fleet refresh is the friendliest large rollout there is: known sites, known quantities, no new capability risk. Which is exactly why it gets under-planned and then surprises everyone. The method holds: audit the estate (records will be wrong), stage per store, deliver in waves around the trading calendar, evidence every site, and pull the old fleet the same night into wiping and remarketing. One visit per store, in and out.

The refresh is also the cheapest moment to fix accumulated sins: cable chaos, unstandardised builds, asset-register drift. Fold the cleanup into the playbook and the estate exits the program healthier than any audit could make it mid-cycle.

The planning rule

Decide the cycle from fleet data, then hold it. Estates that refresh on evidence spend less than estates that refresh on pain, and they spend it in planned waves instead of emergency call-outs. The fleet-wide view lives in our piece on hardware lifecycle management.

Fleet approaching the bend in the curve? Speak to an expert.