Store performance analysis: a triage method for which stores need attention

Why triage beats ranking

A store performance analysis that ranks 300 stores from best to worst produces a long list and few decisions. Regional managers do not need to know which store is 147th; they need to know which five stores to call this week and what to talk about. Triage answers that directly.

What you need

• Every store's reviews for the current window and the previous one, classified into the same themes.
• Each store's peer group: stores of similar format and location type.
• A minimum number of reviews below which a store's rates are treated as uncertain.
• A short list of serious issues that always trigger action, such as safety, hygiene, discrimination or fraud.

The triage rules

Apply these rules to every store each month. The thresholds are starting points; tune them to your network.

• Act now: any serious issue appears in recent reviews, whatever the volume.
• Act now: a theme's negative rate is clearly above the peer group (for example, at least one and a half times the group's rate), it rose compared with the previous window, and the store has enough reviews to trust it.
• Watch: the same pattern, but on a sample below your minimum.
• Watch: above peer rates that are stable rather than rising, or rising rates that are still within the normal range.
• Fine: everything else. No action, but keep the data.

A worked example

• Store A, 220 reviews in the window: staff-behaviour complaints 14% against a peer average of 6%, up from 7%. Act now.
• Store B, 35 reviews: waiting complaints 23% against a peer average of 10%, up from 12%. Watch; the sample is small, so read the reviews and check again next month.
• Store C, 400 reviews: cleanliness complaints 9% against 8%, stable. Fine.
• Store D, 60 reviews: one review describes food poisoning. Act now, whatever the rates.

Four stores, two calls this week, one to keep an eye on, one to leave alone.

What act now means

1. Read the reviews behind the issue before the call. Bring two or three in the customer's own words.
2. Agree the likely cause with the store and area manager: staffing, a process, training, stock, a policy.
3. Name an owner and a date to look again, usually four to eight weeks later.
4. Record the action, so the next month's review can check whether the issue moved.

What watch means

Watch is not a softer act now. It means no action yet, but read the reviews and look again next month. Many watch stores return to fine on their own; a few move to act now with enough evidence to be confident. That is the point: you avoid spending a regional manager's time on noise.

When many stores move together

If the same issue sends many stores to act now in the same month, it is probably not a store problem. Look for a common cause: a new policy, a supplier change, a system outage, a pricing change. That goes to the central owner of the theme, not to each store manager.

Making it a monthly habit

• Same day each month, same rules, same view.
• Start with last month's act-now stores: did the issue move?
• Then this month's new act-now stores, then a quick look at the watch list.
• Close with network-wide patterns that need a central owner.

Setting the thresholds

The rules depend on three numbers you choose once and keep stable: the minimum reviews for a store to be ranked, how far above the peer rate counts as clearly above, and how big a rise counts as rising. Start with simple values, run the triage on the last six months, and check whether the stores it flags are the ones your regional managers would have picked. Adjust once, then leave the thresholds alone, so month-to-month changes reflect stores, not rules.

• Too many stores in act now: raise the above-peer multiple or the minimum sample.
• Known problem stores missing: lower the multiple, or add the issue to the serious list.
• Stores bouncing between groups every month: lengthen the window.

Pitfalls

• Treating watch as act now. It wastes visits on noise and teaches stores to distrust the list.
• Leaving serious issues to the rates. One credible safety review is enough.
• Triage without evidence. Every flagged store should come with the reviews that put it there.
• No follow-up. A store that leaves act now without a recorded result has not been fixed; it has been forgotten.

Doing this with Falcon

Falcon classifies every public Google Maps review for each store into themes, sub-themes and issues with its sentiment. Experience signals show negative share by topic and month, findings built on small samples are flagged rather than ranked, and every number opens the reviews behind it, so a regional manager can compare stores on what customers actually said.