A thumbs up is not evidence

An owner-operator running about 165 womenswear stores used to run her sign-offs through a chat group. Ask the question, collect the answers. Every store, every time, a thumbs up.

Then she would walk into a store and the work wasn't done.

"If I asked everyone in my chat group have you all signed off every store for VM… everyone had to give me a thumbs up. And at any point in time if we went and did store visits and something wasn't done, I would go back and I would say, but you gave me a thumbs up for that."

Somewhere else, in a very large hardware and home improvement chain, photo evidence was mandatory. Every task, a photo, sent up. One store manager worked out that nobody was looking. He photographed his carpet and submitted it for everything. It ran.

Neither retailer lied to head office. Both answered the question they were actually asked. One asked for a confirmation and got a confirmation. The other asked for a photograph and got a photograph. Neither asked for evidence, and neither governed the evidence it did ask for.

A thumbs up is a social confirmation, not a record

When you ask twelve people in a group chat whether the work is done, you are not running a check. You are asking them to publicly position themselves in front of their peers and their boss. What comes back is a statement about the relationship. It tells you nobody wants to be the one holding up the thread. It tells you nothing about the shelf.

An ops leader at a multi-brand apparel group with more than 500 stores put the gap in one sentence:

"There was no way of being able to confirm where a store had or hadn't put up signage, activated the campaign, whatever it might be… I can't run a report and say, okay, yep, these stores have completed all the checkpoints to ensure that this campaign is now live."

Read that again, because it is the whole problem in miniature. Not "I can't see who's behind." I cannot produce a document that says the campaign is live. The thing being sold, in every store, right now, and there is no artefact anywhere in the business that confirms it.

Ask an operator to describe how it works today and you get the same answer in different words. Here it is from a large sporting goods and outdoor retailer:

"Excel spreadsheets, emails, clunky back and forths, unable to see execution, follow-ups, chase ups. Does that give you a flavour?"

Two separate failures are hiding in that sentence. The first is that you cannot control what goes down. The second is that you cannot see what comes back. Most retailers try to fix the first one and are surprised when the second one doesn't move.

Photo evidence you don't govern becomes a photo of the carpet

The carpet story has a coda. The same operator drew the rule out of it himself:

"Don't ask for it if you're not going to govern it and check it. It loses credibility really quick."

Not "photos don't work." Ungoverned photos don't work — and worse, they cost you something, because now the retailer knows the ask is theatre and applies that lesson to the next ask.

There are subtler versions of the same failure. A fashion retailer of around 60 stores had a VM team that required approval before a store's window photo could be uploaded. The intent was quality control. The effect was that the support office could not see whether the work had been done at all, because the evidence was sitting in an approval queue. Evidence held hostage by a sign-off is not evidence.

A national liquor chain has third-party installers going into stores, photographing their own work, and sending it back to a different team inside the business. The operator's summary: "and I have no idea what they're doing."

Photo capture is the cheap part. Anyone can bolt a camera roll onto a checklist. Governing the photo — routing it, checking a sample of it, closing the loop when it's wrong — is the part that decides whether any of it means anything in six months.

Misexecution gets found at the till, on a store visit, or by your supplier

Three ways retailers are actually discover the work wasn't done. All three are too late.

On a walk-in. At a national liquor chain, an area manager instructed twelve stores to place price signage on the floor rather than at eye height. An entire region, uniformly wrong, for as long as it took someone to visit: "I couldn't even see the price when I walked into their store."

In the monthly audit. A large apparel group only surfaces issues on an audit cycle — and knows it's inadequate. Their own words on the exception: "when it comes to safety and risk, you really need to know right then and there."

From the supplier. A large convenience was told by a confectionery supplier that an off-location placement the retailer had been paid to run in 500 stores was actually live in 300. The supplier ran its own check. The retailer found out from the counterparty.

An optical and audiology franchise retailer described the general case:

"We know it's going to just live in the systems, but we don't actually know if they've gone behind it in store. We don't know if they've merchandised it effectively unless somebody actually visits and goes to see it."

That's the real state of most retailers. Your discovery mechanism is a human being walking through a door. It works. It does not scale past roughly fifteen stores per field leader, and every hour spent on it is an hour not spent coaching.

Reported, verified and proved are three different claims

Most conversations about execution collapse three separate things into one word. Pull them apart and the tooling question gets much easier.

Reported — someone said it was done. A tick, a thumbs up, a reply-all.

Verified — the work was confirmed against a standard. A photo, a form, a timestamp, a location, checked or scored against what was asked for.

Proved — the completed work is tied to what happened next. Traffic, shopfront conversion, basket, sales in that store in that window.

Almost every retailer has the first one. Almost every buyer assumes they are shopping for the third.

Verification is ordinary, by the way. A veteran of a large discount department store group made the point better than any vendor could: it works "the same way that when someone comes into your home to change your oven, they take a photograph that the oven's been changed and they send it back to their parent organisation." Nobody thinks the oven technician is under surveillance. It's a trade record.

But stop at verified and you have built a stricter form of admin. Your stores will read it as exactly that, and they will be right.

Verification only survives if the store gets something back

This is where most execution programmes die, and the corpus is unambiguous about it.

That same optical and audiology retailers had already tried photo submission and abandoned it:

"In the past, we have used some software… where we've asked them to take photos, send it in, all that sort of stuff, but the feedback is it's just a bit too micromanagey — they don't like that approach."

The owner-operator with the chat group: "if the team think you're just using it as a carrot and stick approach, it doesn't really work."

A pet specialty operator named the endpoint precisely: "what my concern is for some of them is whether they become a tick and a flick, just to get them done."

Notice that none of those are objections to being measured. They are objections to being measured in one direction only.

The answer showed up in the same conversations, from the operators rather than from us. One retailer's own hypothesis about their business: "generally what we see is a store that is run more efficiently and well executed has a better shrink result." Another operator's test for any platform he'd consider: "the ability to connect a benefit or an outcome to an action."

A store will photograph a window every week for a year if the photograph comes back attached to what that window earned. It will not do it for a percentage on somebody else's dashboard.

Six questions to ask any system that claims to show you execution

Take these into the demo. Every one of them was asked by a retail operator on a real call.

1. How do you stop a store just ticking everything off and saying "yeah, I've done it all"?
2. Can a regional leader see one task's progress across her ten stores on one screen, without collecting anything by hand?
3. At the end of a season, can she pull "my area completed this on time, we achieved 97%" — without rebuilding it manually for a performance review?
4. If the work isn't up to standard, can it be sent back and redone — and is the redo visible?
5. Does the record survive a person leaving? Retailers llose whole campaigns this way, because the instruction lived in one inbox.
6. Can you see what the completed work did to that store's numbers, in that window?

The short version

Reported gets you a percentage you don't believe. Verified gets you a defensible record and a store team that resents it. Proved gets you a record and a reason for the store to keep producing it.

Most retailers are buying the third and installing the first.

Frequently asked questions

What is verified completion? A task confirmed against a standard rather than self-declared. A tick tells you someone pressed a button. Verified means there is a photo, a form response, a timestamp and a location attached to it, and that someone or something checked it against what was actually asked for.

How do you stop a store ticking everything off at once? Make the confirmation cost more than a tap — a photo, a short form, a count — and then check a sample of it. Evidence nobody ever looks at gets gamed within a fortnight. That is how you end up with a photograph of a carpet.

Does photo verification feel like surveillance to store teams? It does when nothing comes back. Retailers who have tried and dropped photo submission describe it as micromanaging. The ones it sticks for are the ones where the store sees what the work it evidenced actually did.

What is a realistic verified completion rate? Ask a better question first: what is your current reported completion rate worth? Retailers routinely report above 90% against tick-based confirmation, then discover on a store visit that an entire region did the same thing wrong. Set a baseline before you set a target.

Is this different from a digital checklist? Yes. A digital checklist moves the tick from paper to a screen — one senior operator called it "a version of the Sharpie with the tick, just online." Verification adds evidence and a check. Proof adds the outcome the work drove.

Every quote in this piece comes from a recorded conversation with a retail operator. Names, brands and identifying details have been removed.