In this guide
What Gets Measured on a Visit
A visit measures three things: performance against a scorecard, behaviour inside a defined scenario, and whatever evidence was captured to support both. The scorecard is a list of observable facts with an unambiguous answer, and every line on it must be answerable by somebody standing in the outlet without forming an opinion. Was the licence displayed. Was the price on the shelf the price at the till. Was a receipt offered without being asked. The scenario is the situation the visitor creates so that behaviour under a specific condition can be observed rather than guessed at. Evidence is the photograph, the receipt or the timestamp that turns a score into something a regional manager cannot simply deny. None of this works unless the standard exists first, in writing, because a scorecard is only ever a translation of a standard, and a measurement taken against an unwritten expectation measures the visitor's assumptions.
Building the Scorecard
A scorecard is a translation of an existing standard, and it fails first where no standard exists. Sections are weighted against business priority rather than evenly, because a programme that gives statutory display the same weight as shelf tidiness will produce a total nobody can act on. Deciding the weights forces the organisation to say what actually matters, which is frequently the most useful part of the design. Binary items and scaled items serve different purposes and should not be mixed carelessly. Anything observable as a fact belongs as a binary: the notice was present or it was not. Anything requiring degree, such as how thoroughly a product was explained, needs a scale with each point described, otherwise two visitors will score the same interaction differently and the variation will be read as outlet performance. Keeping it short enough to complete on site is the constraint everything else has to fit inside. A visitor cannot hold sixty questions in their head while behaving like a customer, and a scorecard completed from memory in a car park is a different instrument from one completed from observation. Most workable scorecards are considerably shorter than the first draft.
Designing the Visit Scenario
The scenario is the situation the visitor creates so that behaviour under a defined condition can be observed rather than guessed at. It needs a realistic reason to be there. A visitor asking questions no ordinary customer would ask, or lingering without an apparent purpose, is noticed, and a noticed visitor is measuring how staff behave when they suspect an assessment. The reason should be ordinary enough to be unremarkable and specific enough to produce the interaction you want to observe. Timing, day and duration are part of the design rather than logistics. A programme run entirely on weekday mornings measures weekday mornings, and outlets vary enormously between a quiet Tuesday and a Saturday afternoon. Spreading visits across the trading pattern is what makes the result describe the outlet rather than one shift. What the visitor may and may not do has to be settled in writing before anybody goes. Whether they may make a purchase, attempt a return, ask for a manager, record anything, or press a refusal are all decisions with legal and practical consequences, and leaving them to the visitor's judgement produces inconsistent visits and occasionally worse.
Evidence That Supports a Score
A score without evidence is an assertion, and assertions do not survive contact with a regional manager who disagrees. Photographs, receipts and timestamps are the three ordinary forms. A photograph supports anything physical: the display, the shelf, the queue, the notice that was or was not there. A receipt independently corroborates the visit, the time and what was actually charged, which makes it the single most useful item on a pricing scorecard. Timestamps establish the window. Geo-fenced check-in is the strongest available proof of attendance, because it ties the submission to the location rather than relying on the visitor's account of where they were. It does not prove what happened inside, but it removes the most damaging possible challenge, which is that the visit did not occur at all. What makes a score contestable is worth understanding before the report is issued rather than after. Findings resting on judgement, on a single unsupported observation, or on a scorecard line that was never traceable to a written standard will be argued about and generally should be. Findings resting on a photograph, a receipt and a timestamp are discussed rather than disputed.
Reporting So It Gets Acted On
Findings change behaviour only when they reach the person who can act on them in a form they can use, which means the same data has to be presented three ways. The outlet view is what a store manager needs: what was observed at this site, on which visits, against which standards, with the evidence attached. The regional view aggregates outlets so a manager can see which sites are dragging and whether a problem is local or general. The network view shows whether the standard is being met at all, and it is the only view that answers whether a policy is working or merely published. Repeat findings across cycles form the second layer. A fault appearing once is an incident; the same fault at the same outlet across three cycles is a process that does not work there. The third requirement is separating a bad visit from a bad outlet, which is a question of how many visits sit behind a score. Reporting that shows the visit count alongside the score lets a reader judge for themselves how much weight the number will carry.
Running Your First Cycle
Establish the baseline before setting any targets. A first cycle exists to find out where the network actually stands, and targets set in advance of that are guesses which will either be met without effort or missed by a distance that discredits the programme. Run the cycle, read the distribution, then set targets against it. Communicate the programme internally before it starts, and be straightforward about it. Announcing that visits will occur across the network over a period does not compromise the method, because the visitor is still anonymous and the date is still unknown, and it removes the impression that the exercise is designed to catch people out. Programmes introduced quietly and revealed through their first bad result generate resistance that outlasts the finding. Be equally clear about what the findings will and will not be used for. An independent provider is worth using where the results will affect incentives, where the network is too large for internal resource, or where an internal team would be recognised, and mystery audit programmes are generally run externally for the first of those reasons.
