Retail audits and store checks: finding out what happens to your product on the shelf

1. The problem: sales to the distributor, silence at the shelf

A beverage producer sees in its internal reports that shipments to distributors are steady. Yet sales fall the following quarter. The explanation is not in the company's data but in the shops: in some of them the product has been missing for weeks, in others it sits on the bottom shelf behind the competition, and in a few it sells above the recommended price. The producer measures what leaves its warehouse, not what the shopper sees. A retail audit is the method that closes this gap: trained auditors visit a sample of stores and record systematically whether the product is there, how it is displayed and what it costs.

The problem is not confined to fast-moving consumer goods. A maker of over-the-counter medicines wants to know how many pharmacies stock its products. A telecoms company wants to check whether its promotional materials have reached partner outlets. A public institution may want to monitor the availability and price of basic goods. In every case the question is the same: what actually happens at the point of sale.

2. What a retail audit is, and what a store check is

An audit is a form of structured observation: the data do not come from anyone's answers but from what the auditor sees and counts, using a fixed grid. In its classic form, the audit is repeated at regular intervals on the same panel of stores and, by counting stock and deliveries, can also estimate in-store sales. In the form companies most often commission, it concentrates on the shelf: product presence, visibility and prices.

The store check is the one-off version: a single visit to a limited number of shops that captures the situation at a given moment. It is useful before a launch, after a campaign, or when the sales team has a suspicion that needs checking quickly. The repeated audit is the monitoring tool: it shows trends, compares regions and channels and measures the effect of commercial actions between waves, as part of monitoring and testing services.

The difference from a survey is fundamental. A survey asks shoppers what they saw or bought and gets reported behaviour. An audit asks nobody anything: it records the state of the shelf. For that reason it cannot explain why shoppers choose a product, but it describes precisely the conditions in which they choose.

3. What is measured: the core indicators

• Numeric distribution: the percentage of stores in the universe where the product is present. If it is found in 60 of 100 audited stores, numeric distribution is 60%.
• Weighted distribution: the same presence, weighted by each store's importance in category sales. A product found only in large stores can have low numeric and high weighted distribution. The calculation needs an estimate of category sales per store, usually from the audit itself or from a proxy such as floor area.
• Out-of-stocks: stores that list the product but where it is not on the shelf on the day of the visit. This is different from a distribution gap and calls for a different fix.
• Shelf share: the proportion of facings or shelf length the brand occupies within its category.
• Position and planogram compliance: shelf level, proximity to competitors, adherence to the layout agreed with the chain.
• Shelf price and promotions: the displayed price, deviation from the recommended price, competitors' prices and active promotional mechanics.
• Point-of-sale materials: the presence and condition of posters, stands and branded fridges.

Not every indicator is needed in every project. The list is built from the decision: a sales team negotiating listings needs distribution; a marketing team paying for display space needs shelf share and compliance.

4. When it is the right choice, and when it is not

A retail audit is the right tool when the decision depends on in-store execution: deploying the sales force, negotiating with chains, checking distributors, evaluating a trade marketing campaign, tracking competitors' prices. It also fits when chain sales data are unavailable or do not cover traditional trade, small shops and markets, which remain invisible to scanning systems.

It is the wrong tool when the question is about the shopper: why they choose, what they think of the brand, how satisfied they are. Those call for surveys, focus groups or ethnographic observation in store. Nor is it right when what needs assessing is the interaction with staff, the quality of advice or compliance with service standards. There the correct method is mystery shopping, in which the evaluator plays the part of a customer.

5. The store sample

Every audit starts with the store universe: the list of outlets relevant to the category, with type, address and ideally a measure of size. The sources are distributor databases, public registers and, where these fall short, a field census of shops in a set of selected localities. Without a universe, numeric distribution has no denominator and cannot be interpreted.

The sample is drawn from the universe, stratified by channel (hypermarkets, supermarkets, convenience stores, traditional trade, petrol stations, pharmacies, depending on the category) and by region, with Chișinău and Bălți usually treated separately from the rest of the country. Within each stratum, stores are selected at random. Strata that are small but carry a large share of sales, such as hypermarkets, can be oversampled or even included in full, and results are then weighted back to the universe. For example, a hypothetical audit of 300 stores might include every hypermarket plus a random sample from the other channels.

For monitoring, the same store panel is kept from wave to wave, with replacements only for shops that have closed. Comparisons between waves are then more precise than with a fresh sample each time, because differences between stores no longer get mixed up with real change.

6. Fieldwork: auditors, grids, evidence

The instrument is an audit form on a tablet or phone, with the exact product list (by code and pack size), fields for presence, number of facings, price and materials, and clear recording rules. Ambiguities are settled in training, not in the field: what "present" means when the product is in the back room but not on the shelf; how facings are counted in a chiller; which price is recorded when a promotional price is also displayed.

Each visit produces evidence: shelf photographs stamped with date, time and location, which allow later verification. A share of visits is checked independently, by revisiting or by comparing the photos with the data entered. Auditors need the store's permission to take photographs, and with large chains this is usually arranged centrally. People who happen to appear in photos are not identifiable in the deliverables.

7. Analysis: from the shelf to decisions

Results are reported by stratum and by wave: numeric and weighted distribution by channel and region, out-of-stocks by distributor, shelf share against competitors, price deviations. The value comes from reading indicators together. High distribution with frequent out-of-stocks points to a supply problem, not a listing problem. Low distribution in traditional trade in one region points to a weak distributor. Shelf share below market share points to a brand that is not getting display in proportion to its sales.

The resulting decisions are operational: which stores sales reps should visit, which distributor needs a conversation, which display clauses to check with a chain, which price to correct. The price data collected in an audit also give a basis for pricing research, and weighted distribution can be combined with sales data to estimate the market, as described in our article on market sizing and segmentation. Combining data from several sources is part of our data analysis services.

8. Limits and common errors

An audit shows the shelf on the day of the visit. A Tuesday out-of-stock says nothing about Friday, and a single visit a month can under- or overstate problems that fluctuate during the week. Visits are therefore spread across days and times according to a plan, not to the auditor's convenience.

The most common error is a convenience sample: easily reached city-centre shops, where the product is usually better represented. The result is inflated distribution that reassures the team precisely where it should alarm it. The second error is having no universe: without one, "present in 70% of the stores visited" cannot be projected to the market. The third is changing definitions between waves, which makes trends impossible to read. The fourth is using the company's own sales reps as auditors: they have a stake in the result and know the shops, so their data are not independent.

9. The retail audit compared with other data sources

10. What to ask the agency

• The definition of the store universe and the source of the list, with the estimated number of outlets by channel and region.
• The sampling plan: strata, stores per stratum, selection method and weighting to the universe.
• A list of indicators tied to the decision, with a written definition for each.
• The exact list of products and competitors audited, by code and pack size.
• Auditor training and a visit schedule spread across days and times.
• Photographs with date, time and location for every visit, and the share of visits independently checked.
• For monitoring: a stable store panel and identical definitions across waves.
• A report that reads the indicators together and proposes actions by channel, region and distributor.

A retail audit turns a vague question, "why isn't the product selling", into a set of verifiable facts about shelf, stock and price. Designed properly, with a universe, a stratified sample and independent auditors, it becomes the tool a sales team uses to put its resources where they count. To build an audit that fits your category, see our monitoring and testing and mystery shopping services.

See also: Monitoring and testing Mystery Shopping Data analysis and desk research Market sizing and segmentation Mystery shopping: measuring customer experience Pricing research

The Perspicuus research team · 24 September 2025