NEW RESEARCH – Retail stock audits: 7 critical questions to ask

You know that retail stock audits consume huge amounts of time, money and attention.

Stores can be disrupted, colleagues diverted from customers, specialist counting teams mobilised and results examined by finance, audit, operations and loss prevention.

Yet, one basic question is often left unasked: are retailers counting inventory because the process delivers value, or because it’s what they have always done?

Providing an answer for that is at the heart of a new ECR Retail Loss research project.

On 15 September 2026, Professors Aris Syntetos, Christoph Glock and Yacine Rekik will introduce their study and invite retailers to help shape it.

The aim is to reshape inventory counting, not to replace it.

Retail stock audits support financial reporting, provide assurance, reveal differences between physical and recorded inventory and can help improve availability.

However, retailers need to distinguish what they need to do from what has grown through convention, organisational preference or legacy process.

Join us for the online research launch: Why Do Retailers Count Inventory? Tuesday 15 September 2026, 1:00–2:30 pm BST. Attendance is free.

Retail stock audits matter, but their purpose is not always simple

A stock count can appear to be a straightforward exercise: establish what is physically present, compare it with the system and investigate or post the difference.

In practice, that count may be expected to do several jobs at once. It may:

• Support statutory accounts and external audit;
• Establish or validate the value of inventory;
• Produce a shrink result for the loss prevention team;
• Correct stock records for replenishment;
• Test the performance of stores or counting providers;
• Or reveal weaknesses in receiving, transfers, returns and other operational processes.

Those purposes overlap, but they are not identical.

The information needed for a financial valuation may not be the same as the item-level information needed to correct a stock file.

A count designed to estimate the value of a department may not provide the detail required to investigate the causes of shrink.

A process that satisfies an assurance requirement may not be the most effective way to improve on-shelf availability.

This is why retail stock audits cannot be understood only as a counting exercise.

The purpose influences the scope, frequency, level of detail, method, ownership, assurance process and use of the data.

If those decisions are not explicit, retailers can end up buying one count and expecting it to answer several different questions.

Ownership can be just as fragmented as purpose:

• Finance may own the final inventory value;
• Operations may organise the count;
• An external provider may collect the data;
• IT may control the adjustment;
• And loss prevention may investigate the variance.

When each function owns only one stage, it can be unclear who is accountable for the overall result or for improving the process next time.

Retail stock audits are more likely to create value when responsibilities, handovers and decision rights are explicit.

The design of retail stock audits should therefore include what happens before and after counting, not just what takes place on the shop floor.

The business case reaches far beyond compliance

Reliable inventory information underpins replenishment, product availability, fulfilment and financial decision-making.

When the system believes stock is present but the shelf is empty, replenishment can be delayed, and a hidden stockout can persist.

When physical stock exceeds the recorded quantity, unnecessary orders, excess inventory or avoidable markdowns may follow.

Our previous research shows why the issue deserves attention.

A study of inventory record inaccuracy metrics finds that more than 60% of inventory records can be wrong in a retail store, depending on how accuracy is defined.

That research also shows that retailers use different definitions and tolerances.

Two organisations can report impressive-looking accuracy figures while measuring materially different things.

Other ECR Retail Loss studies found sales uplifts between 4% and 11% in specific retail settings after inaccurate records were corrected.

A recent peer-reviewed grocery study, by the same authors, found that the effect of an inventory audit is concentrated among products where the system overstated the stock actually available.

While these results don’t promise that every retail stock audit will produce the same uplift, they do show that correcting the right inaccuracies can have commercial value, particularly when a false system record prevents replenishment.

The wider lesson is that value does not come from counting for its own sake. Rather, it comes from what the retailer learns, corrects or changes as a result.

What are retail stock audits designed to achieve?

Before choosing how much to count or how often, a retailer needs a clear answer to a more fundamental question: what decision is this audit intended to support?

The new research will explore questions including:

• Why do retailers conduct stock audits beyond any minimum legal, regulatory or financial reporting requirements?
• Who decides the frequency, scope and method of the count?
• What constitutes a credible result, and who assures the count itself?
• What happens to the data once the count has been completed?
• How are products, locations or discrepancies selected for recounting?
• How do retailers balance compliance with shrink reduction, inventory accuracy, availability and operational performance?
• Which practices appear effective enough to inform emerging guidance for the sector?

These are governance questions as much as operational ones.

Decisions may involve finance, internal and external audit, store operations, supply chain, IT, loss prevention and external inventory counting providers.

Each function may approach the count with a different objective and a different view of what “accurate” means.

Making those assumptions visible is part of the opportunity.

A retailer may discover that different teams commission overlapping counts, that detailed item-level data is collected but not used, or that a result is treated as authoritative without a clear process for checking count quality.

What happens to the count data?

The physical count is only one stage in the process.

What happens next can determine whether the exercise becomes a compliance cost or a useful operational intervention.

• Some retailers use the data to adjust or “true up” their inventory records;
• Some pass variances to loss prevention teams for analysis;
• Others may use results to examine store or category performance, identify process failures, challenge a counting provider, select items for recounting or test the accuracy of prediction models.

The required granularity can therefore vary significantly.

A value-level count may be sufficient for one purpose, while another requires SKU-level detail. Greater granularity can cost more to collect and process, so the choice should follow the intended use.

Buying more detailed data has little value if it is never analysed or acted on.

Equally, a coarse result may save money at the counting stage but leave the business unable to correct records or understand what drove the variance.

The research will examine these differences rather than assume there is a single correct model.

It will ask how retailers turn count results into decisions, who owns the follow-up and whether the process creates a useful feedback loop.

Who checks the checker?

Retail stock audits are intended to create confidence, but the count itself can contain mistakes.

Products can be missed, counted twice, placed in the wrong location or recorded against the wrong SKU.

Movement during the count, unclear cut-off rules and inconsistent methods can introduce further uncertainty.

That raises an important question: how does a retailer know that the new number is more reliable than the one it replaces?

Retailers use a range of assurance practices.

These may include supervisor checks, sample recounts, exception thresholds, comparisons with historical results and targeted reviews of unexpected variances.

Yet the basis for choosing what to recount, and what level of difference is acceptable, may not always be explicit.

The new study will explore how retailers assess count credibility and how they select SKUs for eventual recounting.

It will also bring in the perspectives of external inventory counting providers, auditors, consultants and legal or regulatory experts.

The aim is to understand the assurance system around the count, not merely the actions of the people scanning products.

Technology changes the choices available

Established retail stock audits were designed in a world with less data and less computing power.

Retailers now have richer transaction histories, more frequent inventory signals and analytical methods capable of identifying records that are particularly likely to be wrong.

Our work on predicting wrong inventory records explored how data can direct attention towards higher-risk items.

While our inventory accuracy maturity model places proactive, integrated accuracy management above isolated corrective action.

This does not automatically remove the need for full counts or the assurance they provide.

Legal frameworks, accounting requirements and external audit expectations must be understood in the relevant jurisdiction and reporting context.

Predictive models also need validation, governance and a practical response when they flag a problem.

But technology makes a wider range of approaches possible. Retailers can ask whether:

• Some counting effort could be targeted by risk;
• Full counts and cycle counts could be combined differently;
• Data from RFID or other inventory systems can complement physical verification.

Risk-based inventory counting may allow attention to move towards the records where an error would be most likely or most damaging.

It is not that one method is inevitably better, but that the choices should be made consciously, with a clear understanding of the assurance required, the operational objective and the value of the information produced.

Would you design the same process today?

Many retail processes persist because they are embedded in budgets, calendars, contracts and audit routines.

That does not make them wrong. However, it does make them worth examining.

If a retailer were designing its stock-auditing process today, it might start with questions such as:

1. What do we need to know, and for which decisions?
2. Which parts of the process are required, and which are internal choices?
3. Where do inaccuracies create the greatest financial, operational or customer harm?
4. What level of assurance do finance, auditors and operational teams genuinely need?
5. Could a combination of full, cycle, event-driven and risk-based counts provide better information with less disruption?
6. Are we collecting data at the right level of detail?
7. Do we act on the result, learn from it and improve the processes that created the error?

These questions shift the conversation from “How often do we count?” to “What are we trying to achieve?”

That is a more demanding discussion, but it is also where the potential value lies.

A new study of retail stock audits

Our 12-month research project will investigate how retail stock audits are designed, governed, conducted and evaluated.

It will combine a review of academic, practitioner, legal, regulatory and industry sources with semi-structured interviews across four stakeholder groups:

• Retail companies;
• Inventory counting service providers;
• Auditing and professional services firms;
• And legal or regulatory experts.

We anticipate conducting 30 45-minute interviews (with the final number guided by whether the researchers continue to hear materially new insights).

The legal and documentary review will pay particular attention to Germany, Greece, the UK, the USA and Australia.

Interview participation, however, is neither restricted to those countries nor to a particular retail sector.

The research team wants to understand differences across organisations, sectors and jurisdictions, as well as examples of distinctive or effective practice.

The work will be led by three academics with complementary expertise:

1. Professor Christoph Glock of the Technical University of Darmstadt;
2. Professor Yacine Rekik of emlyon business school;
3. And Professor Aris Syntetos of Cardiff Business School.

They will produce:

• An ECR Retail Loss report;
• A conceptual framework for understanding the decisions that shape retail stock audits;
• Practical guidance or emerging best-practice recommendations;
• An academic paper;
• And an ECR Retail Loss workshop to discuss the findings.

The study is deliberately exploratory to understand why practices differ and what the evidence supports.

Why retailers should take part

The research will be stronger if it reflects the real choices retailers face.

This includes organisations using traditional annual counts, retailers with established cycle-counting programmes and businesses experimenting with RFID, predictive analytics, specialist teams or hybrid approaches.

Participation offers retailers several potential benefits:

• Benchmarking: compare the logic behind your approach with practices used elsewhere, without identifying participating organisations publicly.
• Early insight: gain access to emerging themes and contribute to discussions while the work is still developing.
• Practical influence: help ensure that the eventual framework addresses real constraints rather than an idealised counting process.
• Cross-functional learning: use the research questions to bring finance, audit, operations, supply chain, IT and loss prevention into the same conversation.
• A low data burden: the proposal does not anticipate substantial data provision beyond an interview.

Organisations may share policies or process documents where useful and appropriate, but this is optional.

Any shared information will be treated confidentially, with participating organisations anonymised in the academic output. Non-disclosure agreements can be signed where required unless suitable agreements are already in place.

Businesses do not need to arrive with a perfect model for retail stock audits.

The study also needs to hear where responsibilities are fragmented, where count data is underused, where assurance is difficult or where the current process is simply accepted because changing it feels complicated.

Frequently asked questions about retail stock audits

What is a retail stock audit?

Retail stock audits compare physical inventory with the quantity or value recorded by the retailer.

Depending on their purpose, they can support financial reporting and external assurance, identify shrink or inventory inaccuracies, correct stock records and inform operational decisions.

The scope may range from a full count to a targeted sample or cycle count.

Why do retailers count inventory?

Retailers count inventory for several reasons:

• Support accounts and audit requirements;
• Establish inventory value;
• Measure shrink;
• Correct inaccurate records;
• Improve replenishment and availability;
• And examine operational performance.

The new ECR Retail Loss research will investigate how retailers prioritise these objectives and why their approaches differ.

Could risk-based inventory counting replace full retail stock audits?

That is an open question, not a conclusion of the new study.

Predictive analytics can help identify records that are more likely to be wrong, but retailers must still consider legal, accounting, audit and governance requirements.

The research will explore whether targeted, full and hybrid approaches can play complementary roles.

How can a retailer participate in the research?

The first step is to join the free online research launch on 15 September 2026.

The project team will then recruit retailers and other stakeholders for confidential interviews of approximately 45 minutes.

Substantial data provision is not expected, although organisations may choose to share relevant policies or process documents.