Linkedin
Facebook
Email
more
Black Friday retail: Protect margins with data and AI
The 2026 Black Friday sales period is shaping up to be the pivotal battlefield that will make or break Australian retailers. Between inflation and cash rate hikes, household budgets are tight and consumers are spending more time hunting for the best deal, with online marketplaces increasing competition for physical retail outlets.
While retailers can expect strong sales at the end of November (the Australian Retail Council forecast Australians will spend $7.2bn over the 2026 Black Friday weekend), a decent chunk of that spending will be on Christmas shopping. This means retailers aren’t necessarily getting additional sales, and the end of November is now a critical trading period.
The decisions made for Black Friday can therefore affect stock availability, fulfilment capacity and working capital through Christmas and Boxing Day.
If we look to the latest data from the Australian Bureau of Statistics (ABS), we can see the major pressure points. In the June quarter, retail sales volumes increased by 0.4% and inventories by 1.0%, while retail gross operating profits fell by 5.2%.
This means retailers have more cash tied up in stock while their margins are under pressure. In this environment, decisions about which products to promote, how deeply to discount and where to position stock need to be based on reliable data.
Even with a successful Black Friday campaign, how much value remains after discounts, fulfilment, returns and inventory costs?
Retailers need to ensure that increased sales will deliver sustainable margins. Achieving this means starting with the data.
Effective data analysis starts with strategy, not technology
When it’s time to crunch some data, people often think they need to buy the newest tech and that the tool will simply deliver the answers they need to form a strategy. In reality, defining your strategic direction should be the first step. A tool – AI or otherwise – can only work within the context it is given.
Before investing in a new tool, retailers should be clear on what information they need, and what decisions to base on the answers.
Without defining the parameters, even accurate analysis can produce answers that are irrelevant to your commercial reality or make recommendations that do not suit your brand.
Most retailers already hold useful information across point-of-sale and ERP systems, supplier records, inventory movements and sell-through history.
Retailers preparing for peak sales season need their data to show what is selling, where margin is being lost and which stock could be left on the shelf after the sale is over. In some cases, management should set the decision rules in advance: if demand reaches a defined level, increase stock allocation; if sell-through falls below an agreed threshold, begin markdowns.
AI-enabled tools can help you with the manual labour of reconciling records, mapping product descriptions and exploring scenarios. A model may identify a price that improves short-term margin, but the tool cannot determine whether it supports the retailer’s brand, competitive position or customer expectations. Those judgments still require human direction from management.
The current trend of dynamic or surge-style pricing is a good example.
A tool might highlight the potential increase to your margin. However, in the current cost-of-living environment, any margin benefit needs to be weighed against the reputational risk if customers perceive pricing as unfair or opaque.
Is the commercial upside worth that risk?
Does every retailer need data analytics to prepare for a sale?
Data analysis is not equally valuable for every retailer or every problem.
A focused data review creates the most value for retailers whose scale and operational complexity make important patterns difficult to see without analysis.
| A focused data review is more likely to add value when… | It may add less value when… |
|---|---|
| The retailer manages many SKUs | The product range is small |
| Sales operate across multiple channels | Operations are relatively simple |
| Demand is seasonal or promotion-driven | There is limited variation in demand |
| Significant cash is tied up in inventory | Inventory exposure is low |
| Reliable transaction history is available | Transaction history is limited |
| Management can change pricing, stock or supplier decisions | Pricing or other decisions are supplier-controlled |
| Complexity makes important patterns difficult to see | Existing reporting already answers the key questions |
Questions to ask your data
For retailers with sufficient data and operating complexity, the useful questions are:
Stock and suppliers:
Which suppliers have missed lead times during prior peaks, where are stock-outs most costly and how should reorder settings change across the peak season?
Range and working capital:
Which slower-moving lines continue to absorb cash once storage, handling and markdown costs are considered, and which should be cleared rather than reordered?
Pricing and fulfilment:
Which products genuinely need a discount, which have sold near full price, and are delivery costs, stock visibility or returns creating more friction than price?
External context:
Internal data tells part of the story, while external data can provide context around seasonality, market conditions, competitor pricing and customer behaviour. What external factors could materially change the decision?
Trustworthy recommendations depend on trustworthy data
AI can produce an answer in seconds. That does not mean the answer is correct, or useful to your business.
For most organisations, the biggest barrier is the quality (and governance) of the data you’re feeding your analytics model.
Common data quality issues we see in retailers include:
- inconsistent SKU identifiers
- missing supplier lead times
- duplicate customer records
- gaps in the historical record of past promotions.
The problem can be summed up simply: garbage in, garbage out.
This is a particular issue with AI-enhanced tools, as they can produce outputs that look accurate. If the underlying data is unreliable, acting on that output could lead to the wrong stock, pricing or purchasing decision.
Data governance and privacy
Privacy and security also need to be designed into the approach.
Customer-level or commercially sensitive records should not be entered into public generative AI tools without an approved use case and appropriate controls. If personal information is not required to answer a question, it should be excluded, aggregated or de-identified before the analysis begins.
Management should be able to trace the data, assumptions and transformations behind a recommendation. Outputs should be tested against known results and reviewed by an accountable business owner before influencing pricing, purchasing or customer communications. AI can accelerate the work, but does not remove the need to challenge the result.
If you cannot explain and test the recommendation, should it influence a commercial decision?
Easy steps for retailers to use data analytics this season
Black Friday is close, with Christmas not far behind, but data analytics doesn’t have to be a whole-of-business transformation.
Here are some accessible suggestions to help retailers unlock value from their data.
Start with one decision
Choose a single question where clearer evidence would make a notable impact. How deep should the markdown be for a category? Should priority suppliers’ reorder settings change? Would moving stock between stores prevent costly shortages?
Check your data.
Identify the relevant fields across POS, ERP, e-commerce, supplier and inventory systems. Confirm who owns the data, where definitions differ and whether important information is missing. A focused question still needs reliable inputs.
Set safe-use boundaries.
Agree which information may be used, where it may be processed and which AI tools are approved. If customer-level information is not needed, leave it out.
Test in a contained pilot.
Compare recommendations with prior sales, actual margins and known events, then test a small category, channel or group of stores using clear measures. Did the change improve margin, reduce stock-outs or release working capital?
Carry the learning into your next campaign.
Use the Black Friday results to refine your Christmas and Boxing Day decisions. Then retain the cleansed data, decision rules and repeatable logic for future planning.
Retaining the value of a data review after the sale is over
A focused Black Friday review should leave more than a campaign report. Reconciled data, reusable analysis and clearer decision rules can support forecasting, inventory allocation, supplier performance, pricing and working-capital management well beyond the peak season.
The key question is therefore not simply how much to discount. It is which decisions should change, based on evidence, before the next promotion is locked in.
The outcome should be a better buying and pricing process, not an AI project for its own sake.
How RSM can help
RSM can work with management to identify where analysis is likely to influence margin, inventory or working capital, and whether the available data is strong enough to support the decision.
Where there is a sound case to proceed, we can help reconcile the data, assess pricing and inventory scenarios, design practical dashboards and develop repeatable tools for management. The objective is not to recommend AI in every case, but to help weigh the financial, operational and reputational trade-offs and build capability that remains useful beyond Black Friday.
HAVE A QUESTION?
Get in touch
Ideas and insights
