As of 9 November 2026, a Xero survey of 500 small business owners puts participation in Black Friday and Cyber Monday at 53 per cent, up from 43 per cent a year earlier, while 55 per cent of past participants funded extra stock and staff with credit or deferred payments. The same survey finds that 81 per cent of participating businesses know their maximum profitable discount, but only 30 per cent set prices using real time profit or margin data, and the average maximum discount planned for this year is 34 per cent. For anyone who owns listings, stock cover or promotional pricing, the number to watch is not the 53, it is the gap between 81 and 30.
Participation rose while planning did not
Ten points of extra participation sounds like a healthy market. It is also ten points of extra demand for the same three weeks of warehouse capacity, the same courier network and the same pool of temporary staff. Seven in ten businesses said they felt pressure to compete with larger retailers, and 54 per cent said the period materially affects whether they meet annual profit targets, which suggests the decision to join is being driven by fear of missing out rather than by a margin calculation.
That matters because the cost structure of a small seller does not scale down with the discount. Payment fees, pick and pack labour, outbound freight and returns handling are close to fixed per order, so a smaller basket at a deeper discount compresses contribution from both ends. The businesses that come out ahead are usually the ones that decided in advance which stock lines were allowed to lose money and which were not.
A 34 per cent ceiling priced without margin data
The survey numbers on discount planning are the most useful part of the release. While 81 per cent of participants or intending participants said they know their maximum profitable discount, only 30 per cent used real time profit or margin data to set prices, and 23 per cent of past participants reported lower margins. A further 31 per cent either fell short of the sales they needed or had not set a sales target at all, which means a meaningful share of the group cannot tell whether the event worked.
| Measure | Share | Reading |
|---|---|---|
| Know maximum profitable discount | 81 per cent | Confidence is high in the abstract |
| Set prices with real time margin data | 30 per cent | Execution lags the stated knowledge |
| Reported lower margins | 23 per cent | Depth outran cost cover |
| No sales target or fell short | 31 per cent | No baseline to judge the result |
The credit line behind the stock
Funding stock on credit is rational when the sell through is predictable. It becomes expensive when the discount is deep, because the margin that was meant to repay the facility is the same margin being given away. The Reserve Bank of Australia raised the cash rate to 4.85 per cent this week, the fifth increase this year, with annual inflation at 4.3 per cent, so the carrying cost of that stock has risen at exactly the moment the discount has to be deeper to compete.
Listings that agents cannot read
The survey also asked about AI shopping agents. Thirty eight per cent of businesses expect agents to lift holiday sales, but 54 per cent have taken no steps to make their products discoverable by AI tools, and only 19 per cent have added structured product data. That is a listing problem before it is a marketing problem, because an agent cannot evaluate a product it cannot parse, and it will default to whichever catalogue returns clean attributes.
Structured data is unglamorous: pack size, material, compatibility, dimensions, availability, return window, all in a machine readable field rather than in a photograph of a label. Sellers who have done it describe the same sequence, first the feed, then the copy, then the creative. Doing it in the reverse order produces pages that read well to a human and return nothing to an agent.
Availability beats price in a feed
An agent asked for a specific item will filter on availability before it compares price, because recommending something out of stock is a failed answer. That makes stock accuracy a discovery issue, not just an operations issue. A listing that says in stock when the shelf is empty does more damage than a listing that is slightly more expensive.
A pre-peak checklist for sellers
With the event weeks fixed and the discount ceiling decided, the remaining levers are operational. The checklist below is deliberately narrow: it covers the four items that most often turn a planned promotion into an unplanned loss, and each one is verifiable from data the seller already holds.
FAQ
Is joining Black Friday always better than sitting it out?
No. Participation only pays if the promoted lines can absorb the discount without losing contribution. Sellers without a documented plan are the ones most likely to report lower margins afterwards.
What is a safe discount ceiling?
There is no universal number. The ceiling is the point at which contribution after fees, freight and expected returns reaches zero. It differs by category, basket size and return rate, which is why the 34 per cent average is a description, not a target.
Does structured product data really affect discovery?
Yes. An agent filters on parseable attributes such as pack size and availability before it compares price. Without those fields a product can be listed and still be invisible to the tools shoppers are starting to use.
How should the event be judged afterwards?
Against a baseline set before it started. Realised discount depth, contribution per order and the share of new versus repeat customers are the three measures that separate incremental volume from demand pulled forward from December.
