Ahead of last year's holiday season, Deloitte's survey found consumers planning to spend about 10% less than the year before, with Gen Z intending to cut back by a third. That season, US holiday sales passed one trillion dollars for the first time in recorded history.
Sentiment stopped predicting spending
That gap isn't a one-off, and it's the single most important thing to understand before reading any holiday forecast.
Consumer sentiment has been sitting near historic lows — the University of Michigan index at levels you have to go back to 1980 to match. More than three-quarters of consumers surveyed expected higher prices; well over half expected the economy to weaken, the least optimistic reading since that survey began in 1997.
And spending kept setting records. US retail sales grew close to 4% in 2025 to around $5.4 trillion, and the NRF's 2026 forecast, built with Oxford Economics, projects 4.4% growth to roughly $5.6 trillion — comfortably above the pre-pandemic ten-year average near 3.6%.
The NRF's own phrasing Sentimentally weak, but fundamentally sound. People believe the economy is deteriorating and spend according to their own circumstances instead — and their own circumstances have held up.
The mechanism is straightforward once stated. Sentiment surveys ask how you feel about the economy. Spending follows how your household is doing. Someone can be convinced things are going badly nationally while their own income is stable, their job is secure and their balance sheet is fine — and the second set of facts is what determines whether they buy.
So the practical instruction is blunt: do not plan inventory, budget or discount depth on the basis of stated intent. It has been wrong, in the same direction, for several consecutive years.
What the fundamentals actually say
Set the surveys aside and the supporting picture for 2026 is reasonably constructive, with named risks.
| Factor | Reading |
|---|---|
| Retail sales forecast | +4.4% for 2026, above the ~3.6% long-run average |
| Real versus nominal growth | Goods inflation expected in a lower band, so a meaningful share is real volume |
| Household balance sheets | Cited as a core support for continued spending |
| Labour market | Weaker, with muted employment growth — but unemployment expected below 4.5% |
| Early-year stimulus | Larger tax refunds provided a first-half boost |
| Inflation path | Above target, with relief anticipated later in the year |
| Excluded from forecasts | Significant geopolitical disruption is explicitly not priced in |
That last row deserves attention. Forecasts of this kind are constructed on the assumption that nothing dramatic happens, and the organisations publishing them say so. A forecast is a base case, not a prediction, and the honest way to use one is as an input to a range rather than a number to plan against.
The bifurcation is the real story
An aggregate forecast of +4.4% describes an average consumer who doesn't exist, and the distribution matters far more than the headline.
By income: higher-income households have been driving the majority of spending growth. That's a concentration, not a broad-based lift, and it means brands serving mid-market and value segments face a different market from the one the national number implies.
By generation: the intended pullback was heavily skewed. Gen Z planned reductions of around a third year on year, millennials low double digits, while Gen X was the only cohort intending to increase spending. Even allowing for intent overstating the pullback, the relative ordering is informative.
Which produces a practical instruction: work out which side of that split your customers sit on before applying any national forecast to your plan. A brand selling to affluent forty-somethings and a brand selling to twenty-two-year-olds are reading the same headline about two different markets.
Loox
Collects and displays photo reviews building social proof that drives more sales
Best for: Product Reviews
They'll spend — but they'll make you justify it
The most reliable finding across the survey data isn't about how much people will spend. It's about how they'll go about it.
Value-seeking behaviour is near-universal: roughly nine in ten intending to search for deals, over three-quarters willing to trade down on brands and retailers, and substantial numbers planning to reuse, recycle, or make gifts rather than buy them.
That's not a demand problem. It's a justification problem — and it changes what you should optimise.
Willingness to trade down is the sharpest signal here. More than three-quarters saying they'd switch brand or retailer to save money means loyalty is unusually soft this season. That cuts both ways: your customers are more winnable by competitors, and their customers are more winnable by you. Acquisition economics improve and retention economics worsen simultaneously.
Deal-seeking doesn't mean deepest discount wins. It means the offer has to be legible. A clear, comparable, believable saving beats a larger one that requires arithmetic or arrives with conditions. Bundles and thresholds work well here precisely because they let a shopper feel resourceful rather than merely subsidised.
Trading down is defeated by evidence, not by price matching. Reviews, specifics, durability claims that can be checked — the things that make a shopper conclude the cheaper option is a false economy. That's largely a product page problem rather than a pricing one.
What to watch instead of surveys
Since stated intent has been unreliable, here are behavioural signals that lead the season and come from data you already hold.
- Wishlist and saved-item volume through September and October. People park things they intend to buy on discount. Rising volume is demand accumulating, not demand disappearing.
- Email list growth rate in autumn. Signing up before a sale season is a purchase intention with a delay attached — one reason list building in September is worth more than list building in June.
- Returning-visitor share. People researching now and buying later show up as repeat sessions without conversions. That pattern looks like weak performance and is often the opposite.
- Full-price average order value. If it's holding, price sensitivity is lower than the surveys imply for your specific audience.
- Product page versus category page traffic. A shift toward specific products indicates people have narrowed their consideration and are waiting for a trigger.
The common thread: these measure what people do rather than what they told a surveyor, and they describe your customers rather than a national average. Both properties make them more useful than any published forecast for the decision you're actually making.
Where the trade-down risk actually lands
One refinement worth making, because "everyone is trading down" is too blunt to act on.
Trading down concentrates in categories where the shopper can't easily tell the difference before buying — commodity goods, basics, anything where the specification is the whole story. It's far weaker in categories where the purchase carries meaning: gifts intended to signal thought, items bought for someone else's judgement, anything where being seen to have economised is itself the cost.
So if you sell giftable products, the trade-down statistic is less threatening to you than the headline implies. If you sell utility products, it's more threatening. Segment your catalogue on that axis before deciding where to discount, because uniform markdowns across both types give away margin on the items that didn't need it.
The other place it bites is cart abandonment: a shopper comparing your item against a cheaper alternative in another tab is exactly the scenario a well-built cart recovery flow is designed for, and its value rises in a soft-loyalty season.
Timing signals worth acting on
Two structural patterns have held steady enough across recent seasons to plan around.
The season starts earlier than the name suggests. A substantial share of BFCM-period revenue lands before Black Friday itself, and Thanksgiving Day traffic reaches a large fraction of Friday levels. Early access, particularly for existing customers, is where a meaningful slice of the opportunity sits — and it's operationally easier than competing on the peak days.
The calendar is tight this year. Black Friday 2026 falls on 27 November with Cyber Monday on the 30th, which leaves roughly 25 days to Christmas. Shipping cut-offs arrive sooner relative to the sale and December has less room for a second push. The operational consequences of that compression are worked through in our 90-day BFCM timeline, which is the companion piece to this one.
Where the money is going, not just how much
A distributional note that matters for budgeting as much as the spending forecast does.
Media investment has been shifting toward environments that can connect an ad to a purchase inside their own ecosystem — the reason retail media budgets keep growing — and peak season concentrates that effect, because measurable performance is what survives a budget review in a nervous quarter.
The practical implication is competitive rather than strategic: if your category's spend is consolidating into a few auction environments, expect those auctions to be more expensive in November than your historical CPMs suggest. Build that into the plan rather than discovering it in week one.
A caution about the forecasts themselves
Worth stating plainly, given how confidently these numbers circulate.
Retail forecasts are published months ahead and revised. The 2026 annual figure is an annual figure — holiday-specific forecasts typically arrive in the autumn and often differ. Survey intent, as established, systematically overstates pullback. And macro forecasts explicitly exclude tail risks that have repeatedly not stayed excluded.
There's also a source-quality issue in this category worth naming: a great deal of BFCM forecasting content is published by companies selling marketing software to retailers, whose interest runs toward "this season will be enormous, buy our tools." The NRF, Deloitte and Adobe figures are independent and methodologically transparent. Vendor blog projections frequently aren't, and the two get quoted interchangeably.
Use forecasts to set a range and a contingency, not a plan. If your plan only works at the top of the range, it isn't a plan.
That discipline matters most for smaller operators, who feel a demand miss immediately and have less room to absorb one — and who also have the compensating advantage of being able to change direction mid-season, as set out in how smaller brands compete against bigger competitors.
What this means for your season
- Plan for volume, prepare for value-seeking. Demand is likely to be there. It will arrive looking for a reason to feel good about the purchase.
- Segment before you forecast. Your cohort mix determines whether the national number is optimistic or conservative for you.
- Lean into acquisition. Softer loyalty means competitors' customers are more winnable this season than in a typical year. That's a real, time-limited opportunity.
- Defend retention deliberately. The same softness applies to your own base, and the mechanism that protects it is the automated flows and post-purchase experience you build now rather than in December.
- Watch behaviour weekly from now. Wishlists, list growth, returning visitors. Adjust on data rather than on headlines.
- Set a floor, not just a target. Know what you'd do if demand comes in 15% below plan, and decide it before the season rather than during it.
One measurement note for the season itself: peak-period attribution is unusually unreliable, with multiple channels touching the same buyer within hours and platform-reported conversions exceeding reality. Judging channels against one another during BFCM will mislead you, for the reasons in why attribution keeps getting harder. Compare total revenue to total spend and leave the channel autopsy for January.
The short version
Consumers said they'd spend 10% less and then set a record — so stated intent is a poor basis for planning, and has been for several years running. The fundamentals underneath are constructive: growth forecast above the long-run average, with a meaningful share of it real rather than inflationary. But the aggregate hides a sharp split by income and generation, so check which side your customers sit on before applying any national number. Expect the demand to arrive value-seeking rather than absent: nine in ten hunting deals and three-quarters willing to trade down, which softens loyalty in both directions. And watch your own behavioural signals weekly, because they describe your customers rather than the country's.
Planning peak season on headlines or on your own numbers?
We build e-commerce programmes that win the trade-down season and keep the customers afterwards.
Explore E-commerce Marketing →Frequently asked questions
Will consumers spend less this holiday season?
They will probably say so and then spend anyway, based on the recent pattern. Deloitte's survey ahead of the 2025 season found consumers intending to spend around 10% less year on year, with Gen Z planning a 34% reduction. Actual holiday sales that season passed one trillion dollars for the first time on record. Stated intent has become a poor predictor of behaviour, which is why the NRF characterises the current consumer as sentimentally weak but fundamentally sound — sentiment near historic lows, spending steady.
What is the retail sales forecast for 2026?
The NRF, working with Oxford Economics, forecast US retail sales growth of 4.4% in 2026, reaching around $5.6 trillion on its core definition that excludes autos, petrol and restaurants. That compares with a pre-pandemic ten-year average near 3.6%, so it represents a stronger than typical year. Because goods inflation is expected to stay in a lower band, a meaningful share of that growth should be real volume rather than price increases. The forecast does not price in significant geopolitical disruption.
Why has consumer sentiment stopped predicting consumer spending?
Sentiment surveys capture how people feel about the economy in general, which has become increasingly detached from their own household circumstances. Someone can believe the economy is deteriorating while their own income is stable, their employment is secure and their household balance sheet is sound — and spending follows the second set of facts rather than the first. The University of Michigan sentiment index has sat near levels last seen in 1980 while retail spending set records, which is about as clear a decoupling as the data offers.
Which shoppers are actually pulling back?
The picture is heavily bifurcated rather than uniform. Higher-income households have been driving the majority of spending growth, while younger cohorts show the sharpest intended reductions — Deloitte's survey found Gen Z planning to cut spending by around a third and millennials by low double digits, with only Gen X intending to spend more. This matters more than the headline forecast, because an average that blends a spending high-income segment with a retrenching young segment describes neither accurately.
What should retailers watch instead of sentiment surveys?
Behavioural signals from your own data, which lead the season more reliably than published intent. Useful ones include wishlist and saved-item volume, email list growth rate through autumn, returning-visitor share, average order value trend on full-price items, and traffic to product pages versus category pages. These reflect what people are actually doing rather than what they told a surveyor, and they are specific to your customers rather than to a national average that may not describe your segment at all.
