Why the smartest peak season retailers get more from the same promo budget

Peak season isn’t one day. For most retailers, it runs from Black Friday through Cyber Monday, then keeps going into December. A few weeks, but they carry a disproportionate share of the year’s revenue, the tightest margins, the heaviest competition, and the highest pressure to discount just because everyone else is.

That’s what makes the promo budget decisions made in this window so expensive to get wrong. A discount that’s slightly too generous, or shown to the wrong shopper, doesn’t just cost you on one order.

Multiplied across peak season traffic, it’s the difference between a record sales number and a record profit number, and this year’s data makes that gap easier to see than ever.

The scale involved makes small inefficiencies expensive fast. The global eCommerce market reached $6.3 trillion in 2024, with 8.8% growth projected, and US eCommerce alone is closing in on $1.2 trillion, up 8.7% year on year.

Peak season sits at the sharp end of that growth, concentrating a huge share of annual spend into a few high-pressure weeks. Get the promo budget right in that window, and it compounds. Get it wrong, and so does the waste.

Peak season 2025, in numbers

Cyber Monday 2025 brought in $14.25 billion in online sales, up 6.7% on the year before, from 75.9 million shoppers. Cyber Week 2025 (the Thanksgiving-to-Cyber-Monday stretch) totalled $44.2 billion, up 7.2%, and around 75% of shoppers now make a purchase during the Black Friday-to-Cyber Monday window. Mobile drove 57.7% of Cyber Monday 2025 sales, worth $8.2 billion, up from 57% the year before.

That builds on an already-record base. US online holiday sales for November and December 2024 hit $241.4 billion, up 8.7% year on year, and Black Friday online sales grew 10.4%, more than six times the 1.7% growth seen in stores.

None of that says which retailers actually made more money from it. That comes down to how efficiently the promo budget was spent, not how much of it got spent.

Blanket discounts are an efficiency problem, not just a margin one

The average discount across Black Friday 2024 was 28% (27% in electronics, 28% in toys, 23.2% in apparel). Everyone drops roughly the same amount, at roughly the same time, for roughly the same reason: everyone else is doing it too.

That’s not a strategy, it’s a reflex, and reflexes are rarely efficient.

Free delivery is still the biggest purchase motivator, cited by 50.6% of shoppers, ahead of coupons and discounts at 39.3%. Shoppers aren’t only asking for money off, they’re asking for friction to go away, and a discount can create friction of its own: 32% of shoppers abandon their basket if a code doesn’t work, and 36% abandon over an unexpected delivery fee at checkout.

Spend that isn’t targeted has to be spread across every shopper to reach the ones who need it, which means most of the budget goes to people who were buying anyway. That’s the efficiency problem.

A 20% discount on a 40% margin item halves your profit on that sale, so you need double the volume just to land the same total profit, and a blanket rate has no way of knowing which shoppers actually needed that push.

What a more efficient promo budget looks like

Retailers who moved from calendar-based promotions to AI-driven targeting saw a 3% annualised margin improvement within three months. Targeted promotions deliver a 1–2% sales lift and a 1–3% margin improvement compared with blanket discounts, and 65% of shoppers name a targeted offer as a top reason they bought. That’s a better outcome on both sides of the ledger, sales and margin, from the same or a smaller pot of money.

Expectation backs this up. 71% of shoppers now expect some level of personalisation, and 76% say they get frustrated when a brand doesn’t deliver it. A blanket discount, by definition, can’t be personalised. It’s the same offer for the shopper who was always going to buy and the one who needed a reason to stay, which is exactly where an efficient budget starts leaking.

SportsShoes saw this play out directly. Instead of a blanket promotion, they used small Stretch & Save incentives for shoppers close to a spending threshold, alongside exit-intent offers for genuine abandoners. Result: a 10% lift in revenue per user and a 6% rise in conversion rate, without discounting shoppers who didn’t need the nudge, which is the same budget working harder rather than a bigger budget.

Loyalty data is a targeting signal you already own

Before reaching for a new discount, it’s worth checking what you already know about a shopper. Loyalty programmes are among the most efficient targeting signals available because the data already exists and the shopper has told you they’re engaged.

72% of consumers say loyalty programmes make them more likely to spend with a brand, and 56% say they actually spend more once they’ve joined. Peak season is when that engagement peaks too: new loyalty sign-ups rose 119% during BFCM 2024 compared with the rest of the year. A loyalty member who’s already told you they plan to keep buying doesn’t need the same incentive as a first-time, price-sensitive visitor.

Treating them the same isn’t generous, it’s inefficient.

The principle underneath all of this: incrementality

Efficient promo spend comes down to one question, asked of every offer: would this shopper have bought anyway? If yes, the discount was pure margin given away. If not, it earned its cost. That question is called incrementality, and it’s the difference between an offer that looks like it worked and one that actually did.

Redemption rate can’t answer it. High redemption is easy; you just give away enough margin, and everyone who was already buying will happily take the discount, too. A valid incrementality test can be conducted, whether that’s a control group, a holdout segment, or a split test comparing offer formats.

Our own research with IMRG, tracking 35 retailers over two years, found a clear correlation: the worse the trading month, the more discounting activity across the industry, with little evidence that most of it was doing anything incremental.

Dynamic pricing, closer to the airline and hotel model, is gaining traction for the same reason: it’s another way to ask the incrementality question in real time, rather than setting a single rate for everyone and hoping.

Before you commit peak season budget

Three checks are worth running now, while there’s still time to act on the answer.

Measure spend against incremental outcomes, not redemption

A high redemption rate tells you the offer was popular. It doesn’t tell you it was needed.

Segment before you discount

A blanket rate assumes every shopper needs the same push. Most don’t, and treating them the same is where the budget stops being efficient.

Set your allowable discount before peak starts

Decide the number and what counts as a genuinely incremental result in a planning meeting, not mid-campaign under pressure.

Peak season is bigger than it’s ever been. The retailers who profit most from it won’t be the ones who spent the most on promotions. They’ll be the ones who spent it on the shoppers who actually needed it.