Every promotion produces a sales spike. The harder question is what share of that spike would have happened without the discount.

We surveyed 1,000 U.S. consumers in November 2026 to see promotions from the shopper’s side. Three in four (76%) told us they would have bought the same product at full price at least some of the time. Nearly as many (73%) said their last promotion still changed what they did. Both findings hold at once, and the space between them is where promotional budget gets won or lost.

Here’s what we found, and what it means for how you measure promotions.

Promotions change behavior, but not always in the way the dashboard suggests

When we asked shoppers how their most recent promotion affected them, the answers spread out:

  • 25% bought sooner than they planned
  • 19% bought something they hadn’t considered
  • 15% switched from their usual brand
  • 14% bought more than they intended
  • 23% bought what they had already planned, and just paid less

Only some of those outcomes are new revenue. A brand switch or a purchase the shopper hadn’t considered is demand the promotion created. A purchase made two weeks early is mostly the same sale on a different date. And the 23% who were already planning to buy received a discount on revenue the brand would have earned anyway.

A standard sales report counts every one of those purchases as promotional lift. That’s why a promotion can look like a winner while reducing profit.

Shoppers have learned to wait

79% of consumers delayed a planned purchase in the past year because they expected a sale. Half do it every few months or more.

Ask the same shoppers directly and 61% say they’re likely to pay full price rather than wait. The gap between stated intent and reported behavior is a warning for anyone forecasting from surveys alone, and a bigger one for anyone reading promotion results at face value.

When shoppers hold off, the weeks before a sale go quiet and the sale week absorbs demand that was already there. The promotion looks stronger than it is, and the baseline it’s compared against looks weaker than it should. Repeat that cycle a few times a year and the business starts to depend on discounts to hit numbers it used to hit at full price.

Loyalty is negotiable, especially for Gen Z

61% of consumers switched away from their usual brand for a better discount every few months or more. Faced with a hypothetical $100 purchase from a favorite brand, 44% said a 25%-off offer from a comparable competitor would change their mind.

Younger shoppers move fastest. 51% of Gen Z said a competitor’s discount would sway them, compared with 34% of Baby Boomers. Gen Z is also far more likely to buy on impulse from their feed: 54% made an unplanned purchase prompted by social media, versus 7% of Boomers.

For a brand, this cuts both ways. A discount that pulls a shopper from a competitor is real growth. The identical discount sent to a loyal customer who was about to reorder is margin given away. The offer is the same and the economics are opposite, which is why promotion results have to be read by audience.

Frequent discounting has a long tail

Only 29% of shoppers say frequent discounts make a brand look like better value. 48% report at least one negative effect: they expect future discounts, doubt the regular price or question product quality.

That cost doesn’t show up in the campaign report. It shows up months later as softer full-price weeks and a baseline that needs more promotional support to hold steady.

In-store still leads, and the channel mix depends on who you’re reaching

In-store promotions and displays are the top driver of unplanned sale purchases (44%), followed by social media (34%), retailer websites and apps (29%), online ads (25%), search (24%) and email (23%).

The mix shifts by generation. Boomers lean on in-store (52%). Gen Z leans on social (54%) and is the only group where influencer content registers meaningfully (18%). A single blended promotion ROI hides those differences.

How to measure what your promotions really do

None of this means brands should stop running promotions. It means promotions deserve the same causal measurement as any other marketing investment. Five practices help:

  1. Measure against a true baseline. Estimate what would have sold with no promotion, controlling for seasonality, pricing and other marketing. Marketing mix modeling built on causal methods does this across all channels at once.
  2. Run holdout tests. Offer the promotion in some regions or audiences and withhold it in comparable ones. Geo-based incrementality testing gives a direct read on how many purchases the offer created.
  3. Watch the weeks after. A dip after the spike is a sign of pull-forward. Net it out before calling the promotion a success.
  4. Break results out by audience and channel. New versus existing customers, social versus email, Gen Z versus Boomers. Incrementality varies widely across all of them.
  5. Count margin, not just revenue. Incremental profit is the number that matters. A promotion can add revenue and still lose money.

As Lifesight Co-Founder and CEO Tobin Thomas put it, “Sales lift alone can’t tell brands whether a promotion generated a new purchase or simply discounted one a shopper would have made anyway. Without measuring incremental impact, marketers risk rewarding existing demand rather than investing in sustainable growth.”

Get the full report

The full survey results include generational, income and household breakdowns for every question, plus a framework for measuring promotional incrementality. Download The Discount Paradox: 2026 Consumer Promotions Report

Want to know how much of your promotional lift is incremental? Book a demo to see how Lifesight’s Agentic Unified Marketing Measurement Platform combines causal MMM, geo-based incrementality testing and incrementality-adjusted attribution in one view.

Methodology: Lifesight surveyed 1,000 U.S. consumers ages 18+ via Dynata from November 5 to 11, 2026. The overall margin of error is ±3.1 percentage points at the 95% confidence level. Generational subgroups include 192 Gen Z and 200 Baby Boomer respondents (margins of error of roughly ±7.1 and ±6.9 points).