Tipping fatigue in 2026 — what the data actually shows.

Pew says tipping is expected in more places. Bankrate says fewer people are tipping. Both are true. The story underneath is the touchscreen.

The two most-cited numbers on American tipping in 2026 sound like they contradict each other. The Pew Research Center, in its November 2023 survey of 11,945 adults, reported that 72% of Americans say tipping is expected in more places now than five years ago. Bankrate, in its June 2025 survey, reported that the share of Americans who say they "always" tip at sit-down restaurants fell from 77% in 2019 to 65% in 2024 — the lowest figure since Bankrate began the survey in 2018.

So which is it? More tipping or less?

Both, and the explanation is the iPad on the counter. There are two American tipping economies now, and the data is telling you about both at the same time.

The old economy: still steady

Sit-down restaurant tipping, the canonical case, has not collapsed. Square's aggregate transaction data — every tip processed through a Square terminal at a full-service restaurant — shows the median tip steady at 19.4% in Q4 2024, essentially flat against 19.6% in 2019. Reservations platform OpenTable's diner sentiment surveys put the same number at 18%. Bankrate's "always tip 65%" headline is misleading in isolation: the median tip among those who tip is up slightly, even as the share who "always" tip drifts down. The decline is in the consistency, not the amount.

If you eat at a sit-down restaurant once a week, the customary tip on your bill is what it was in 2019. 18–22% is still the range. The cooks and bussers still get their cut. Nothing has structurally changed about the relationship.

The new economy: prompted, contested

What has changed is the universe of prompted transactions. In 2019, you ordered a coffee at a counter and were charged $4.50; the tip jar was glass and the prompt was gravity. In 2024, you ordered the same coffee and the iPad rotated to show three buttons: 18%, 22%, 25%. The fourth button, "No tip," is smaller and on the far right.

This is where Pew's 72% lives. It is empirically correct that more transactions present a tip prompt than five years ago — Square's own reports document the rollout of suggested-tip features to bakeries, drive-throughs, dry cleaners, and self-checkout kiosks. The norm did not change. The infrastructure did.

Pew also asked whether respondents actually tip at these new prompted contexts. Only 25% of US adults said they always or often tip at counter service. Only 12% at fast-food. Only 13% at coffee shops. Cross-referenced against the Square data, this looks like an equilibrium settling: the prompt is everywhere, the tip is sometimes, the customary amount at a counter — when it happens — is still $1 or 10–15%, not 25%.

What "fatigue" actually means

"Tipping fatigue" is a tabloid coinage and it gets used to mean two different things. Sometimes it means resentment of being asked. Sometimes it means actually reducing the tip. The first is up sharply; the second, the harder thing to measure, is barely up at all in established service categories.

The Pew survey asked respondents directly what they think about touchscreen tip prompts: 32% said the prompts are "annoying" and another 41% said they are "inappropriate" in at least some contexts. Bankrate's qualitative responses skew the same way — the open-ended complaints are not "tipping is too expensive," they are "I should not be asked to tip when I picked up the food myself."

This is a useful diagnosis if you're trying to figure out what to do at the screen. The honest answer is the unsexy one: tip the customary amount for the situation you're in, not the situation the screen suggests. If the situation is sit-down, that's 20%. If it's delivery, 15–20% with a $5 floor. If it's counter coffee, $1 if you want, zero if you don't, and the screen's 25% preset is a merchant marketing choice, not a moral test.

Where this lands in 2026

The pressure on the customary tip — at sit-down, where it actually matters for the worker — is structurally upward. Minimum wage in tip-credit states has not moved much. Card-processing fees have. The cost of running a restaurant in 2026 is the highest it has ever been in inflation-adjusted dollars, and tipping is what backstops the model. None of the headlines about "fatigue" have softened the floor under the sit-down norm.

The pressure on the touchscreen tip is downward, and the discomfort is the source of it. The screen prompt at the bakery counter is not stable equilibrium. Either the screens will quietly drop the 25% preset (some chains already have) or the tip behavior at counters will eventually move closer to it. Three years from now, one of those two things will be true. I would bet on the first.

Until then, the heuristic is: the customary tip is the answer the data has been giving for ten years, not the answer the screen suggests today. A guide that tries to update either upward to match the iPad or downward to match the resentment will be giving you a number that doesn't match anyone's actual behavior. The numbers we publish are the median, with sources cited and reviewed annually. They have moved less than the discourse has.

Sam Levenson built TipCalc and writes about service, money, and the small ways software changes the way we behave at the register.