A pre-checked box, a reminder at the right hour and a line that says most customers pick the middle plan are all nudges. Marketers use them because they cost little and leave every option open. The evidence says they work, with smaller effects than early studies claimed. This entry covers what nudge marketing is, how well it works and how a loyalty team can use it without crossing a line.
What is nudge marketing?
Nudge marketing is the use of small changes in how choices are laid out to steer customers toward an action. Every option stays open, and nobody is paid to act. The term comes from Richard Thaler and Cass Sunstein. They define a nudge as any feature of the choice setup that shifts behavior in a predictable way. It bans no option and does not change the economics much (Hansen and Jespersen, 2013).
A nudge is not a discount and not a rule. A discount changes the economics of the choice. A rule removes an option. A nudge does neither, which is why it is cheap to try and easy to reverse.
Thaler coined the word. The Nobel committee noted in 2017 that he showed how nudging can help people exercise better self-control when saving for a pension (Nobel Prize press release, 2017). Marketing borrowed the idea for sign-up flows, checkout pages, emails and apps.
What are the main types of nudge?
The UK government’s BIT unit sums up its approach in four words. Make a behavior Easy, Attractive, Social and Timely (BIT’s EAST framework). BIT first published EAST in 2014 and revised it in 2024. The table maps each principle to a marketing use and a loyalty use.
| Principle | Marketing use | Loyalty use |
|---|---|---|
| Easy | Fewer form fields; a default option already selected | Automatic enrollment of new account holders |
| Attractive | A clear, visible reward at the point of choice | A bonus shown next to the action that earns it |
| Social | "Most customers choose" messages | Tier and leaderboard comparisons |
| Timely | A prompt when the customer is ready to act | A push notification at the usual purchase time |
Defaults are the strongest of the four. Brigitte Madrian and Dennis Shea studied a large U.S. company that changed its 401(k) plan to automatic enrollment. About 86% of employees hired under automatic enrollment took part, against about half of those hired just before the change (Madrian and Shea, 2000). No economic feature of the plan changed.
Social nudges have field evidence too. Goldstein and colleagues ran two hotel field experiments. Signs saying most guests reuse their towels beat the standard appeal about the environment. Signs that named guests in the same room worked best (Goldstein et al., 2008).
Do nudges work?
They work, but less than early studies suggested. A large meta-analysis found a small to medium effect (the meta-analysis). Its original abstract gave a d of 0.45. A reply found that no evidence for nudging remains once publication bias is adjusted for (the reply).
The kind of nudge matters. Mertens and colleagues found that changing the structure of a choice, such as a default, beat changing the information shown. Maier and colleagues found evidence against an effect for information nudges and for nudges that reinforce intentions. For structure nudges, the evidence was undecided. Results vary widely, so some nudges work even if the average does not.
The field data are clearer. DellaVigna and Linos studied 126 trials covering over 23 million people in two U.S. nudge units. In published papers, a nudge raised take-up by 8.7 percentage points. In the nudge unit trials, it raised take-up by 1.4 points, still sizable and highly significant (DellaVigna and Linos, 2022). In their analysis, publication bias and low power explain the gap.
Plan on the 1.4 points. The 8.7 comes from published papers.
For a loyalty team, the lesson is practical. Expect lifts of a point or two. A jump of eight is rare outside published papers. Favor structure nudges such as defaults, since that is where the evidence is least weak. And measure the change in repeat purchase against a holdout.
When does a nudge become a dark pattern?
A nudge turns into a dark pattern when it removes the option to say no. The FTC’s 2022 report describes dark patterns as designs that trick or manipulate consumers. It lists pre-checked boxes, hard-to-find disclosures and confusing rules for canceling (FTC press release). A default is fair when opting out is as easy as opting in. The FTC release also names making it hard to cancel subscriptions or charges. That matters to paid loyalty tiers, where the path to cancel is part of the design.
The test is the original definition. A nudge works without forbidding any option. A loyalty program that makes it hard to leave or to decline an offer has moved from nudge to trap.
How do loyalty programs use nudges?
Starbucks shows several. A new account is enrolled in Starbucks Rewards at the Green Level, which is a default. Offers go out by app push notification, which is a timing nudge. A member can be asked to tap an email or app notice to activate a Double Star Day. That adds one small step the member chooses to take (Starbucks Rewards Terms).
Notice what each does. The default sets the starting point. The push notification asks at a chosen moment. The activation step turns passive interest into a small commitment. Each leaves the member free to decline.
Four nudges are worth a first test in a program. A default can enroll new account holders, with a clear way to opt out. A progress cue can sit beside the action that earns it. A timely message can arrive at the member’s usual purchase time. A social cue can show how many members reached the next tier this month.
How should a loyalty team test a nudge?
Size the test for the effect that is likely, which is small. For illustration, a program wants to know if a reminder lifts 30-day repeat purchase from 20% to 21.4%, a 1.4-point lift like the nudge unit average. A standard two-group test at 80% power and a 5% significance level needs about 13,100 members per group. A 2-point lift needs about 6,500 per group. An 8-point lift needs only about 450.
Run a clean holdout, keep it fixed for the whole window and read the result on repeat purchase. Clicks are a weak guide. Then run it again on a second segment. A nudge that wins in one segment can lose in another.
Log the exposure, the segment and the outcome for every member. Keep the message, timing and offer identical across groups, so only the nudge changes. Fix the metric before launch. Report the result even when the lift is zero.
How GRAVTY handles nudge marketing
Nudges are rules and messages, so a team needs to build and test them fast. The GRAVTY technology page says teams set loyalty rules without code. On the retail page, we describe AgenticTest as watching lift in flight, with live A/B simulation before launch. Our personalization guide covers how to tailor each nudge.
Frequently asked questions
What is an example of a nudge?
A default option is the classic example. At one large U.S. company, about 86% of employees hired under automatic 401(k) enrollment took part, against about half of those hired just before (Madrian and Shea, 2000). The plan’s economics did not change.
Who invented nudge theory?
Richard Thaler and Cass Sunstein developed nudge theory. Thaler coined the term (Nobel Prize press release). A nudge shifts behavior in a predictable way. It bans no option and does not change incentives much (Hansen and Jespersen, 2013).
Do nudges actually work?
Yes, but the effect is smaller than early studies claimed. In two U.S. nudge units, the average nudge raised take-up by 1.4 points, against 8.7 in published papers (DellaVigna and Linos, 2022). One reanalysis finds no evidence for nudging after correcting for publication bias (Maier et al., 2022).
What is the difference between a nudge and a dark pattern?
A nudge leaves the choice open, and a dark pattern tricks or traps the customer. The FTC lists pre-checked boxes, hidden disclosures and confusing rules for canceling as dark patterns (FTC press release). A fair default is one that is easy to decline.
How is nudge marketing used in loyalty programs?
Programs use defaults, timely prompts and small commitment steps. Starbucks automatically enrolls new accounts at the Green Level and sends offers by push notification (Starbucks Rewards Terms). Teams then test each nudge with a holdout sized for a small effect.
