什么是助推营销？定义与指南

助推营销是通过微妙的影响来引导特定消费者行为的做法。

Marketing Fundamentals

Nudge marketing is the practice of shaping how choices are presented to customers, through defaults, order, framing, and effort, without removing options or materially changing prices or incentives. A real nudge stays easy and cheap to decline.

That sounds simple. The evidence isn't. A widely cited 2022 meta-analysis pooled more than 200 studies and reported a small-to-medium average effect (Mertens et al., 2022, PNAS), but independent reanalyses of the same literature found that the effect shrinks sharply, or leaves no clear evidence at all, once publication bias is corrected (Maier et al., 2022; Szaszi et al., 2022). So this entry treats nudges as small adjustments at specific points in the customer journey, and it proposes a stricter test than conversion lift: measure a nudge by what it costs a customer to say no.

Nudge Marketing, Defined by Its Limits

The term comes from behavioral economics, a field that studies how real people decide rather than how idealized models say they should. Richard Thaler and Cass Sunstein popularized it in their 2008 book Nudge, which uses the term for a change to how choices are presented that steers behavior while keeping the options open and leaving economic incentives largely unchanged. Marketers borrowed the idea. They didn't always borrow the limits.

The Three-Part Test: Options Kept, Incentives Unchanged, Easy to Avoid

Before calling a tactic a nudge, check three things.

Options kept. Every alternative the customer had before is still available and visible.
Incentives unchanged. The price, the reward, and the penalty stay the same. Only the presentation moves.
Easy to avoid. Choosing differently takes little time, money, or effort, which is the condition Thaler and Sunstein attach to the term.

Fail the first test and you've built a restriction. Fail the second and you've changed the offer. Fail the third, and the design starts drifting toward the kind of practice that US and UK regulators now penalize.

Nudge vs. Discount, Restriction, and Dark Pattern

A discount changes the economics of the decision, so it isn't a nudge, even when it lifts conversions. Removing a cheaper plan from a pricing page is a restriction. Dark patterns are different again. The US Federal Trade Commission's 2022 staff report describes design practices that trick or manipulate people into choices they wouldn't otherwise make, including countdown timers on offers that aren't actually time-limited (FTC report on dark patterns, Bringing Dark Patterns to Light, 2022). The technique rarely decides the label. Honesty and reversibility do.

Where Nudges Sit in the Customer Journey

Researchers group choice architecture interventions into decision information, decision structure, and decision assistance (Mertens et al., 2022). Weinmann, Schneider, and vom Brocke define digital nudging in their paper as steering choices through the way a digital interface presents them (Weinmann et al., 2016, Business & Information Systems Engineering).

Decision Structure: Defaults, Order, and Effort

Structural nudges change the shape of the choice. A preselected plan, the order of shipping options, and the steps needed to change a setting all count. In the Mertens meta-analysis, interventions that targeted the organization and structure of options consistently outperformed those that only changed how options were described. Defaults are the best-known structural tool, covered separately below.

Decision Information: Framing, Salience, Social Cues

Information nudges leave the structure alone and change what the customer notices. Examples include price framing, highlighted attributes, and showing what other buyers chose. That last category overlaps with social proof signals, and it carries a specific risk. A social cue only helps when it's true. Invented demand figures belong on the dark pattern side, so keep information nudges accurate and easy to verify.

What the Evidence Says, and Why the Headline Number Shrank

Nudges earned their reputation from striking individual studies. The pooled evidence is more modest. Each synthesis below is reported by author, because the methods differ and the numbers shouldn't be averaged.

The 0.45 Figure and the Publication-Bias Debate

Mertens and colleagues analyzed more than 200 studies reporting over 450 effect sizes, with 2,149,683 participants in total. They reported an overall effect of Cohen's d = 0.45, with a 95% confidence interval of 0.39 to 0.52, and they also found moderate publication bias in the literature (Mertens and colleagues' 2022 meta-analysis, PNAS).

Three readings of that bias now sit side by side.

Mertens et al.: their own sensitivity analysis put the effect much lower under a severe publication-bias scenario.
Maier et al.: after applying a Bayesian correction for publication bias, they concluded that no evidence for the effectiveness of nudges remains (Maier and colleagues in PNAS, 2022).
Szaszi et al.: their letter notes that Mertens and colleagues' own bias adjustment put the average between d = 0.03 (severe bias) and d = 0.31 (moderate bias), reports its own bias-corrected estimates of d = -0.01 to 0.08, and argues there's no reason to expect large and consistent effects (Szaszi and colleagues' PNAS letter, 2022).

The fair summary is narrow. Average effects look small after correction. They also vary widely between studies. That undercuts the claim that nudges reliably work everywhere; whether a specific nudge works has to be tested case by case.

Published Results vs. Results at Scale

The clearest test of scale comes from government nudge units in the United States. DellaVigna and Linos assembled 126 randomized trials covering 23 million people, all run by two of the largest US nudge units, and compared them with nudge trials published in academic journals (DellaVigna and Linos in Econometrica, 2022). In journal papers, the average nudge raised take-up by 8.7 percentage points. In the nudge units, the average was 1.4 percentage points. The authors attribute about 70 percent of that gap to selective publication, made worse by the low statistical power of smaller academic studies.

For marketers, the lesson is practical. A published effect may overestimate the result in a new setting; it is not a forecast. Domain matters too: the Mertens analysis found food choices especially responsive, with effect sizes up to 2.5 times larger than in other domains, but that finding describes food decisions, not digital marketing.

Defaults: Stronger on Average, Still Unpredictable

Defaults are the exception critics tend to name. Jachimowicz and colleagues meta-analyzed 58 default studies with a pooled sample of 73,675 and found a considerable average effect of d = 0.68, with a 95% confidence interval of 0.53 to 0.83, alongside substantial variation between studies, and defaults in consumer domains were more effective than those in environmental ones (meta-analysis of default effects, Jachimowicz et al., 2019, Behavioural Public Policy). Even the Szaszi letter lists defaults among the few exceptions to its skepticism. Strong on average, then. Unpredictable case by case. A default that works on one checkout can fail on the next, so it still needs its own test.

When a Nudge Becomes a Dark Pattern

The legal line is getting sharper in the United States and the United Kingdom. Regulators in both have moved from describing manipulative design to fining it. Two problems keep coming up: fake urgency, and choices that are easy to accept but hard to refuse.

False Urgency and Fake Scarcity

The FTC's 2022 staff report lists countdown timers on offers that aren't really time-limited, along with claims that an item is nearly sold out when stock is plentiful, as dark patterns that create false beliefs. A Congressional Research Service dark patterns brief gives false limited-time messages and countdown timers as typical examples. Scarcity marketing only stays on the nudge side when the scarcity is real.

Pre-Ticked Add-Ons and Hard-to-Decline Offers

In the UK, the Competition and Markets Authority fined appliance retailer Marks Electrical £720,000 and ordered refunds of around £600,000 to nearly 40,000 customers after the company automatically opted shoppers into additional paid services (CMA press release on Marks Electrical, June 18, 2026). The charges were pre-selected. As the case was reported, UK consumer law doesn't let a business infer consent to a paid extra from a box the customer never ticked.

In the US, Amazon settled FTC allegations over Prime enrollment and cancellation on Sept. 25, 2025. The settlement totals $2.5 billion, made up of a $1 billion civil penalty and $1.5 billion in refunds for an estimated 35 million consumers, under the FTC Act and the Restore Online Shoppers' Confidence Act (FTC press release on Amazon). Amazon must now offer a clear and conspicuous button to decline Prime, and it can no longer use a button that says "No, I don't want Free Shipping."

Both cases turned on how consent was obtained or how hard it was to cancel, not on persuasion itself. That's where the idea of sludge helps. Sludge means excessive or unjustified frictions, such as paperwork burdens, that cost people time or money, and Sunstein argues that firms and private and public institutions should regularly conduct sludge audits to catalogue those costs and decide how to reduce them (Sunstein's Sludge Audits article, Behavioural Public Policy). For anyone building a landing page with an upsell, the practical rule is plain. If the extra costs money, the customer should opt in, not opt out.

Testing a Nudge Without Overclaiming

Given the evidence above, any nudge deserves a controlled test before wide rollout. Published averages can't predict your own pages.

Measure the Decision, Not Just the Click

A nudge that lifts first clicks but also lifts refunds, cancellations, or complaints hasn't helped. Track the whole decision. That means the initial choice, whether customers keep it, and whether they understood what they agreed to. Run a randomized comparison where you can, set the success metric before launch, and don't borrow an expected lift from academic studies, given the journal versus practice gap DellaVigna and Linos documented. This is editorial guidance. It isn't a benchmark.

Audit the Path to "No"

This entry recommends one test above the others. Measure a nudge by the cost of saying no. Walk the refusal path the way a customer would, and time it.

Can the customer see the alternative without scrolling, hovering, or opening a menu?
Does declining take one action, or several?
Is the decline option worded neutrally, without guilt or shame?
Can the customer reverse the choice later, using the same method they used to accept it?

If the honest answer to any of these is no, the design is sliding from nudge toward sludge, and that's exactly where the US and UK cases above landed. Clear choice architecture usually depends on clear copy, so this audit fits naturally alongside content marketing that explains choices, where plain language does most of the work. You can browse past projects in the published project portfolio.

Nudge Marketing FAQs

What is nudge marketing in simple terms?

It's changing how a choice is presented so one option becomes easier or more noticeable, while every option and price stays the same. Defaults, ordering, and framing are the usual tools.

Is nudge marketing manipulative?

It doesn't have to be, but the label doesn't make it ethical. A nudge that's truthful and easy to decline sits on the acceptable side. Hidden costs and hard cancellation drew penalties in the US and UK cases above (FTC, 2025; CMA, 2026). The FTC's 2022 staff report also lists fake urgency as a dark pattern.

What is the difference between a nudge and a dark pattern?

A nudge keeps options open and easy to choose. A dark pattern tricks or obstructs, for example with countdown timers on offers that aren't really time-limited (FTC, 2022). Honesty and the effort needed to refuse usually decide which side a design is on.

Do nudges actually work?

Sometimes, and usually less than headlines suggest. Pooled estimates shrink after publication bias is corrected, and results vary widely (Maier et al., 2022; Szaszi et al., 2022). In large US government trials, the average effect was 1.4 percentage points, compared with 8.7 in academic journals (DellaVigna & Linos, 2022).

What is sludge?

Sludge is excessive or unjustified friction, such as paperwork or a cancellation process far longer than the signup, that costs people time or money. Sunstein recommends regular sludge audits to find and reduce it.

Are pre-ticked boxes legal?

For paid extras in the UK, the Marks Electrical case shows they can lead to fines and refunds (CMA, 2026). This entry covers only US and UK enforcement, so check local rules elsewhere and treat this as general information, not legal advice.

Related Terms

Behavioral targeting and audience selection - which decides who sees an offer, not how the choice looks.
A/B testing for design changes - the standard way to check whether a nudge helps on your own pages.
Behavioral Economics - the field behind the nudge concept.
Conversion Rate Optimization - the wider practice of raising goal completion.
Neuromarketing - which studies consumer responses rather than designing the choice itself.