Defaults, the right to say no, and urgency that’s true, for brands that want customers to come back on purpose.
Every DTC store makes dozens of choices for its customers before they make one. Subscribe or one-time. Every 30 days or 60. Shipping protection on or off. Somebody picked each, usually an app’s setup wizard, and the pick decides more than any copy on the page.
The first number is why defaults matter: changing only which answer was preselected roughly doubled agreement. The second is why the obvious move backfires. The auto-renewing trial made more money in the short run. It also put people off starting at all, and 20 months later the newspaper had fewer subscribers than with trials that ended quietly.
That’s the argument of this guide. Defaults are the strongest tool you have, stronger than any email. Used for the customer, they remove work and build a habit. Used against the customer, they pull money forward from next year and send the bill later as refunds, disputes, unsubscribes and regulators. Urgency works the same way: a real deadline helps people act on what they already want, and a fake one teaches them to ignore you.
The customer who chose you on purpose is worth more than the one who forgot to untick a box.
Cancel flows live in The Standing Order and reading tests in The Honest Test. This guide covers the choices in between: what’s preselected, how you ask, and whether your clock tells the truth.
Start with The Default Audit, or with the one-page map just below. Or follow a path:
Three tools and a scored audit run in the page. Nothing you type leaves your browser.
Examples that open with Say or Picture use made-up round numbers. Every source is listed in Appendix C. The chapter on the rules is an operator’s summary as of September 2026, not legal advice.
What this guide argues, and what would prove each claim wrong.
A position says what would prove it wrong. Test each on your own store.
Seven places a DTC stack decides for the customer. Each has a common default, an honest one, and a number that tells you which you have.
Walk your store on a phone as a new customer and write down every choice that arrives already made. Then find its row.
The whole book is above and always will be. These are the same chapters addressed individually, for linking to one idea rather than to ninety.
| Choice | Common default | Honest default | The number that tells you |
|---|---|---|---|
| Subscribe or one-time | Subscription ticked, terms below the fold | Either, with terms beside the button | First-renewal refunds and disputes |
| Reorder cadence | The shortest interval | The interval customers actually use | “Too much product” cancels per 100 subscribers |
| Paid add-ons | Ticked | Unticked, one plain line on what it covers | Refunds and tickets naming the add-on |
| Email sign-up | Ticked at checkout | Unticked, with a reason to tick it | Complaints and unsubscribes in the first 30 days, by source |
| SMS sign-up | Folded into another box | Its own unticked box and disclosure | Opt-outs per send, by source |
| Skip, pause or cancel | Cancel hidden, skip missing | Skip and pause one tap away; cancel as easy as signup | Disputes per 1,000 renewals |
| Urgency claims | Timers that reset, stock that never runs out | Real dates and real stock | Days until the “ended” offer returned |
The rows interact. A pre-checked subscription raises the subscription share this month and the dispute rate in two months, and the disputes land with your payment processor, not in your retention report. A timer that resets lifts this week’s email and weakens the next. The cost of a bad default shows up somewhere other than where it lives. That’s why it survives.
A bad default rarely shows up where it lives. It shows up in support, in disputes and in next quarter’s list.
Twelve checks on what your store decides for customers, how it asks, and whether its clocks tell the truth. About fifty minutes with a phone, your subscription app, your email platform and your support inbox.
You can’t avoid defaults: a box is either ticked or it isn’t. The audit asks whether each one is what a customer would pick if asked plainly, and whether its costs show up anywhere you look.
Have the screenshots from the one-page map to hand. Score each check 0 to 2: 0 if it failed or nobody can answer it, 1 if partly true, 2 if clean. “The app set it up that way” scores 0.
Every default was chosen by someone. The audit finds out who, and whether they’d choose it again.
Score as you go; your band appears when all twelve are in.
| Score | What it means | Read next |
|---|---|---|
| 20–24 | Your defaults work for customers. Keep them that way as apps update, and start testing real deadlines against your old habits. | Testing a Real Deadline, then The Choice Scorecard |
| 14–19 | Mostly sound, with a few boxes doing quiet damage. Fix any zero on checks 5, 6, 8 or 11 first: those carry legal risk as well as cost. | The chapter linked from your lowest check, then Good Defaults and Bad Ones |
| 8–13 | Your store is borrowing from next quarter. Take-up looks good because customers haven’t noticed yet. | Where a Default Becomes a Trap, then The Checkout Box |
| 0–7 | Stop adding boxes and timers. Untick everything paid, make every claim true, and put cancel where signup is, this month. | The Rules, September 2026, then The First Thirty Days |
No subscriptions? Score checks 3, 4, 7 and 8 as 2. A clean 16 on the other eight is a strong start.
The same people, the same choice, a different box ticked: agreement roughly doubles. Why that happens decides how you should use it.
Two studies started the modern interest in defaults. What they found is subtler than the headline.
In 2003 Eric Johnson and Daniel Goldstein asked 161 people online whether they would be organ donors. The only thing that changed between groups was the default. When people had to opt in, 42% agreed. When they had to opt out, 82% stayed in. With no default at all, forced to choose, 79% agreed Published. Across European countries the gap was the same shape and bigger: effective consent of 12% in Germany, where people opt in, against 99.98% in Austria, where they opt out Published.
Two years earlier, Brigitte Madrian and Dennis Shea had looked at a large US employer that switched its 401(k) from opt-in to automatic enrollment. Among employees with similar tenure, participation was 37% before the switch and 86% after Published. Only the default changed.
PublishedJohnson and Goldstein, Science, 2003; Madrian and Shea, Quarterly Journal of Economics, 2001.
Both studies carry a warning the famous chart hides. Six in ten automatically enrolled employees did nothing to change the company’s choice of a 3% contribution in a money market fund, and 80% of their contributions went to that fund Published. The default raised participation and anchored people at a savings rate few had chosen. The authors put it down to inertia and to employees reading the default as advice.
And consent is not donation. Johnson and Goldstein estimated that an opt-out default was associated with a 16.3% rise in actual donation rates, from 14.1 to 16.4 donors per million Published. A real gain, far smaller than a 60-point gap in consent. Between the box and the outcome sat families, doctors and practical steps.
A default moves the box. What happens after the box is still up to the customer.
That’s the first lesson for a store. A preselected box moves take-up. Whether it moves reorders, margin or a list that buys depends on what happens next, which is where bad defaults fail.
A meta-analysis of 58 default studies by Jon Jachimowicz and colleagues tested three explanations Published:
Endorsement matters most for a brand. Preselect subscribe-and-save and the customer reads it as advice. If the advice is good, the default is a service. If not, you’ve spent credibility, and the customer finds out at the first renewal.
In a German experiment with online energy orders, making the pricier green tariff the default increased its purchase nearly tenfold Published. In 13 million New York taxi rides, suggested tips moved what riders tipped, but higher suggestions pushed more riders to leave no card tip at all Published. Push a default too far and some people opt out of the whole thing.
Defaults hold up when the evidence is checked for publication bias. Most other nudges shrink, and the ones that look like your emails shrink most.
Behavioral science had a hard decade. Famous findings failed to replicate, and journals turned out to have published the lucky results. Before you plan around a nudge, know which effects survived the checking.
Jachimowicz and colleagues’ 2019 meta-analysis of defaults found a large average effect, d = 0.68, with wide variation: most studies found positive effects, several found none, two found negative ones. They tested for publication bias, estimated about eight studies were likely missing, and found the effect held Published. Defaults in consumer settings were stronger than average.
In 2022 Stephanie Mertens and colleagues pooled over 200 studies of all kinds of nudges, with more than two million participants, and found d = 0.43 overall, with “decision structure” nudges, which include defaults, strongest at d = 0.54 Published. Months later Maximilian Maier and colleagues re-analyzed the same data with a method that corrects for publication bias. The corrected average for all nudges was d = 0.04: no evidence of an effect. For the structure category, which holds the defaults, the evidence was “undecided,” and the authors noted that the spread of results means “some nudges might be effective, even when there is evidence against the mean effect” Published.
PublishedJachimowicz, Duncan, Weber and Johnson, Behavioural Public Policy, 2019; Mertens, Herberz, Hahnel and Brosch, PNAS, 2022; Maier, Bartoš, Stanley, Shanks, Harris and Wagenmakers, PNAS, 2022. The bias-corrected estimate for structure nudges alone was reported as undecided, not as a number.
The most useful study for a DTC operator is the one that compared journals with the real world. Stefano DellaVigna and Elizabeth Linos collected every trial run by two large US government nudge units, 126 trials covering 23 million people, and compared them with nudges published in academic journals. In the journals, the average nudge raised take-up by 8.7 percentage points, a 33.4% lift. In the nudge units’ full set of trials, it was 1.4 points, an 8.0% lift Published. About 70% of the gap came from selective publication combined with small samples.
Look at what the nudge units sent. About 90% of their nudges were emails, letters and postcards. Only one trial used defaults, and it was left out of the main analysis Published. So the realistic benchmark for a reminder email is 8%, not 33%, and it says nothing about defaults, a different and stronger tool.
Plan a reminder email at single-digit lifts. Save the big expectations for the choices you preselect.
Say a brand sends a replenishment reminder to 20,000 customers a month and 5% would reorder that week anyway. An 8% lift, strong by the nudge-unit standard, adds 80 orders. Someone expecting 33% was counting on 330, and will spend months rewriting copy to chase a number that was never there.
Defaults are your biggest lever, for better or worse, so they need the most care. Messages are small levers: test them with a holdout (The Honest Test covers sizing) and distrust any double-digit claim for a phrase. Chapter 11 has an example.
Three questions sort them. And one large field experiment shows what a bad default does to growth when you wait long enough to see.
Every default has two effects: the take-up it produces now, and the behavior of the people it enrolled. A good default produces customers who act like they chose. A bad one produces customers who act like they were caught.
Most defaults pass the first question for some customers and fail it for others. The fix is to present both options clearly, or to preselect and make the terms and the way back impossible to miss.
The best evidence on this comes from a field experiment run with a large European newspaper and reported in a January 2026 working paper by Klaus Miller, Navdeep Sahni and Avner Strulov-Shlain. Over a million readers were offered trial subscriptions of two or four weeks, free or at €0.99. Some trials renewed into a paid subscription automatically. Others ended unless the reader chose to continue Published.
The auto-renewing trial did what everyone expects in the short run: more paying subscribers right after the trial. But fewer readers started a trial when it auto-renewed. The authors’ summary: auto-renewal “benefits the firm in the short-term,” but “lowers subscriptions take-up by 35% and total subscribers by 23% over 20 months” Published. Many readers were inert, the paper finds, “yet most anticipate and account for their inertia.”
Customers price in your defaults. The ones who expect to be caught don’t start.
It’s a working paper, not yet peer reviewed, and one newspaper isn’t every brand. But your customers have seen plenty of auto-renewing trials. When they suspect a default is set against them, some decline the whole offer.
Liran Einav, Benjamin Klopack and Neale Mahoney used payment-card data to show that cancellations jump in months when a card is replaced and the subscription must be actively renewed. Their models estimate these frictions “roughly double seller revenues on average, holding fixed initial subscribers” Published.
Put the two side by side. Inertia is worth a lot to the seller, which is why bad defaults exist. Customers know it, which is why bad defaults cost more over time than they show at launch. Revenue that depends on customers not noticing is borrowed from next year’s growth, and in 2025 a regulator called in one very large loan (chapter 9).
Say a coffee brand preselects “every 3 weeks” because its one-time buyers reorder at a median of 22 days, and shows the price, next shipment date and “skip or cancel any time” beside the button. Most customers would pick it, nobody is surprised, and moving the date is one tap. That default earns its take-up.
Now say it preselects “every 2 weeks” to lift revenue per customer, with the terms in a tooltip. It fails all three questions. The second bag arrives while the first is half full, and “I have too much” becomes the top cancel reason.
The two defaults on the product page that decide most of a subscription program’s first year: which option is ticked, and how often the box comes.
A subscription toggle makes two choices before the customer does: one-time or subscribe, and how often. It’s easy to set both when you install the app and never look again.
Preselecting subscribe-and-save will raise your subscription share, partly because customers read it as your recommendation (chapter 2). So preselect it only where the recommendation is true.
Whatever you preselect, the law cares about what the customer sees. The Restore Online Shoppers’ Confidence Act requires online sellers of anything that renews to disclose all material terms before taking billing information, get the customer’s express informed consent before charging, and give a simple way to stop the charges Published. California’s automatic renewal law, as amended for contracts from July 1, 2025, requires the renewal terms to be clear and conspicuous before the purchase and the customer’s “express affirmative consent” to them, and requires the business to keep proof of that consent for at least three years Published. A ticked toggle with the terms in a tooltip is the weakest position under both. Have counsel review the product page and checkout together.
My rule: preselect subscribe only when all four are true.
The interval default gets less attention and does more damage. Ship faster than people use the product and every subscriber builds a pile, which becomes the reason to cancel. The right default is the interval customers actually use, readable from one-time buyers: the median days between first and second orders of the same product. The Second Order covers replenishment timing in depth.
Say a supplement lasts most people 45 days, and the app’s default is 30. Say subscribers on the 30-day default cancel at 12% of shipments, because the bottles pile up, and subscribers on a 45-day default cancel at 7% of shipments. Over a year at $34 a bottle, the 30-day group receives about 6.5 shipments per starting subscriber, $222 of revenue, and 22% are still subscribed at the end. The 45-day group receives about 6.3 shipments, $214, and 56% are still subscribed. Year-one revenue is nearly the same. The 45-day default goes into year two with two and a half times as many subscribers.
A cadence that’s too fast doesn’t earn more. It spends next year’s subscribers to hit this year’s number.
Show the reason beside the default: “Most people finish a bottle in about six weeks.” That turns the endorsement into advice the customer can check, and a reason to change it if they use more or less.
Shipping protection, gift wrap, donations and warranties. A ticked paid add-on looks like free margin. Count what comes back.
The most common preselected choice in DTC checkout is a small paid extra: package protection, a donation, priority handling. It’s often installed by an app that may be paid on take rate, and ticked because a ticked box sells more. The question is what the extra takers cost you afterward.
In the EU this was settled in 2011. The Consumer Rights Directive requires “express consent to any extra payment,” and a charge added through “default options which the consumer is required to reject” must be refunded Published. The UK has the same rule in its Consumer Contracts Regulations.
The UK regulator has now enforced it with its new direct powers. In June 2026 the Competition and Markets Authority ordered Marks Electrical, a UK appliance retailer, to refund nearly 40,000 customers about £600,000 for two pre-selected paid services at checkout, recycling an old appliance and unwrapping and recycling packaging, and fined it £720,000 Filed. The CMA’s Emma Cochrane put it plainly: “The law is absolutely clear that automatically opting customers into extra charges is never ok.”
The US has no single rule banning a ticked add-on. But the FTC’s 2022 report on dark patterns lists pre-checked boxes among the old tricks dark patterns have refined Published, and its 2025 Amazon settlement required a clear decline button that can’t say “No, I don’t want Free Shipping” Filed. Have counsel review what applies in the states you sell into.
Even where it’s allowed, it might not pay. People who take an add-on because it was ticked aren’t the people who choose it. Some email support. Some ask for a refund. A few dispute the whole order, which costs the fee, the order and a mark on your dispute ratio. Some decide you’re a brand that slips things into the cart. None of that appears in the app’s dashboard. The tool puts it next to the margin.
With the defaults, a store with 10,000 orders a month and a $3 add-on gets 4,500 extra takers from the tick and $8,100 of extra add-on margin. Six in a hundred of them complaining, plus three disputes in a thousand, costs back about $8,050. The tick earns about $54 a month, and it stops paying at a complaint rate just over 6% Derived. Unticked, the same add-on earns $4,500 a month from customers who wanted it, with none of the complaints.
The margin from a ticked box is visible in one dashboard. Its costs are scattered across four. That’s the only reason it looks good.
The complaint rate decides this, and most stores have never counted it. Search a month of tickets and refund notes for the add-on’s name and divide by default-only takers, not all orders. Better still, untick it for half your traffic for four weeks and compare add-on revenue, tickets and refunds.
An unticked add-on still sells when explained in one plain line near the total: “Package protection, $3: if it’s lost or damaged in transit, we’ll replace it.” Customers who tick it won’t complain.
A ticked sign-up box grows the list and shrinks what the list is worth. For text messages it also creates legal risk that no list size justifies.
The email box at checkout is the default brands defend hardest. The case for ticking it is list size. The case against is what a list is for: people who read what you send.
In the US, CAN-SPAM is an opt-out law. You can email customers who didn’t ask, as long as every message carries a working unsubscribe that you honor within 10 business days Published. So a ticked email box at checkout is generally lawful for US customers. In the EU it isn’t a form of consent at all. The GDPR’s recitals say that “silence, pre-ticked boxes or inactivity should not therefore constitute consent” Published, and in the 2019 Planet49 case the EU’s top court ruled that a pre-ticked checkbox is not valid consent Published.
Lawful isn’t useful. Subscribers who never chose you tend to complain more, and complaints are what mailbox providers watch when deciding whether you reach the inbox (The Whole Machine covers deliverability). Say a ticked box adds 4,000 subscribers a month and an unticked one adds 2,500. If the extra 1,500 mark you as spam at several times the others’ rate, the list grows and your reach shrinks. Tag subscribers by source and compare 30-day clicks, unsubscribes and complaints before you argue about it.
Text messages are different in kind. Under the Telephone Consumer Protection Act rules, marketing texts sent with automated systems need the customer’s “prior express written consent”: a signed agreement, which can be electronic, that clearly authorizes the messages and tells the customer they don’t have to agree as a condition of buying anything Published. A pre-ticked box, or SMS consent folded into the email box, is a weak way to show that. Customers can revoke “by using any reasonable method,” honored within 10 business days Published. Give SMS its own unticked box and disclosure, and have counsel review it.
A list is worth what its readers would pay to stay on it. Nobody pays for a box they didn’t tick.
An unticked box needs a reason about the customer. Weak: “Sign up for news and offers.” Better: “Email me when it’s time to reorder, and when you restock the 8-ounce.” Whoever ticks that has told you what they want.
The easiest “no” keeps the most customers. A customer who can say “not now” in one tap stays reachable; a customer whose only exit is cancel leaves.
It’s natural to design for “yes.” The “no” gets whatever the app provides: cancel three menus deep, a reminder with one button. Each turns “not now” into “not ever.”
A customer who wants a break and can’t find one cancels, disputes the next charge or stops opening your messages. All three cost more than a skip. Stripe’s guidance on preventing disputes recommends a quick, easy way to cancel, clear billing terms up front and reminders before charges Reported.
The law has moved the same way. Minnesota’s automatic renewal law, which took effect in January 2025 according to the law firm Kelley Drye, restricts retention offers once a customer has asked to cancel, but explicitly allows a seller to describe “downgrading, pausing, or suspending” options Published. California allows save offers during an online cancellation only while a “click to cancel” button stays visible next to them Published. The direction is clear: offer alternatives, never block the exit.
There’s little rigorous public evidence on how much a visible pause reduces churn; most numbers come from vendors that sell cancel flows. The case rests on simpler ground: a skip is a customer you keep, a cancel one you must win back, a dispute one you pay to lose.
Say 1,000 subscribers reach a renewal they don’t need yet. With cancel the only visible exit, 300 leave, 20 of them by disputing the charge. With skip and pause one tap away, 250 skip and 60 leave, 5 of them by dispute. If two-thirds of skippers are still subscribed two months later, that’s over 150 more subscribers from one cycle and 15 fewer disputes. Measure your own numbers; the direction isn’t in much doubt.
“Not now” is information. “Cancel” is a customer walking out because you gave them no other door.
Dark patterns are common, sold as plug-ins, and effective in exactly the way that makes them dangerous: the less the customer understands, the better they work.
There’s no bright line between a helpful default and a trap, but there is a reliable test. A good default works because it’s what the customer would choose. A dark pattern works because the customer didn’t notice, didn’t understand or couldn’t find the way out.
In 2019 a Princeton team led by Arunesh Mathur crawled about 53,000 product pages on 11,000 shopping sites and found 1,818 dark patterns on 1,254 sites, about 11%, with 183 sites outright deceptive. They also found 22 third-party companies selling dark patterns to retailers as ready-made plug-ins Published. Many of your store’s tactics weren’t chosen by you. They came with an app.
Regulators found the same. The FTC’s September 2022 staff report, “Bringing Dark Patterns to Light,” described countdown timers with no real deadline, pre-checked boxes and cancellation paths built to wear people down Published. In January 2023 the European Commission and national authorities screened 399 online shops and found 148 using at least one manipulative practice, including fake countdown timers on 42 sites and hidden subscription terms on 23 Published.
Jamie Luguri and Lior Strahilevitz offered a representative sample of 1,963 Americans a paid “data protection” plan. With a plain offer, 11.3% accepted. With mild dark patterns, 25.8%. With aggressive ones, 41.9% Published.
A second study compared tactics. Hidden information, trick questions and obstruction were the most effective. Loaded language and social proof worked moderately. And “must act now” messages “did not make consumers more likely to purchase a costly service” Published. Keep that in mind for chapter 12.
If an option sells better the less clearly you explain it, it’s selling confusion.
| Pattern | What it looks like in DTC | The honest version |
|---|---|---|
| Sneaking into the basket | A paid add-on already in the cart | Unticked, one line of value |
| Preselection against interest | Subscription or fastest cadence ticked, terms hidden | Terms beside the button |
| Hard to cancel | Cancel by phone only, or behind a login maze | Cancel the way you signed up |
| Confirmshaming | “No thanks, I don’t like saving money” | “No thanks” |
| Hidden information | Ongoing price in gray small print | Same size and place as the intro price |
| False urgency | Timers that reset, stock that never falls | Real dates and real stock |
The largest dark-patterns case so far is about a default. In June 2023 the FTC sued Amazon, alleging it enrolled customers in Prime without consent and made canceling hard: buying without Prime was harder to find, and some purchase buttons didn’t clearly say they also joined Prime. The FTC’s announcement noted press reports that Amazon called its cancellation process “Iliad,” read as an allusion to Homer’s poem of “twenty-four books and nearly 16,000 lines” Filed.
On September 25, 2025, Amazon agreed to pay $2.5 billion: a $1 billion civil penalty and $1.5 billion in refunds to about 35 million customers. It didn’t admit wrongdoing. It agreed to a clear button to decline Prime, which can’t say “No, I don’t want Free Shipping,” clear disclosure of cost, billing dates and renewal, cancellation through the same method used to sign up, and a third-party supervisor for the refunds Filed. By September 2026 it had paid out more than $845 million, and the court had approved raising the cap to $200 per customer Filed.
The FTC quoted an Amazon document calling subscription driving “a bit of a shady world” Filed. Every tactic in the complaint is available to a Shopify store with the right apps.
The federal click-to-cancel rule is gone for now. The obligations it described mostly aren’t, because older federal law, the states, the card networks, the UK and the EU cover the same ground.
An operator’s summary of the rules on defaults, consent and urgency, as of September 2026. Not legal advice: the rules change often, and what applies depends on where your customers live. Have counsel review your flows.
Click-to-cancel. The FTC finalized its revised Negative Option Rule, known as click-to-cancel, in October 2024. On July 8, 2025, six days before most of it was due to take effect, the Eighth Circuit vacated it in Custom Communications v. FTC, because the agency hadn’t done a required preliminary regulatory analysis Filed. The ruling was about procedure, not easy cancellation. The FTC restarted with an advance notice of proposed rulemaking in March 2026; as of September 2026 its rule page shows no new proposed rule Published. No federal click-to-cancel rule is in force.
ROSCA still applies. The 2010 Restore Online Shoppers’ Confidence Act covers anything sold online with a negative option. It requires clear disclosure of all material terms before taking billing information, express informed consent before charging, and “simple mechanisms” to stop recurring charges Published. The Amazon settlement’s terms, a clear decline button and cancellation by the same method as signup, show how the FTC reads “simple.”
Section 5 and the pricing guides. The FTC Act’s ban on deception covers fake urgency. The Guides Against Deceptive Pricing require a “was” price to have been “openly and actively offered” for “a reasonably substantial period” Published. The 2024 reviews rule bars incentives conditioned on sentiment Published.
Text messages. Marketing texts sent with automated systems need prior express written consent, which can’t be a condition of purchase (chapter 7).
The practical standard is the strictest state you sell into, and for most national brands that’s California.
Visa and Mastercard act like regulators on disputes. Visa’s monitoring flags a merchant whose disputes and fraud reports reach 0.5% of transactions, with fees above 1.5% in the US. Mastercard’s program starts at 100 chargebacks and 1.5% in a month. Refunds after a dispute don’t remove it from the count Reported. Surprise charges are a fast way toward those lines.
The federal rule fell. The duty it described didn’t: disclose, get consent, and let people leave the way they came in.
Commands provoke resistance; choices lower it. The famous phrase that illustrates it works best face to face, weakly in writing, and less reliably than its reputation.
Lifecycle copy is full of orders. Reorder now. Don’t miss out. Leave a review. Each pushes against something people guard closely: the feeling that they decide what they buy.
The psychologist Jack Brehm named it in 1966: when people feel a freedom is threatened, they move to restore it, often by doing the opposite. A 2026 meta-analysis by Zixi Li and Jingyuan Shi pooled 33 studies. Language high in threat to freedom (“you must,” “don’t”) raised reactance and anger, and anger was linked to less persuasion. Gain versus loss framing, “save 20%” against “don’t lose 20%,” made no difference Published. The words that matter are the ones telling people what to do.
In 2000 two French researchers, Nicolas Guéguen and Alexandre Pascual, had people ask strangers in the street for bus fare, sometimes adding “but you are free to accept or refuse.” With the phrase, people were more likely to give, and gave more, as the 2023 re-examination below describes it. It became a staple of the persuasion literature.
| Evidence | What it found |
|---|---|
| Carpenter, 2013: 42 studies | A small positive effect on saying yes (correlation .13) |
| Same, decision made on the spot | Correlation .18 |
| Same, decision made later | Correlation .07, a much weaker effect |
| Fillon and colleagues, 2023: pre-registered review of 52 experiments | A medium average effect (g = 0.44), stronger face to face |
| Same, only the 7 studies at low risk of bias | No detectable effect (g = 0.11, interval −0.18 to 0.40); estimated replicability very low |
PublishedChristopher Carpenter, Communication Studies, 2013; Adrien Alejandro Fillon, Lionel Souchet, Alexandre Pascual and Fabien Girandola, Meta-Psychology, 2023.
The first meta-analysis says the phrase works, less so when the decision isn’t made on the spot, which describes most emails and texts. The second, pre-registered, finds the best-designed studies can’t detect an effect. Plausible, probably small in writing, not proven.
Use choice language because it’s respectful and costs nothing, not because a phrase doubles your revenue.
The downside of commanding copy is better supported than the upside of any magic phrase. Removing the orders is likely to help a little and very unlikely to hurt.
| Flow | Commanding | Choice |
|---|---|---|
| Replenishment | “Time to reorder! Don’t run out.” | “Running low? Reorder in one tap, or we’ll check back in two weeks.” |
| Subscription reminder | “Your order ships Friday.” | “Your order ships Friday. Skip it, move it or change it here, whatever suits.” |
| Winback | “We miss you! Come back now for 20% off.” | “If it still fits your routine, here’s where to pick it up. If not, no hard feelings.” |
| Review ask | “Leave us a 5-star review!” | “Would you tell other people how it went? Good or bad, it helps them decide.” |
| Sign-up pop-up | “Don’t miss out! Enter your email.” | “Want 10% off your first order? Or keep browsing.” |
Three rules make the right column work: name the easy no as a real option with a link; say it once, in one line; and mean it, because a choice line above a countdown timer reads as a trick.
Review asks carry a legal edge: the FTC’s 2024 rule prohibits incentives conditioned on a review’s sentiment Published. Ask every buyer the same way. More on reviews is in The Proof File, and on winback offers in The First Offer.
Scarcity makes things more desirable when it’s real. Fake scarcity is common, weaker than it looks, and now the subject of court orders.
Urgency is a claim about the world: this ends Friday, three are left, twelve people are looking. True, it helps customers act on what they already want. False, it’s a lie with a timer attached.
In a classic 1975 study, Stephen Worchel and colleagues gave people cookies from a jar and asked them to rate them. Cookies from a nearly empty jar were rated more desirable than the same cookies from a full one. They were rated higher still when the jar went from full to nearly empty, and highest when people were told the cookies had run low because others wanted them Published. Scarcity works because it usually carries information: something is limited, or other people value it.
That’s why fake scarcity is corrosive: each fake signal teaches the customer that your signals carry no information.
The Princeton crawl from chapter 9 found 393 countdown timers, and 157 of them, about 40%, were deceptive: the timer reset, or the offer it said was expiring stayed live after it hit zero Published. It found 632 low-stock messages; for 17 it could prove the number was fake, including 16 sites that counted stock down in the same repeating pattern Published. The European screening in 2023 found fake timers on 42 of 399 shops Published.
Less than you’d think. In Luguri and Strahilevitz’s second study, “must act now” messages did not make people more likely to buy a costly service Published. Your tests may show a campaign-window lift, but part of it is pulled forward, and the scoreboard never counts the customers who noticed.
The UK’s competition regulator began investigating the mattress brand Emma Sleep in 2022 and went to court in 2024. In May 2026 the High Court confirmed Emma’s undertakings to stop misleading countdown timers, false “high demand” messages and “limited time” sales whose deals carried on after the deadline. Its “was/now” discount claims were set for a separate trial Filed. The regulator’s senior director: “using fake countdown clocks or misleading ‘discounts’ to push people into spending is illegal” Filed.
Emma is an online mattress brand, and every tactic in the case is one a DTC store can install in an afternoon.
The cost lands on the customers who notice: the one who reloads and sees the timer restart, the one who gets “final hours” and then the same offer two weeks later. The tool weighs the lift you can see against the orders you can’t. The noticing inputs are estimates; the point is how small they need to be.
With the defaults, the clock adds 60 truly new orders and $1,800 of contribution. If 15% of the 40,000 customers who saw it notice and buy 5% less next year, the brand loses $6,000: a net loss of $4,200. The clock stops paying if the customers who notice buy just 1.5% less Derived. No published figure says how much a noticed fake costs in future orders. The break-even says how little it takes.
A fake clock has to fool almost everyone, almost forever, to pay. It won’t.
Most DTC businesses have more real deadlines than they use:
The scorer turns twelve months of urgency claims into one number and names the weakest. Pull the counts from your promotion calendar, apps and inventory system.
With the defaults, half the timers resetting, five of eight “last chance” offers returning within a month, two extended deadlines, three in ten low-stock messages false and activity messages that are pure decoration, the store scores 41 out of 100, with false instances in all five kinds of claim. The weakest is the activity messages Derived.
Short, real deadlines help people act on intentions they’d otherwise put off. Whether yours pays depends on 60 days of orders, not the campaign week.
The best argument for a real deadline isn’t pressure. It’s procrastination: people put off even what they want.
Suzanne Shu and Ayelet Gneezy studied gift certificates and gift cards with different expiry dates. People with long deadlines procrastinated more and ended up redeeming at lower rates than people with short deadlines, the opposite of what participants predicted they’d do Published. A longer window felt kinder and produced less use. A short, true deadline can help the customer get what they meant to get. But it only works if they believe that after Friday the chance is gone. Once they’ve seen your “final” offer come back, Friday means nothing.
Say a brand splits 30,000 customers into three groups of 10,000. In the campaign week the deadline group places 420 orders and the no-deadline group 300: a 40% lift. Over 60 days the holdout places 600, the no-deadline group 760 and the deadline group 800. The offer alone added 160 orders; with the deadline, 200. The deadline added 40 orders, a quarter more, not 40%.
And is even that real? The deadline group converted at 8.0% over 60 days and the no-deadline group at 7.6%. At 10,000 customers each, the 95% range for that difference runs from about −0.3 to +1.1 points, and the p-value is about 0.29 Derived. The test can’t tell, a common and useful answer: rerun it bigger before building the calendar on it. The Honest Test shows how to size it.
A deadline that wins the campaign week and ties over 60 days didn’t create demand. It moved it.
Use real deadlines sparingly and protect them. Tie each to something true, and never run the same “final” offer twice in 30 days. How often and how deep to run offers belongs to The First Offer. Your job is that when you say an offer ends, customers believe you.
One page, every month. Each default’s take-up next to what it costs downstream, so nobody can celebrate one without seeing the other.
Defaults survive because their gains are reported and their costs aren’t. The take rate is in the app; tickets, disputes and unsubscribes are in three other tools. The scorecard puts them on one page.
| Number | Defined as | What it catches |
|---|---|---|
| Subscription share, by toggle | Subscriptions over orders on subscribable products, preselected or not | How much of your program the default is carrying |
| First-renewal outcome, by arrival | Canceled, refunded, disputed or renewed, by preselected versus chose | Subscribers who didn’t mean to be subscribers |
| Cadence fit | Skips and “too much product” cancels per 100 active subscribers | A default interval faster than real use |
| Add-on take rate and tickets | Take rate, and tickets or refunds naming it, over default-only takers | Whether the ticked box is paying (chapter 6) |
| List quality, by source | 30-day unsubscribes and complaints by sign-up source, email and SMS | A sign-up box nobody chose |
| Easy-no usage | Skips, pauses and “not yet” taps, against cancels | Whether customers can find the other doors |
| Dispute ratio | Disputes over transactions, monthly, by cause | Surprise, and distance from the card networks’ lines |
| Urgency honesty score | From the scorer in chapter 12, re-run monthly | An app update that switched a fake timer back on |
| Offer recurrence | Days until each “final” offer, or a better one, ran again | Deadlines customers have learned to ignore |
Put the take rate and its cost on the same line, or the take rate will win every argument.
First, every take-up number sits on the same row as its downstream number; a take rate without tickets is half a fact. Second, compare by arrival: the customers a default brought in against those who chose. The blended average hides exactly the people the default affects.
Expect blanks in the “by arrival” rows at first; start tagging now (chapter 5) and they fill within a quarter. When a number moves, check that month’s app updates first, because defaults change without anyone deciding. And read the tickets behind the complaint counts: “I never signed up” and “I didn’t add this” are the clearest signal on the page.
Legal risk first, then the boxes, then the words and clocks, then the measurement that keeps them honest. Four weeks, in that order.
Whether you’ve just run the audit or just inherited a store, the order is the same: remove what could get you in trouble, fix the costly defaults, change the words and clocks, then measure so it stays fixed.
At day thirty you won’t have results; renewals and deadline tests take 60 days. You’ll have a store where every preselected choice is one you’d defend to a customer, every claim is true, and the costs are on one page.
Remove the risk, fix the boxes, then the words. Measure so it stays fixed.
Seven things the person who owns your defaults needs on the first day.
At many brands nobody owns defaults; apps and a theme developer set them. Whoever takes this on needs seven things on day one, or the first month goes on finding out who set what.
The books and papers this guide leans on, and what to take from each.
Full references for everything else are in Appendix C.
Andrew Lauchner runs Growth Legend, embedding inside consumer brands to own lifecycle, email and SMS, and revenue operations. He is the author of The Second Order, on turning first-time buyers into second-time buyers, and The Whole Machine, on the fundamentals of DTC growth, along with a series of field guides for DTC operators at andrewlauchner.com.
As Senior Director of Growth and Retention Marketing at Gallery Furniture, he rebuilt the customer journey and the sales playbooks together. He has worked on growth and retention at Binance and 3Commas, and has been Head of Growth and Retention at Greatness Wins and at Nexus Agriscience.
“Andrew led retention, lifecycle, and email/SMS, but what separates him from most in this space is how deeply he understands the role retention plays in the overall growth engine.”
Akram Khan, Head of Marketing at Gallery Furniture, senior to Andrew but didn’t manage Andrew directly
Andrew answers every note from operators working on this, including those looking for someone to own it. Write to andrew@growthlegend.com or message him on LinkedIn.
The formulas behind the three tools, and four queries that fill most of the scorecard.
| For | Formula | Notes |
|---|---|---|
| Default-only takers | E = N × (tticked − tunticked) | N: orders; take rates from a split test. |
| What the tick earns | E·P·m − E·s·(P·m + c + L) − E·d·F | P price, m share kept, s complaint rate, c handling, L lost future contribution, d dispute rate, F cost per dispute. |
| Break-even complaint rate | (P·m − d·F) / (P·m + c + L) | Above this, the tick loses money. |
| Fake clock, gain | O × lift × (1 − pulled forward) × contribution | O: campaign orders without the clock. |
| Fake clock, loss | seen × noticed × V × drop | V: next-12-month contribution per customer who saw it. |
| Fake clock, break-even drop | gain / (seen × noticed × V) | Above this, the clock loses money. |
| Urgency honesty score | 100 − 25a − 25(c/b) − 10·min(f,4)/4 − 20d − 20e | a, d, e as shares; b: last-chance offers; c: those that returned within 30 days; f: extended deadlines. |
-- orders in a split test of the add-on box: ticked vs unticked
-- order_lines: sku 'PKG-PROTECT' marks the add-on
-- checkout_variants: order_id, variant ('ticked' / 'unticked')
-- tickets: one row per ticket, with order_id and body
SELECT v.variant,
COUNT(DISTINCT o.order_id) AS orders,
COUNT(DISTINCT ol.order_id) AS took_addon,
COUNT(DISTINCT r.order_id) AS addon_refunded,
COUNT(DISTINCT t.order_id) AS addon_tickets,
COUNT(DISTINCT d.order_id) AS disputes
FROM orders o
JOIN checkout_variants v ON v.order_id = o.order_id
LEFT JOIN order_lines ol ON ol.order_id = o.order_id AND ol.sku = 'PKG-PROTECT'
LEFT JOIN refunds r ON r.order_id = o.order_id AND r.line_sku = 'PKG-PROTECT'
LEFT JOIN tickets t ON t.order_id = o.order_id
AND t.body ILIKE '%protection%'
LEFT JOIN disputes d ON d.order_id = o.order_id
WHERE o.created_at >= CURRENT_DATE - INTERVAL '35 days'
GROUP BY v.variant;
Default-only takers are the difference in took_addon between variants. Divide the difference in refunds and tickets by it to get the complaint rate for the tool in chapter 6. Allow a week for late tickets and disputes.
-- subscriptions: arrival = 'preselected' or 'chose' (tag at checkout)
-- subscription_charges: one row per renewal attempt, with charge_number
SELECT s.arrival,
COUNT(*) AS subscribers,
AVG(CASE WHEN s.canceled_at < c.scheduled_at THEN 1 ELSE 0 END) AS canceled_before_first_renewal,
AVG(CASE WHEN c.status = 'paid' THEN 1 ELSE 0 END) AS first_renewal_paid,
AVG(CASE WHEN c.refunded THEN 1 ELSE 0 END) AS first_renewal_refunded,
AVG(CASE WHEN c.disputed THEN 1 ELSE 0 END) AS first_renewal_disputed
FROM subscriptions s
LEFT JOIN subscription_charges c
ON c.subscription_id = s.subscription_id AND c.charge_number = 2
WHERE s.created_at BETWEEN CURRENT_DATE - INTERVAL '150 days'
AND CURRENT_DATE - INTERVAL '60 days'
GROUP BY s.arrival;
Charge number 2 is the first renewal. Refunded and disputed renewals are the clearest sign of a default nobody chose.
-- median days between first and second order of the same product,
-- one-time buyers only
WITH firsts AS (
SELECT o.customer_id, ol.product_id, o.created_at,
ROW_NUMBER() OVER (PARTITION BY o.customer_id, ol.product_id
ORDER BY o.created_at) AS n
FROM orders o
JOIN order_lines ol ON ol.order_id = o.order_id
WHERE o.subscription_id IS NULL
)
SELECT a.product_id,
PERCENTILE_CONT(0.5) WITHIN GROUP (
ORDER BY EXTRACT(DAY FROM b.created_at - a.created_at)) AS median_days,
COUNT(*) AS repeaters
FROM firsts a
JOIN firsts b ON b.customer_id = a.customer_id
AND b.product_id = a.product_id AND b.n = 2
WHERE a.n = 1
GROUP BY a.product_id
HAVING COUNT(*) >= 50;
This counts only people who came back, so treat it as the fastest sensible default and round up.
-- promotions: promo_id, offer_key (what's offered), discount_pct,
-- starts_at, ends_at, called_final (true if billed as final/last chance)
SELECT p.promo_id, p.offer_key, p.ends_at,
MIN(q.starts_at) AS next_same_or_better,
EXTRACT(DAY FROM MIN(q.starts_at) - p.ends_at) AS days_until_it_returned
FROM promotions p
LEFT JOIN promotions q
ON q.offer_key = p.offer_key
AND q.discount_pct >= p.discount_pct
AND q.starts_at > p.ends_at
WHERE p.called_final
AND p.ends_at >= CURRENT_DATE - INTERVAL '12 months'
GROUP BY p.promo_id, p.offer_key, p.ends_at
ORDER BY days_until_it_returned NULLS LAST;
Rows under 30 days go into the scorer in chapter 12. Expect to build the promotions table by hand from the email calendar the first time.
Messages, copy and briefs to adapt. Every marketing message carries a working opt-out, and consent wording needs counsel’s review.
SUBJECT Your Next Order Ships Friday
PREVIEW skip it, move it or change it in one tap
Hi {first_name},
Your next {product} ships Friday, {date}, for {price}.
Want it then? You don't need to do anything.
Not yet? [Skip this one] [Move the date]
Need a break? [Pause for a month]
Something different? [Change product or frequency]
If you'd rather stop, you can [cancel here] in one step.
{brand}
You're receiving this because you have an active subscription.
Marketing emails: [unsubscribe]. {postal address}
{brand}: your {product} ships Fri {date} for {price}. Not yet?
Skip, move or pause: {short_link}. To cancel: {short_link2}.
Reply STOP to opt out.
SUBJECT Running Low On {Product}?
PREVIEW reorder in one tap, or we'll check back later
Hi {first_name},
Most people finish {product} in about {median_days} days, and it's
been {days_since} since yours arrived.
[Reorder in one tap]
Not yet? [Remind me in two weeks]. We'll check back then, and not
before.
{brand}
[Unsubscribe] {postal address}
SUBJECT How Did It Go?
PREVIEW good or bad, it helps the next person decide
Hi {first_name},
You've had your {product} for {days} days. Would you tell other
people how it's going? Good, bad or mixed, an honest review helps
the next person decide.
[Write a review]
Totally up to you. Every customer gets this same email.
{brand}
[Unsubscribe] {postal address}
SUBSCRIPTION TOGGLE (beside the button)
( ) One-time {price}
( ) Subscribe {sub_price} every {interval} days
Most people finish one in about {median_days} days.
After your first order: {ongoing_price}. Skip, change or
cancel any time from your account or your reminder text.
CHECKOUT CONFIRMATION (above the pay button, subscription only)
[ ] I want {product} every {interval} days at {ongoing_price},
charged on shipment until I cancel. I can cancel online any time.
ADD-ON (unticked, near the order total)
[ ] Package protection, {price}: if it's lost or damaged in transit,
we'll replace it.
EMAIL SIGN-UP (unticked)
[ ] Email me when it's time to reorder, and when you restock.
SMS SIGN-UP (unticked, separate; have counsel review)
[ ] Text me offers and reorder reminders. By checking this box I agree
to receive recurring automated marketing texts from {brand} at the
number provided. Consent is not a condition of purchase. Msg
frequency varies. Msg and data rates may apply. Reply STOP to
opt out, HELP for help. [Terms] [Privacy]
1. A deadline is a real date and time, the same for every customer.
It is never extended, and the timer never resets.
2. An offer called final, last chance or ending does not return at the
same or a better price for at least 30 days.
3. Stock messages come from inventory, update live, and appear only
below a threshold we can defend.
4. Activity messages ("12 people viewing") are real and current, or
they are off.
5. Every "was" price is a price we charged openly for a substantial
period. In the EU, the prior price is the lowest of the last 30 days.
6. New apps that show urgency are checked against this first.
Owner: {name}.
OFFER what, how deep, to whom:
THE TRUE REASON shipping cutoff / batch / price change / season:
GROUPS A real deadline ({hours}h) / B no deadline / C holdout
SIZE PER GROUP customers: (sized in advance)
PRIMARY METRIC orders per customer over 60 days
GUARDRAILS unsubscribes / complaints / discount per extra order
RULE no extension; no return of this offer for 30 days
IF A BEATS B we will:
IF IT CAN'T TELL we will:
OWNER name, and the date it will be read (day 60):
Every external source, by chapter. Web sources were read in September 2026.