The Conversion Tactics That Cost You the Second Purchase

Your checkout conversion rate went up 8% last quarter.

You added a countdown timer, a low-stock badge, and a pre-checked shipping-protection box. The test came back green. You shipped it. The number moved.

Here’s the question nobody ran a test on: what did it do to the people who bought?

The tactics work. That’s the problem.

Let’s get the easy objection out of the way. Urgency and scarcity cues are not snake oil. They work, and the research is not close.

Barton, Zlatevska and Oppewal ran a meta-analysis in the Journal of Retailing covering 416 effect sizes across 131 studies. Their finding: cues signalling that a product is running out, or is in high demand, generally increase how much people want it. They also found it isn’t one-size-fits-all — demand-based scarcity (“300 people bought this today”) does best on utilitarian products, supply-based scarcity (“only 3 left”) does best on experiences, and time-based deadlines do best on high-involvement purchases.

A separate experiment by Koh and Seah tested four specific ecommerce patterns on 195 shoppers: the low-stock message, the activity message, the countdown timer, and the limited-time message. All four significantly increased the likelihood of choosing the product carrying them. The limited-time message was the strongest.

So no, this isn’t a post arguing that urgency doesn’t convert. It converts. That’s precisely why it’s dangerous — a tactic that didn’t work would have removed itself from your site by now.

The test measured the session. You’re running a business.

An A/B test has a boundary, and the boundary is the visit.

You measure whether this session ended in an order. You do not measure whether that person opens your next email, comes back in March, or tells someone about you. Those outcomes land outside the window, in a cohort report somebody looks at quarterly, if that.

So the dashboard can only ever tell you half the story. It reports the lift, and it is structurally incapable of reporting the cost.

Moreira, Carlos and Lopes put numbers on the other half. Their study — a focus group plus a survey of 162 consumers, published in the Marketing and Smart Technologies proceedings — found that perceived exposure to manipulative design significantly decreases trust, loyalty, and purchase intention. The qualitative half is the part worth sitting with. Shoppers reported frustration, a sense of lost control, and an active preference for brands that don’t do this.

Read those two bodies of evidence together and you get the actual finding:

The tactic lifts the order. The perception costs the relationship.

Both are true at once. They just show up in different reports, months apart, and only one of them has a green arrow next to it.

The word doing the work is “perceived”

This is the part most CRO advice gets wrong, in both directions.

The purist says never create urgency. That’s not right either — a real deadline is real information, and withholding it is its own kind of bad service. If the sale genuinely ends Sunday, say so. If there are genuinely four left, that’s a fact the shopper needs.

The damage isn’t caused by urgency. It’s caused by urgency the customer eventually figures out was fake.

And they do figure it out. They come back Tuesday and the timer has reset. They clear their cookies and “only 2 left!” becomes “only 2 left!” again. They notice the shipping-protection box was pre-checked. Nobody files a complaint. They just quietly stop being a customer, and it never shows up in an attribution report, because there’s no event for decided you were slightly cheap.

This is what makes it expensive. The cost is invisible, delayed, and lands in a different metric than the gain.

The five-question test

Both friction removal and pressure application show up green in your test dashboard. You need a way to tell them apart before you ship, because the data won’t do it for you.

Run any conversion tactic through these five.

1. Is it true?

Not “is it technically defensible.” True. If the timer is on a loop, if the stock number is a fixed string, if “127 people are viewing this” is a random integer between 90 and 200 — you already have your answer. Stop here.

2. Does it survive discovery?

Imagine the customer comes back tomorrow and sees the mechanic reset. Do they feel informed, or do they feel worked? If it only functions on people who don’t look twice, it’s a tactic with an expiry date on every individual customer.

3. Who does it serve at the moment of the click?

Free shipping over $75 serves both of you — they save money, you raise AOV. A pre-checked add-on serves exactly one party. Shared-interest tactics compound. One-sided ones deplete.

4. Does it change the decision, or only the speed of it?

Helping someone decide faster on something they wanted is service. Getting someone to buy something they didn’t want is a return, a refund, a support ticket, and a person who won’t be back. Check your return rate on the products carrying your most aggressive badges. It’s usually the fastest version of this audit you can run.

5. Would you say it out loud?

The sentence in the UI — would you say it, in those words, to a repeat customer standing in front of you? “This offer expires in nine minutes” is fine if it does. “No thanks, I don’t want to save money” is not a sentence a person says to another person.

Where this showed up this week

The clearest illustration of the whole problem didn’t come from ecommerce at all. It came from a tip screen.

Writing in Fox News on August 16, Ted Jenkin described what those iPad checkout prompts actually are: instead of charging what something actually costs, businesses have “outsourced the uncomfortable conversation to an iPad.” The menu says one thing, the checkout screen says another — and that, he argues, isn’t transparency. It’s behavioral marketing. On the design of the prompt itself, he notes those screens are built to make you hesitate before pressing “No Tip,” while an employee watches.

Every one of those prompts tested well. Average tip went up. That’s why they’re everywhere.

But look at what got built in exchange: an entire category of customer who now feels a small flinch at the moment of payment. That feeling doesn’t stay attached to the iPad. It attaches to the business.

Your checkout is the same screen with better fonts.

What to measure instead

You don’t have to give up rigour to stop doing this. You have to move the boundary of the test.

• Extend the window past the session. Judge a checkout change on the 90-day repeat rate of the cohort that saw it, not the conversion rate of the session. Slower — and it’s the only version that answers the actual question.
• Watch returns and refunds by tactic. If a badge lifts orders 6% and returns 9%, it isn’t a conversion win. It’s a rounding error wearing one.
• Run the reset test on yourself. Once a quarter, go through your own checkout as a stranger. Then do it again the next day. Anything that reset is something a customer will eventually catch.
• Separate friction removal from pressure application in your own roadmap. Guest checkout, visible shipping cost, saved payment methods, a clear returns policy — those raise conversion and raise trust. They’re the ones with no back end. Most stores have not exhausted them and have already moved on to timers.

The strategic version of this

Here’s the thing to take away, and it’s bigger than any badge.

If your conversion rate only moves when you apply pressure, the pressure isn’t the finding. The finding is that the offer isn’t carrying its own weight.

A conversion rate is a symptom. Read it as a diagnostic of three things, in this order: the offer (is what you’re selling, at that price, with those terms, worth it to this person?), the traffic (are these the people the offer is for?), and the trust (do they believe you?). Urgency mechanics don’t touch any of the three. They just borrow against the third to paper over the first.

That’s why “we ran the CRO playbook and the number didn’t move” is such a common sentence. The playbook was applied to the wrong layer. You can’t test your way out of a positioning problem, and no amount of button colour fixes an offer a well-informed buyer would decline.

Every store has a ceiling set by the offer, the traffic, and the trust. Tactics move you toward that ceiling. Only strategy raises it.

If your traffic is holding and your conversion rate isn’t — or you’ve already run the playbook and the number didn’t move — that’s the conversation worth having. Book a fit call.

Sources

• Barton, B., Zlatevska, N., & Oppewal, H. (2022). Scarcity tactics in marketing: A meta-analysis of product scarcity effects on consumer purchase intentions. Journal of Retailing, 98(4), 741–758.
• Koh, W. C., & Seah, Y. Z. (2023). Unintended consumption: The effects of four e-commerce dark patterns. Cleaner and Responsible Consumption, 11, 100145.
• Moreira, R. R., Carlos, V. L. da S., & Lopes, T. S. A. F. de O. (2026). Dark Patterns and Mindful Commerce: Portuguese Consumers’ Perceptions and the Impact on Trust, Loyalty, and Purchase Intention. In Marketing and Smart Technologies (ICMarkTech 2025), Springer.
• Jenkin, T. (2026, August 16). America needs to stop turning every purchase into an uncomfortable tip request. Fox News.