The litmus test: would customers choose this?

Some things no customer would ever choose. Popups that interrupt a task. Terms that let their data be sold. Subscription flows that take one click to enter and a maze to exit. Upgrade pressure in the middle of normal use. Yet all of them ship, constantly, because each one wins a metric somewhere. The litmus test is a single filter that catches them: if customers wouldn't voluntarily choose it for themselves, it is suspect by default.

Why it matters to the business

Manipulative design is now common enough to be an industry. A Princeton crawl of roughly 11,000 shopping sites found dark patterns on more than one in ten, plus 22 vendors selling them as turnkey services. It works in the short term, which is exactly why it spreads: Luguri and Strahilevitz's research found even mild dark patterns more than doubled sign-ups for a dubious service. But the bill arrives from two directions. Customers first: PwC found 32% will walk away from a brand they love after a single bad experience. Then regulators: FTC settlements include $520 million from Epic Games over dark-pattern billing and $2.5 billion from Amazon over its Prime cancellation flow.

How to use it

• Add the question to every launch review: would a customer voluntarily choose this? Record the answer.
• Audit subscription, consent, and cancellation flows for asymmetric friction: easy in, hard out.
• Make cancellation as easy as sign-up; that is the standard regulators now enforce.
• Ground the test in real research, interviews and task observation, so "what customers would choose" is knowledge rather than projection.
• Give one senior owner veto power over experiments that trade customer trust for conversion lift.

Where teams get it wrong

The trap is the mild pattern. Luguri and Strahilevitz found aggressive manipulation triggers customer backlash, but mild manipulation, like a pre-ticked box or a guilt-trip decline button, produces no complaint signal at all. Dashboards show conversion up and nothing wrong, so the practice spreads quietly until a regulator or a churn spike prices it for you.

Ask your team

• Which of our current flows would fail the test, and where do we profit from customers not noticing something?
• How many clicks does it take to subscribe, and how many to cancel?
• Who has the authority to kill a winning A/B test because it manipulates customers?

If customers wouldn't voluntarily choose it for themselves, it is suspect by default.