I've been in product reviews where dark patterns were proposed with genuine conviction. Pre-checked opt-in boxes. Fake countdown timers. Confirm-shaming. 'Skip this offer? (I don't want to save money)'. The data in these meetings always showed the pattern worked.
Here's the problem: the data window is wrong. A 2-week A/B test can show a dark pattern increasing conversion by 15%. What it can't show is what that 15% does to NPS over 6 months, or how many users talk to their friends about feeling tricked.
I've seen this play out specifically. An airline (not IndiGo) added pre-checked travel insurance. Conversion of insurance: +40%. Beautiful. Eighteen months later, their social media was dominated by complaints about 'sneaky charges'. Customer acquisition cost had increased 22% because positive word-of-mouth had dried up.
Dark patterns are a loan against your brand. The conversion is the drawdown. The compounding interest is the trust erosion you can't easily see or measure.
My standard: if explaining how a UI element works to a user would make them feel deceived — don't build it. Not because it's illegal (though increasingly it is). Because it's a bad long-term business decision.
Measure what dark patterns do to your NPS, retention, and organic acquisition over 12 months. Then decide if the conversion uplift is worth it.
