I recently got into a debate with a coworker about dark patterns. He's a numbers guy -- and I made the mistake of bringing up an example of a dark pattern for which I didn't have evidence of any negative consequences.
Here are two examples we talked about:
- Ads disguised as something else — like those banner ads back in the day (and sometimes even today) that looked like search boxes. These definitely increased conversion rates, but I didn't have evidence of what the harm was.
Emails or notifications that intentionally withhold information — so that they can drive "reengagement" on their actual site. An example of this is an email I received from LinkedIn when a poll I had voted on ended:
They could have displayed the survey results in the email, or at least the "winner". But instead, they just included a "View Results" button, so I'd actually have to visit the site to see the results.
I know about Net Promoter Score and the idea that a detractor can undo the positive effects of six promoters, but that didn't convince him -- he suggested that the annoyed users may not be your target audience anyway.
What evidence is there that the costs (annoyed users, lost time doing something positive, etc.) outweigh the benefits (increased conversion rates or "engagement rates")?