Why a confirm dialogue wording decides the click
A cancel button labelled with self-criticism is harder to press than one labelled no. The option is present and its cost has been raised.
Intuition test — answer before you read on
A decline button reading No thanks, I prefer to stay uninformed is pressed less than a plain No. Why?
Two buttons. One reads Subscribe. The other reads No thanks, I prefer to stay uninformed. The second option is fully available, clearly visible and free to choose, and it is chosen less often than a plain No would be. Nothing was hidden and nothing was forced.
What everyone sees
The dialogue looks like a fair choice, and defenders point out correctly that both options are present and legible. Users who decline report mild irritation and do decline. The measurable shift comes from the users near the margin, who are not conscious of having been moved.
What is actually happening
Confirmshaming attaches a social or self-image cost to the declining option, so the two branches stop being equivalent in anything but function. Work on dark patterns by Gray and colleagues catalogues the technique, and Mathur and colleagues measured its prevalence across thousands of shopping sites. The mechanism draws on the same self-consistency pressure that makes people reluctant to describe themselves in unflattering terms even privately.
Why it stays hidden
The manipulation hides because it is legible as humour. The wording is playful, so objecting appears humourless, and the designer can honestly describe it as brand voice. Regulators and users both look for removed or hidden options, and nothing has been removed — the cost was added to a label, which is not where anyone is checking.
Both options remain, and one has been made expensive to say. Price can be added to a choice without touching the price.
Both options remain, and one has been made expensive to say. Price can be added to a choice without touching the price.
Collect this card
Both options remain, and one has been made expensive to say. Price can be added to a choice without touching the price.
0 / 10,000 collected
Sources & further reading 3
- Gray et al., "The Dark (Patterns) Side of UX Design", CHI Conference, 2018
- Mathur et al., "Dark Patterns at Scale: Findings from a Crawl of 11K Shopping Websites", CSCW, 2019
- Brignull, "Dark Patterns: Deception vs. Honesty in UI Design", 2011
Cross-references
Related files
-
Hidden Mechanics Entry #0009Why elevator close buttons do nothing
In most lifts installed since the early nineties, the door-close button is disconnected for everyone except firefighters. It still lights up. That is the point.
NoviceThe hidden part #0009The button is not wired to the door. It is wired to your patience.
Mechanics Open file -
No visual on filePlatform Mechanics Entry #0286
Why moderation is a pricing problem, not a values problem
At scale, every piece of content must be reviewed at a cost that approaches zero. The cost determines what gets caught, and the values follow the budget.
NoviceThe hidden part #0286Values are free to declare. Moderation costs money per decision. At scale, the budget determines what the values actually mean.
Platforms Open file -
No visual on filePlatform Mechanics Entry #0282
Why recommendation systems narrow taste while widening catalogue
The library grows every month and the same titles keep arriving at the top. Availability and discoverability are separate quantities that can move apart.
NoviceThe hidden part #0282Catalogue size and reachable set are different numbers. Recommendation confidence is built from attention already spent.
Platforms Open file -
No visual on fileHabit Loops Entry #0303
Why changing a route changes a routine
A habit feels like a decision stored in the person, but it is actually stored in the environment. Move the person and the habit loses its cue.
NoviceThe hidden part #0303Habits are stored with their surroundings. Change the surroundings and the habit has to be rebuilt from scratch — or it simply does not fire.
Habit Open file
Annotations are reserved for archive members.
Sign in to annotate