When Preferences Lie: The Gap Between What People Say and Do

The gap between what people say and do

Stated preferences are the cleanest data people can give you.

They’re also the easiest data to misunderstand.

A person can sincerely want “quality,” “simplicity,” “value,” or “control”… and still choose the opposite when friction appears, when context shifts, or when their attention is taxed. Digital behaviour exposes the difference between what someone endorses and what someone actually does.

Why the gap exists (and why it’s not just bad people)

People aren’t lying for sport. They’re doing something more human: predicting themselves.

When asked in calm conditions—“Which do you prefer?”—they describe a future version of themselves. But digital behaviour is never a calm condition. It’s cluttered, time-pressured, uncertain, and full of cues that quietly steer choice.

Here are the most common gap-makers:

• Friction swaps priorities. “I value X” collapses when X costs time, steps, or cognitive effort.
• Default effects masquerade as taste. Options presented first, most visibly, or easiest to click can win regardless of stated preference.
• Context changes the decision. Price perception, trust cues, urgency, and social proof re-weight the choice in real time.
• Memory is blurry. People misremember what happened last time—or confuse what they wanted with what they did.
• Social desirability nudges answers. The “right” answer can feel safer than a truthful one.

The deeper point: preferences are not the variable—constraints are

If you treat preferences as the main lever, you’ll keep building against the grain.

But what actually governs choice online is usually a stack of constraints:

• Cognitive load (how much the user must understand)
• Time-to-action (how quickly they can get value)
• Uncertainty (whether they feel safe making a choice)
• Effort asymmetry (what’s easy to do vs hard to do)
• Choice architecture (defaults, ordering, framing, visibility)
• Motivation state (bored, curious, rushed, anxious, hopeful)

So the question becomes less “Do they prefer it?” and more:

What constraints turn preference into action—or prevent it?

What this looks like in practice

You’ll see the gap most clearly where stakes are medium and attention is limited—places like:

• landing pages
• onboarding flows
• checkout and upgrade screens
• subscription prompts
• preference settings (where “later” becomes “never”)
• feature adoption (where intent doesn’t survive first use)

Often the pattern is:

1. People say they want A.
2. They see A.
3. They hesitate.
4. They pick B because B is easier, safer, or faster.

A preference statement can be “true” and still irrelevant to conversion.

The measurement shift: from “what they want” to “what they’ll do”

To close the gap, you need behavioural instrumentation and tests that reflect actual constraints.

Use behaviour to infer the real preference weights

Instead of asking what they prefer, observe what they do when:

• the page is skimmed
• the user is interrupted
• the next step is one click away vs three
• the choice is reversible vs irreversible
• the benefit is immediate vs delayed

You’re not trying to “catch” people. You’re trying to model decision physics.

Learn from “near misses,” not just conversions

Conversion is one signal. But behavioural gaps often show up earlier:

• scroll depth
• dwell time
• bounce patterns
• add-to-cart vs checkout completion
• funnel drop-offs at specific micro-steps
• repeated visits without action
• where users pause before committing
• what they search for after declining

These are preference stories written in friction.

A simple system to work with the gap

The goal isn’t to eliminate the difference between stated and revealed preference.

The goal is to design so that what people actually do aligns with what you want them to do.

Tabs: how to respond depending on what you’re holding

The practical takeaway

The next time someone tells you what they prefer, don’t dismiss them. Respect the signal.

Then ask the better question: Under what constraints will that preference still survive?

Because digital behaviour doesn’t reward ideas. It rewards decisions that happen at the speed of reality.