Context Collapse: Why Heuristics Work Differently in the Real World

The laboratory is a lie—not intentionally, but structurally. When Kahneman and Tversky documented how people rely on mental shortcuts to navigate uncertainty, they were observing behaviour in controlled conditions where context had been deliberately stripped away. The heuristics they identified—availability, representativeness, anchoring—are real. But the way they operate in the messy, layered reality of actual customer decisions is fundamentally different from what the research suggests.

The problem isn't the heuristics themselves. It's that we've been treating them as universal rules when they're actually context-dependent tools that shift meaning depending on what surrounds them.

The Thing Everyone Gets Wrong

Most teams applying behavioural science to product and CX decisions treat heuristics as fixed cognitive patterns. A customer sees a price, anchors to it, and their valuation is locked. A product is available in memory, so it feels representative of its category. The logic is clean, testable, and mostly wrong in practice.

What actually happens is that heuristics don't operate in isolation. They collide with competing information, emotional state, social signals, and prior experience. A price anchor only works if the customer hasn't already formed a strong reference point through lived experience. Availability bias only drives choice if the available option is also contextually appropriate. When these conditions aren't met—which is most of the time in real products—the heuristic either weakens or inverts entirely.

Consider how a customer evaluates a subscription service. In a lab study, you might show them a price and measure anchoring effects. But in their actual decision, they're simultaneously processing: whether they've used similar services before, what their friends are paying, whether they trust the company, how the pricing compares to their current spend, and whether they even need the service. The anchor exists, but it's competing with a dozen other reference points. Context collapses the heuristic's power.

Why This Matters More Than People Realise

Teams often invest heavily in optimising for a single heuristic—making something more "available," creating a stronger anchor, increasing perceived similarity to a trusted category. But without understanding the full context in which the decision happens, these optimisations either fail or create friction.

A classic example: companies try to increase product adoption by making their offering more "available"—more prominent, more visible, more frequently mentioned. But if the customer's actual decision context includes scepticism about whether they need the product at all, availability becomes noise. It might even trigger reactance. The heuristic that was supposed to help actually creates resistance.

The deeper issue is that context collapse reveals something uncomfortable: customers aren't following the decision rules we've documented. They're making sense of their specific situation using whatever mental tools seem relevant. The heuristics are there, but they're subordinate to context, not dominant over it.

This changes what you should actually measure. Instead of testing whether a heuristic "works," you need to understand whether it's relevant to the decision context your customer is actually in. That requires moving beyond lab-style A/B tests and into observation of real decision-making—how people actually talk about choices, what information they seek, what they ignore, what surprises them.

What Actually Changes When You See It Clearly

Once you accept that context shapes how heuristics operate, your approach shifts. You stop trying to engineer heuristics and start designing for the contexts where they're naturally relevant.

This means building products and experiences that help customers understand their own context first. What's their reference point? What do they already know? What are they actually trying to solve? Only then do the heuristics become useful—not as manipulation, but as natural shortcuts within a decision framework that makes sense to them.

It also means accepting that some heuristics won't work in your context, no matter how elegantly you apply them. That's not a failure of the science. It's a sign that you're finally seeing the decision as your customer does.