Cognitive biases in buying, with evidence and limits

What are cognitive biases in buying?

Cognitive biases are systematic patterns in judgement that can cause decisions to differ from a particular normative benchmark. In buying, the relevant pattern may involve a reference price, a gain or loss frame, or the surrounding choice set.

The label should follow a precise explanation. A customer choosing an expensive product may value its features; that is not automatically an anchoring effect. Someone declining a new service may have sensible switching costs; that is not automatically irrational loss aversion.

The broader marketing psychology field guide puts these effects alongside other research. Marketers also have biases. Teams can overvalue a memorable case study, search for evidence supporting a preferred campaign or interpret random variation as proof of a favourite theory. Applying behavioural ideas only to customers misses half the practical value.

The useful question is not “Which bias can we trigger?” It is “Which part of this decision might depend on its presentation, and how would we know?” That question leads to a testable design rather than an impressive vocabulary.

Why distinguish anchoring, framing, decoys and loss aversion?

They describe different relationships between information and choice, so evidence for one does not establish another.

Anchoring concerns the influence of an initial value on judgement. Tversky and Kahneman’s 1974 paper is a foundational source on judgement heuristics. In a marketing application, the reference point must be specified: a previous price, a competing offer or a customer’s own expectation.

Framing concerns how equivalent or closely related outcomes are represented. Tversky and Kahneman’s 1981 experiments examined preference changes under different formulations of decision problems. A commercial test must distinguish a genuine framing change from a change in the underlying information.

The decoy or attraction effect concerns changes in preference when an asymmetrically dominated option enters a choice set. Huber, Payne and Puto’s 1982 paper introduced the influential finding. Frederick, Lee and Baskin’s 2014 research documented limits, including sensitivity to how attributes are represented.

Loss aversion concerns the valuation of losses relative to gains around a reference point. Kahneman and Tversky’s 1979 prospect-theory paper is the original theoretical source. It is not a licence to assign every customer the same loss sensitivity.

These research programmes have different evidence and boundary conditions. The replication and boundary-condition questions in the decoy literature are especially important for marketing: a clean experimental choice does not establish that adding a poor plan to a real pricing page improves useful demand.

How should a marketer test a behavioural hypothesis?

Define one mechanism, change one relevant feature of the decision and measure outcomes beyond the immediate click.

1. Describe the existing decision

Record what customers know, which alternatives they see and what they are trying to achieve. Ask whether price is the main uncertainty or whether service, trust, compatibility or delivery is more important.

For a hypothetical Indian software buyer, a lower monthly price may be less important than implementation effort and support availability. A pricing experiment that ignores those concerns could produce a neat behavioural story about an incomplete decision.

2. State the proposed mechanism

Write a sentence such as: “Showing the total annual cost alongside the monthly figure may help buyers compare plans more accurately.” That is clearer than “use framing to improve conversion.”

Specify what evidence would contradict the hypothesis. If customers understand the plans no better, or choose plans they later regret, the change has not achieved the intended improvement even if initial sign-ups rise.

3. Keep comparisons honest

Use genuine reference prices and clearly comparable features. If one plan excludes an important service, show that exclusion close to the price. A comparison table should reduce the work needed to evaluate an offer, not hide material differences behind visual emphasis.

If a new option is introduced, ensure it serves a plausible customer need. A deliberately confusing offer can alter the choice set while also increasing support costs and distrust. The fact that a psychological term describes a mechanism does not make the design good.

4. Separate choice share from total demand

Measure how many people purchase as well as which option purchasers choose. A premium plan gaining share among a smaller set of buyers can look attractive in a partial report while reducing total contribution.

Inspect cancellations, returns and complaints over the relevant period. Where pricing affects ongoing use, assess whether customers can successfully use what they selected. The best immediate response may not be the best customer or business outcome.

5. Preserve uncertainty in the conclusion

Report the comparison, sample, observation period and practical size of the result. Do not call an inconclusive test a win because the direction matches the theory. Avoid searching through many segments until one produces an appealing story.

If the treatment changes several things at once, describe the result as evidence about the package of changes. It cannot establish that one named bias caused the outcome without a design that separates competing explanations.

Where do behavioural ideas go wrong in India?

They go wrong when findings from narrow experimental tasks are treated as universal descriptions of Indian consumers.

Rupee prices, household budgets, payment methods and local trust can change the decision context. A reference price meaningful to one audience may be irrelevant to another. Language can also change what counts as an equivalent frame; a literal translation may add urgency or ambiguity.

Customers beyond the metros should not be described as a single behavioural segment. Product familiarity, category experience and access to service may be more relevant than city classification. Identify the actual uncertainty instead of explaining every difference through culture.

For consequential purchases, prioritise comprehension and appropriate review. A choice that appears efficient for the seller may expose the customer to obligations they do not understand. Clear total costs and conditions should survive every creative treatment.

When should these ideas influence a campaign?

Use them when the team has a specific presentation problem and enough control to evaluate a proposed solution.

In week one, inspect customer questions and the existing choice journey. Select one hypothesis and define the outcome before launch. Run the comparison for a period appropriate to the purchase cycle and expected volume, rather than adopting a universal test duration.

Recheck after changes to the product range, prices or audience. Behavioural effects depend on context, and the comparison set is part of that context. A result from an earlier offer may not transfer to a later one.

Who should interpret the results?

A marketer and analyst should interpret the experiment together, with product, finance and customer teams checking the consequences.

The analyst protects the comparison from avoidable statistical errors. Product explains whether the options meet different needs. Finance assesses contribution. Support identifies confusion. The interpretation should remain open to ordinary explanations such as a clearer feature description or a more suitable offer.

My view

I would use behavioural research to make buying easier to understand. An honest comparison can help customers recognise value without treating them as targets for hidden tricks.

The trade-off is that a clear design may lose some purchases driven by confusion. That is a reasonable price for more informed demand. A business should be able to explain its pricing page without relying on the customer failing to notice something.

A behavioural explanation is a hypothesis to test, not a result to assume.

Start with the comparison customers actually make

Ask a few prospective customers to explain how they understand the existing options. Note the information they use and the gaps they encounter. Build the next test around that decision, with the mechanism and success measure written down first.