Behavioural economics, nudges and defaults in marketing

What is choice architecture?

Choice architecture is the way a decision is arranged: which options appear, how they are described, what happens if someone does nothing and how difficult it is to change course.

This is one practical area in the wider psychology of marketing. Every interface has some arrangement. A form must put fields in an order, a comparison must choose attributes and a service must decide what happens when a customer makes no selection. The existence of those decisions does not mean any arrangement is equally acceptable.

A nudge changes part of this environment while preserving meaningful choice. It might make a useful next step easier to understand or draw attention to information people overlook. A price discount, a ban and a deceptive obstacle are different interventions even if all change behaviour.

For marketers, the useful starting point is a customer difficulty. People may forget an intended action, struggle to compare plans or misunderstand a renewal date. Naming that difficulty is more informative than declaring that the campaign needs more behavioural science.

Why are defaults influential but not universally effective?

Defaults can affect choices through effort, interpretation and inertia, but the strength and meaning of the effect depend on the setting.

Johnson and Goldstein’s 2003 paper examined defaults in organ-donation decisions. It is a foundational source, but a consequential public-policy choice is not equivalent to selecting a retail bundle. The research should inform questions about a default rather than supply a sales forecast.

DellaVigna and Linos’s 2022 study of nudge units compared evidence from interventions at scale with the academic literature. It highlights why implementation setting and the evidence available for publication matter when estimating practical effects.

The broader literature is contested. Mertens and colleagues’ meta-analysis reported evidence on choice-architecture interventions, while Maier and colleagues’ 2022 reanalysis challenged the aggregate conclusion after adjusting for publication bias. The disagreement is a reason to avoid a universal “nudges work” claim, not proof that every individual intervention is ineffective.

Replication and scale are separate questions. Repeating a procedure tests whether its result holds under specified conditions. Deploying a related intervention in a new market tests transfer and implementation. Both matter, but they should not be described as the same evidence.

How do you design a useful behavioural intervention?

Start with observed friction, choose a transparent change and evaluate whether it improves the customer’s decision as well as the business outcome.

1. Identify the intended action

Ask what the customer is trying to do and why the current journey makes it difficult. Use support questions, observation and interviews. Avoid assuming that a low completion rate means people need stronger pressure; some may have reasonably decided the offer does not fit.

Distinguish inability, uncertainty and lack of interest. A reminder may help forgetting. A clearer comparison may help uncertainty. Neither is a good solution to an unsuitable product or an unaffordable price.

2. State the mechanism

Describe how the proposed change addresses the problem. “Show the delivery conditions before payment so customers can evaluate them earlier” is specific. “Use choice architecture to increase conversion” is not.

List alternative explanations for a possible result. A revised page may work because it loads faster, uses clearer language or shows a better offer. Naming a behavioural mechanism does not eliminate those possibilities.

3. Make the choice visible

If a default is used, explain what is selected and let people change it without unnecessary effort. Give consequences and costs the prominence needed for understanding. Avoid making acceptance visually obvious while hiding refusal in ambiguous wording.

For consequential commitments, an active choice may be more appropriate than a default. The decision should be reviewed in its legal and customer context. A convenient implementation is not sufficient justification for preselecting an obligation. Check the distinction between helpful design and dark patterns that impair customer choice.

4. Choose the outcome before the test

Measure the action of interest, but also assess comprehension, regret, complaints and later completion where relevant. A form that becomes easier to submit can still collect poor information or create unwanted follow-up.

Specify the comparison and observation period. Do not select only the metric that improved after seeing the results. If the intervention affects repeated behaviour, give the evaluation time to observe that behaviour rather than relying entirely on an immediate response.

5. Test the actual implementation

Review the live design on relevant devices and in each language. A theoretically transparent default can become unclear when the explanation is truncated on mobile. A reminder can become intrusive when several systems send it independently.

Check exceptions. What happens if the user leaves halfway through, returns later or changes a preference? A good design should preserve the person’s choice consistently, rather than restoring the preferred business default at the next visit.

6. Keep a record of unsuccessful tests

Document the design and result even when the intervention does not help. Otherwise the organisation creates its own publication bias, remembering a handful of successful experiments and forgetting the denominator of attempts.

An unsuccessful test can still be useful if it eliminates a plausible explanation. It can also reveal that the team targeted a minor friction while overlooking a larger product or service problem.

Where does choice architecture matter in India?

It matters in mobile forms, product comparisons, renewals, payments and service journeys where small presentation choices can affect understanding.

Indian audiences use different languages, devices and support channels. A concise English label may not fit or carry the same meaning after translation. Test whether customers understand the action, not merely whether the translated words are grammatically correct.

For customers with intermittent connectivity, preserving entered information may be more valuable than adding persuasive copy. For a service with local availability constraints, showing serviceability early may prevent wasted effort. These are proposed design improvements to test, not claims of universal effect.

Avoid using broad categories such as “Bharat users” as a substitute for research. Identify the relevant conditions: device, language, familiarity with the service and access to assistance. A useful intervention addresses those conditions directly.

When should a nudge be changed or removed?

Change or remove an intervention when it does not improve the intended outcome, creates misunderstanding or no longer fits the customer’s situation.

In week one, document the problem and the current journey. Prepare a modest comparison and review the implementation before launch. Continue only after enough relevant observations exist to support a decision. The required duration depends on the event and volume, not on a fixed behavioural-science timetable.

Revisit defaults when products, costs or policies change. A selection that was reasonable for an earlier offer can become unsuitable after the terms evolve. Customers should not inherit a consequential change simply because an old interface setting remains in place.

Who owns the customer outcome?

The product or service owner should own the outcome, with marketing, research, engineering and relevant specialists sharing explicit responsibilities.

Marketing helps explain the choice. Research tests the mechanism and outcome. Engineering preserves the selected preference. Customer support reports confusion. Legal reviewers assess consent and obligations where relevant. The accountable owner should resolve conflicts between immediate conversion and informed customer choice.

My view

I would use behavioural economics first to remove avoidable difficulty. Clear comparisons, useful reminders and understandable defaults can help without turning the customer journey into a contest of pressure.

The trade-off is that a transparent design may reveal that some people do not want the offer. That is valuable information. A nudge should help people carry out a reasonable intention, not manufacture agreement that disappears once the consequences are understood.

Better choice architecture makes the decision easier to understand and easier to own.

Start with one misunderstood choice

Find a decision that repeatedly generates support questions. Observe how customers interpret it, then test a clearer arrangement. Record both the commercial result and whether people understand what they chose.