The research on friction is consistent across contexts: small changes to the effort required to perform a behaviour produce disproportionately large changes in whether that behaviour occurs. The disproportionality is the key finding. The effort cost of an additional click in a checkout process is objectively trivial — it takes less than a second. The abandonment rate it produces is not trivial. The disconnect between the objective effort cost and the behavioural consequence is explained by how the System 1 evaluation of effort works: it is not calibrated to objective time or energy expenditure, it is calibrated to relative ease, and relative ease is as sensitive to the addition of one step as to the addition of ten.

The System 1 effort-cost mechanism

Kahneman’s (2011) dual process framework establishes the foundational mechanism. System 1 continuously evaluates the effort cost of available actions and automatically biases toward the lowest-effort option. This evaluation is not deliberate; it is automatic, rapid, and operating below the threshold of conscious decision-making. When an action requires more effort — more steps, more navigation, more information to enter, more decisions to make — the System 1 effort-cost evaluation registers the increase and biases toward alternatives that require less effort.

The consequence that makes friction so commercially significant is that this evaluation is disproportionately sensitive to small changes. The additional form field, the additional confirmation screen, the additional click — each of these increases the perceived effort cost by more than its objective time cost warrants, because the System 1 evaluation is not measuring seconds but registering friction as a signal about how easy or difficult the action is to complete. The signal from a single additional step is “this is harder than the default path,” and that signal is sufficient to shift the behaviour toward the easier alternative.

The meta-analytic evidence for friction’s commercial magnitude

Mertens, Herberz, Hahnel and Brosch’s (2022) meta-analysis confirmed that friction interventions produce effect sizes across commercial and policy contexts that are comparable to or larger than financial incentive interventions of equivalent monetary value. The finding is important precisely because it challenges the assumption that behaviour change requires financial incentive: changing the ease of a behaviour, rather than its financial cost or benefit, produces comparable behavioural effects at substantially lower cost.

Friction removal from desired behaviours — simplification of onboarding, one-click purchase, pre-populated forms, reduced navigation — consistently produced the largest conversion improvements per intervention dollar. Friction addition to unwanted behaviours — extended cancellation processes, required confirmation steps, cooling-off periods, multi-step opt-out designs — consistently produced the largest retention improvements. The dual application is what makes friction the most powerful single lever in choice architecture: it can be deployed in both directions simultaneously, increasing the conversion rate on desired behaviours while reducing the exit rate on retained customers.

The Amazon one-click patent and what it established commercially

Amazon’s one-click purchase mechanism is the most commercially documented friction removal intervention available. The standard checkout process required the customer to navigate through a cart, a shipping address confirmation, a payment method confirmation, an order review screen, and a final confirmation. Each of these steps was a potential abandonment point — a moment where the customer’s purchase intention had to survive a friction encounter. The one-click option removed all five screens from an alternative purchase pathway, while leaving the full checkout available.

The conversion improvement was substantial enough that Amazon filed and defended a patent on the mechanism — establishing that the competitive value of the friction reduction was sufficient to make it worth protecting through IP. The one-click mechanism is not the result of improved product quality, better pricing, or more compelling communication. It is the result of removing five screens of friction from the path between purchase intention and purchase completion.

The dual application: friction removal and friction addition operating together

The most commercially sophisticated friction design deploys both directions simultaneously. Netflix’s autoplay feature removes the decision friction that would otherwise interrupt viewing between episodes — the transition between “this episode has ended” and “the next episode is beginning” eliminates the decision point where the viewer might choose to stop watching. A single removed decision point produces the binge-watching behaviour pattern that is Netflix’s primary engagement mechanism. The viewer who would have stopped at the end of an episode if prompted to make a decision continues because no decision is required.

The organ donation opt-out framing can be read as a friction design: in opt-out countries, non-donation requires active friction expenditure — the deliberate registration of a preference to not donate. The friction prevents the non-donation for the majority of people who hold moderate or indifferent preferences and would not expend the effort to opt out. The 83-percentage-point difference between opt-in and opt-out donation rates is partly a default effect and partly a friction effect — the two mechanisms are in practice inseparable, because the default defines the path of least resistance, and the path of least resistance is what System 1 takes.

Friction in onboarding and habit formation

Wood and Rünger’s (2016) habit formation research establishes a specific implication for product onboarding: habits form faster when friction is lower, and are maintained more easily when the habitual environment remains stable. The product that reduces friction at the initiation point of its habitual use — the trigger moment — is reducing the effort cost precisely at the point where the habit association is being formed. Lower friction at the initiation moment means the behaviour is more likely to be performed on each trial; more frequent performance across a consistent context means the habit association forms faster; faster habit formation means less dependence on maintained motivation and more dependence on the automatic contextual trigger.

The onboarding design implication is that friction should be minimised specifically at the moments that will become the habitual initiation points — not uniformly across the entire product, but at the specific trigger-action sequences that the product needs to habituate. Friction elsewhere in the product may be acceptable or even valuable; friction at the habit initiation point delays the formation of the automatic association that sustainable engagement requires.

The ethical dimension of friction design

The GDPR cookie consent guidance that requires equal prominence for “accept all” and “reject all” options is explicitly friction-aware: it acknowledges that commercial interests would naturally deploy asymmetric friction to drive consent, and regulates against it. This is the regulatory acknowledgment that friction design is a powerful enough behaviour influence tool to warrant explicit ethical and legal attention.

The ethical framework for friction design parallels the one established for loss aversion and default design: friction that accurately represents the genuine cost structure of an action — a cancellation process that reflects the genuine steps required — is legitimate. Friction that is manufactured specifically to impede a behaviour that the user would otherwise readily perform — a cancellation process designed to be more onerous than operationally necessary — is a dark pattern that exploits the System 1 effort-cost evaluation to override the user’s genuine preferences. The mechanism is identical; the ethical distinction is whether the friction is reflecting or manufacturing the cost.

The practical standard is whether the friction in a cancellation, opt-out, or rejection process reflects the genuine operational requirements of that process, or whether it has been added specifically because it will deter the behaviour. The first is acceptable; the second is the friction equivalent of manufactured scarcity.

Books worth reading on this

The Effortless Experience by Matthew Dixon, Nick Toman and Rick DeLisi is the most directly applicable available account of friction in the customer service and retention context — documenting through large-scale research that reducing customer effort is a stronger predictor of loyalty than exceeding expectations, and providing the most operationally specific available treatment of where friction accumulates in customer journeys and what removing it produces. For the entrepreneur who wants the most practically structured available account of how friction operates in the post-purchase customer experience, Dixon, Toman and DeLisi’s research provides the most directly applicable available complement to the choice architecture framework.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Kahneman, D. (2011), Thinking, Fast and Slow, Farrar, Straus and Giroux. Mertens, S., Herberz, M., Hahnel, U.J.J. & Brosch, T. (2022), The Effectiveness of Nudging: A Meta-Analysis of Choice Architecture Interventions Across Behavioural Domains, PNAS, 119(1), e2107346118. Thaler, R.H. & Sunstein, C.R. (2008), Nudge, Yale University Press. Johnson, E.J. & Goldstein, D.G. (2003), Do Defaults Save Lives?, Science, 302(5649), 1338–1339. Wood, W. & Rünger, D. (2016), Psychology of Habit, Annual Review of Psychology, 67, 289–314. Gollwitzer, P.M. & Sheeran, P. (2006), Implementation Intentions and Goal Achievement, Advances in Experimental Social Psychology, 38, 69–119. Dixon, M., Toman, N. & DeLisi, R. (2013), The Effortless Experience, Portfolio. Eyal, N. (2019), Indistractable, BenBella Books.