Neuromarketing: What the Evidence Actually Supports

What the instruments measure, and the replication problem that sits under all of it.

Search this term and you get the same four studies, in the same order, page after page.

Two of the psychology findings those pages rest on have since failed large pre-registered replications. One widely quoted statistic has no primary source. And almost none of the "neuromarketing tips" on offer involve measuring anyone's brain.

The instruments are real, though, and some published results are solid. Read the field as a way of measuring responses, one level below the choices in your brand strategy. Strategy decides who you are for; neuromarketing, at best, tells you whether the execution landed.

What neuromarketing actually is

Neuromarketing — also called consumer neuroscience — applies the measurement tools of neuroscience and psychophysiology to marketing questions. The defining feature is not the brain: it is that every method records something a person cannot accurately tell you about themselves.

People are reliable witnesses to their opinions and unreliable witnesses to their attention. That gap is why the field exists.

Read that last column again. It is where the honest version of this topic lives.

What it tells you that a survey cannot

Self-report fails in three documented places, and those three are the field's legitimate territory: attention, immediate response and reasons. Ask whether someone noticed your trust badges and they say yes; eye-tracking routinely shows page regions never fixated at all. And the reasons people give for a choice are reconstructions.

Now the side the listicles skip.

• It cannot tell you what to make. Measurement ranks options that exist; it does not generate them.
• It cannot read intent. No brain region means "will buy".
• Reverse inference is not evidence. Region lit up, therefore the person felt X — the commonest error in neuromarketing writing, and the logic runs backwards.
• It cannot replace the work above it. If the offer or the audience is wrong, a cleaner attention map changes nothing — which is what a buyer persona is for.

The honest state of the evidence

Most neuromarketing content is not built on neuroscience. It is built on behavioural psychology from the 1990s and 2000s, and that work went through a severe correction. Take tactics from it without checking which parts survived and you are acting on withdrawn ideas.

The pivotal paper is the Open Science Collaboration's "Estimating the reproducibility of psychological science", in Science (349(6251), article aac4716, 28 August 2015). A hundred studies were replicated. Ninety-seven per cent of the originals had reported significant results; 36% of the replications did, at half the original effect size.

Four favourites of the marketing-blog canon did not hold.

• Ego depletion — willpower as a depleting resource, behind "decision fatigue" copy. Hagger, Chatzisarantis and colleagues ran a 23-lab pre-registered replication, 2,141 participants (Perspectives on Psychological Science 11(4), 2016): d = 0.04, spanning zero.
• Choice overload — "fewer options sell more". Scheibehenne, Greifeneder and Todd meta-analysed 63 conditions, 5,036 participants (Journal of Consumer Research 37(3), October 2010): mean effect D = 0.02.
• Loss aversion — behind every "don't miss out" line. Gal and Rucker argued in Journal of Consumer Psychology 28(3), 2018 that the evidence does not support losses generally outweighing gains.
• Priming by incidental cues — flag and currency priming both came out at roughly zero in the Many Labs project below.

The conclusion is not "psychology is fake". It is narrower: a single study is not a tactic. One paper, one lab, one sample is a hypothesis.

Two claims to stop repeating

"95% of purchasing decisions are subconscious." This figure is repeated everywhere and traces to no primary source. Chris Harvey set out its history in Research World, ESOMAR's publication, on 15 May 2025: it comes from a popularisation of Gerald Zaltman's book How Customers Think, and there is no evidence behind the number. No study design could produce one percentage covering every human decision anyway. Remove it from your decks.

The Coca-Cola and Pepsi fMRI study. Real, and nothing like the version in circulation. McClure, Li, Tomlin, Cypert, Montague and Montague published "Neural correlates of behavioral preference for culturally familiar drinks" in Neuron 44(2), 2004. Anonymously, ventromedial prefrontal cortex activity tracked participants' behavioural preference. With the brand shown, brand knowledge changed both the stated preference and the measured response.

So the finding is that knowing the brand changes the experience — a result about branding, which is why the story attached to a product is not decoration. It is not evidence that Pepsi tastes better, or that a buy button was found. It was a small 2004 study on sugared drinks.

The findings that have held up well enough to act on

Four results clear the bar — named journals, named researchers, and either large replications or tests against real market outcomes.

Note what is absent: nothing about colours, nothing about the word "free", nothing about the brain's supposed three layers. Those are copy, not findings.

The cheap substitutes an Indian SMB can actually run

The decisions neuromarketing informs — noticed, understood, chosen — are reachable with methods that cost nothing but attention.

1. Moderated usability sessions. Six to eight people, one task each, screen shared, you silent. Jakob Nielsen's case for small samples at Nielsen Norman Group (18 March 2000): five users surface roughly 85% of usability problems, and three rounds of five beat one round of fifteen. He sets the bar higher for card sorting (15) and quantitative metrics (20) — small rounds find problems, they do not measure them.
2. First-click tests. Show the page, give the task, record only the first click. If it is wrong, the rest of the journey rarely recovers.
3. Five-second tests. Show the page for five seconds, take it away, ask what the company does and who for.
4. Attention heatmap tools. Several predict likely gaze from an image, using models trained on eye-tracking data. A cheap first pass on hierarchy — held loosely, since real eye-tracking did not beat self-report in the Venkatraman study.
5. A/B tests that measure behaviour, not opinion. The only method here that records what someone did with their money. Google Optimize is gone: its support notice states Optimize and Optimize 360 have not been available since 30 September 2023, and that page carries no last-updated date (read 5 October 2026). Our A/B testing guide covers sample size.

From our own work (first-hand, carrying no measured figure): on CRO engagements we run the five-second test before touching anything, with six to eight people from the client's own customer list rather than colleagues. The recurring finding is not subtle — people describe the hero image and cannot say what the company sells or who for. We have not quantified how often that happens and will not invent a number for it, but it reshapes more first revisions than any analytics report. Fixes then follow our landing page design guide and copywriting guide.

Inside a wider programme this is the measurement layer under our marketing services — though attention fixes change a page in days, while the organic results around it still take roughly three to six months.

The ethics line, and where India's rules land

The line is not whether you used psychology — you cannot avoid that. It is who the change benefits once the person understands it. One question: if the customer could see exactly what you did and why, would they still be glad you did it?

In India the right-hand column is not merely distasteful. It is regulated.

The Central Consumer Protection Authority notified the Guidelines for Prevention and Regulation of Dark Patterns, 2023 on 30 November 2023, under the Consumer Protection Act, 2019. They cover sellers, advertisers and platforms systematically offering goods and services in India, foreign entities included. Thirteen are named:

• False urgency — faked scarcity or deadlines.
• Basket sneaking — items or donations added at checkout unasked.
• Confirm shaming — guilt, shame or ridicule used to steer the choice.
• Drip pricing — price elements hidden until after confirmation.
• Subscription trap — cancellation made impossible or hidden.
• Forced action — unrelated purchases or sign-ups demanded to proceed.
• Interface interference — design that obscures what matters.
• Bait and switch — one outcome advertised, another delivered.
• Disguised advertisement — ads dressed as editorial.
• Nagging — repeated unrelated interruptions.
• Trick question — wording that misleads the answer.
• SaaS billing — recurring charges not clearly disclosed.
• Rogue malware — fake warnings extracting money or installs.

Enforcement has moved since. On 5 June 2025 the CCPA advised e-commerce platforms to self-audit within three months and file self-declarations (Storyboard18, 22 July 2025).

Writing in Bar & Bench on 1 September 2026, Naresh Pareek, Abhishek Nair and Shravan Kalluri of Lex Consult report that 26 platforms had filed declarations by November 2025, and that the CCPA has penalised seven platforms and issued corrective directions to two more — around ₹20 lakh in total. They name basket sneaking and default selections, drip pricing, confirm shaming and false urgency as the recurring grounds. We could not reach primary CCPA orders this run, so treat that as their reported tally, not a verified official total.

Four of the thirteen are standard items in "neuromarketing tactics" listicles, and the regulator has started naming companies. A tactic that works once and produces a refund, a review and a notice is not a tactic.

Where to go from here

1. Audit your pages against the thirteen. Be honest about the countdown timer and the pre-ticked box. Forty-five minutes, highest value here.
2. Run a five-second test, then a first-click test — six people from your customer list, on the page where a wrong first click costs most, against the stage it serves in your customer journey.
3. Turn the best finding into an A/B test, so the decision ends in behaviour, not an opinion about an opinion.