What neuromarketing actually costs, and what skipping it costs more.

Most new products fail after the research is done

Clayton Christensen, Scott Cook and Taddy Hall published "Marketing Malpractice: The Cause and the Cure" in the Harvard Business Review in December 2005. Their finding is direct: roughly 30,000 new consumer products launch every year, and over 90 percent of them fail. The number has circulated in marketing decks for two decades, usually cited as a warning about how hard it is to predict what people want. It deserves a closer read, because the detail most decks skip is the one that matters most.

Every one of those products had a marketing team behind it. Most of those teams had already commissioned research: focus groups and surveys measuring purchase intent. The budget for understanding the customer had been spent, and spent as intended, before the failure happened. Nine launches in ten still failed anyway.

Christensen, Cook and Hall are describing what happens when research measures the wrong layer of a decision. People are asked to explain preferences they cannot fully access, and the explanation gets trusted as if it were the decision itself. We go through that mechanism in detail in why surveys lie: people are fluent narrators of choices they never consciously made, and a conventional survey records the story they tell about the choice, a different thing from the choice itself. The 30,000-product figure is what that gap costs at scale, whether or not a company spent freely on research along the way.

Christensen's broader argument in that same article is that most of this research asks the wrong question from the start. Companies segment customers by who they are, age and income chief among the variables, and then build products for the segment instead of for the job the customer is trying to get done. A survey can describe a segment with precision and still miss the reason a specific person reached for a specific product. That mismatch, between who a person is and what they hired the product to do, is part of why that ninety percent stays so high.

What a study like this costs

Cost is the first question any budget holder asks, and it is a fair one. Orbital's own process is a reasonable proxy for what a study involves in practice. A tailored brief arrives within 24 hours of a written question about what needs testing and why. A full study cycle, covering recruitment, lab sessions, multimodal data collection, analysis and recommendations, typically runs three to five weeks, depending on scope and sample size.

Speed matters as much as price in a budget conversation. A launch date is usually fixed well before anyone commissions a study, so the test has to fit inside the time already on the calendar rather than add to it. A brief inside 24 hours and a result inside three to five weeks fits inside most campaign timelines, which is the practical reason speed gets asked about as often as cost.

What moves is the scope, and scope is where the real cost lives. A single ad tested before a media buy commits is a different job from a multi-market brand-equity tracker run across several territories and re-measured over time. Cost is scoped to the decision being made, ranging from a focused single-asset pre-test up to full multi-market brand-equity tracking. A useful way to read a quote is as a function of what is riding on the answer: a small decision earns a small study, and a national relaunch earns a larger one, scaled to the size of the bet rather than the size of the client asking.

Cost tracks the size of the decision rather than the size of the company asking, which is worth stating plainly, since it answers most objections before they surface. Measured against the numbers in the previous section, even the larger end of that range sits in a different order of magnitude from what a single failed national launch spends on production and media before anyone notices a problem.

The evidence for a better instrument

Two figures are worth putting in the same room. A 2015 meta-analysis from the Advertising Research Foundation found that biometric measures of advertising effectiveness predicted in-market sales lift roughly twice as well as the best survey methods. The finding came from an industry body auditing a wide base of published studies, which carries more weight than any single vendor's case study. Separately, Orbital's own benchmarks suggest roughly a 5x signal lift over surveys on emotion and intent, the two inputs most survey instruments handle worst.

Read together, the two findings answer different questions. The ARF figure says biometric measurement out-predicts survey measurement on the outcome a business cares about: sales lift rather than sentiment. The Orbital figure says the size of that advantage grows again on the two inputs, emotion and intent, that most influence which creative concept wins in market. Surveys still do a competent job on stated awareness, claimed usage, demographics and shifts in declared attitude, the questions they were built to answer. They do a poorer job on the questions a launch budget depends on: whether the creative gets noticed at all, and whether it is still remembered at the point of purchase.

The business case a budget conversation needs

Put the two sections above next to each other and the case makes itself. A pre-test is priced against the size of the decision: a single-asset study at the low end, a full brand-equity programme at the high end, quoted within 24 hours and delivered in three to five weeks. A failed launch is priced against everything that shipped before the market delivered its verdict: the production run and the media plan booked around it, plus whatever it cost to get the product onto shelves. By the time a launch is visibly failing, that spend is already gone.

A pre-test sits ahead of that spend, earlier and smaller, functioning as insurance in the plain sense of the word: a modest, early cost that lowers the odds of a much larger, later one.

A pre-test is priced against the decision it informs. A failed launch is priced against the media and distribution spend that already went out the door.

The comparison only holds if someone acts on the result. Testing a concept and then greenlighting the version that tested worse, because someone senior prefers it, is a familiar failure mode, and it spends the cost of research while forfeiting the benefit. Creative judgement chooses what a brand says and how it says it. A pre-test's job sits underneath that choice: it tells the team which version of the work is likely to land before the media budget commits, and leaves the choice itself where it belongs.

None of this requires a precise dollar figure to make the point. A study quoted inside 24 hours and delivered inside five weeks, scoped to a single asset or a full campaign, sits in a different order of magnitude than a national media plan. A budget conversation that prices the study and never prices the alternative is pricing half the decision.

Smaller markets, smaller margin for error

Market size changes the arithmetic further. A national launch in a market of two or three million consumers does not have the luxury of a slow regional rollout, reading early signals in one city before committing everywhere else. A miscalculated launch in a smaller Caribbean market is often the national result within weeks, because there is no larger domestic market left to iterate into.

The same mistake, made by a brand launching across a market of tens of millions, has more room to recover. A weak region can be recut and relaunched elsewhere in the same country while the balance sheet absorbs the difference. A market that cannot iterate its way out of a bad launch is exactly the market that benefits most from finding the problem while it is still cheap to fix.

The constraint sharpens the case for testing first rather than weakens it. A market that cannot afford to get a national decision wrong is precisely the market where paying a small, early cost to avoid a large, late one matters most.

The same logic applies inside the region, where island markets differ from each other in culture and retail structure as much as they differ from any market outside the Caribbean. Jamaica and Trinidad do not share a single playbook, so a study built for one does not transfer cleanly to the next. Testing once and rolling the result out across the wider Caribbean repeats the same mistake Christensen described at a global scale, trusting a result from elsewhere to answer a question about here.

What testing cannot promise

None of this makes a pre-test a guarantee, and it is worth saying so plainly. A study can find that a concept lands emotionally, tests well on memory and intent, and still watch that concept fail in market because a retailer will not give it shelf space, or because the launch lands in a season nobody chose for it. Price is its own separate risk: a competitor's discount in the same month can undo a launch that no pre-test was ever positioned to see.

The same limit applies to other forms of testing. We make a related argument about click-based experimentation in the limits of A/B testing: a method that measures one layer of a decision well can still miss the layers above and below it. A pre-test's scope stops at what people notice and remember. Whether the sales team can get the product onto a shelf next month is a separate question, and no amount of EEG data answers it.

The honest claim for this kind of testing is a reduction in the risk a pre-test can see: the risk that creative goes unnoticed, or that it is forgotten by the time a shopper reaches the shelf. That share of the failure rate is large enough to move the Christensen, Cook and Hall number when it is addressed properly. It is one share among several. A study that finds a strong emotional response earns confidence going into a launch, while the parts of a launch it cannot see remain exactly as risky as they were before anyone tested anything.

The question this piece opened with tends to show up alone in a budget meeting: what will the study cost. The question that decides the argument rarely gets written down anywhere: what does it cost to skip the study and find out in market, priced in the production and media spend already committed, plus whatever a season of distribution adds on top. Start with a brief and Orbital will scope a study to the size of the decision in front of you, whether that is a single ad or a full relaunch.

Frequently asked questions

What percentage of new products actually fail, and why does that matter here?

Clayton Christensen, Scott Cook and Taddy Hall reported in the Harvard Business Review in December 2005 that of the roughly 30,000 new consumer products launched each year, over 90 percent fail. It matters here because most of those launches had already been through conventional research: focus groups and surveys measuring purchase intent. The failure happened after that work was done, which is the starting point for the argument in this piece.

Does conventional market research already protect against these failures?

Only for the questions it is built to answer. Surveys and focus groups are reliable for stated awareness, claimed usage, demographics and shifts in declared attitude. They are weak at predicting which creative concept people will respond to, because respondents describe decisions they never consciously made, which is the mechanism covered in why surveys lie. That gap is why the Harvard Business Review failure rate holds even in categories where research budgets were large.

How much does a neuromarketing study actually cost?

Cost is scoped to the decision being made. Orbital quotes within 24 hours of a written brief, with study designs ranging from a focused single-asset pre-test through multi-stage creative optimization to full multi-market brand-equity tracking, and a full study cycle typically runs three to five weeks depending on scope and sample size. The size of the decision sets the size of the study, so a single ad and a national relaunch sit at very different points on that scale.

Is neuromarketing worth it for a single ad or a small launch, or only major campaigns?

It scales down as readily as it scales up. A focused single-asset pre-test exists for exactly this case: a single ad or a single pack redesign, tested before the media budget behind it commits. Orbital's own benchmarks suggest roughly a 5x signal lift over surveys on emotion and intent, and that benchmark applies to a single-asset study as much as to a full campaign.

Can testing guarantee a product or campaign will succeed?

No. A pre-test reduces the risk it is built to measure: a concept that leaves people cold, or an ad that goes unnoticed and unremembered. It has no way to measure distribution or price, and no way to catch a launch that lands in the wrong season. A well-tested product can still fail for any of those reasons. The honest claim is risk reduction on the part of a launch a pre-test can see.