It’s been a rough few days for neuromarketers. First, Matt Wall of Slate wrote a thoughtful article, What Are Neuromarketers Really Selling?. Then, PopSci jumped on the bandwagon and writer Shaunacy Ferro published Why Neuromarketing Is A Neuroscam. Ferro quoted Wall’s article but added an even more provocative headline.
The subtitle of Wall’s piece is, “The poor data and shoddy logic behind a hyped business boom” – certainly sounding like an indictment of the fledgling neuromarketing industry. But is this warranted?
In fact, despite the provocative headlines, Wall’s article focuses on one topic, the reliability of EEG data gathered using headgear with a small number of dry contacts. Within this scope, Wall raises some good points. Academic studies using EEG don’t use the convenient little gizmos one can pop on a subject’s head in seconds. Rather, they use messy conductive gels and large numbers of contacts to gather more data and better data.
Wall is also correct in writing that many neuromarketing firms base their techniques and claims of effective ad or product studies on their unpublished proprietary data, making it difficult for outsiders to verify these claims. I’m sympathetic with this sentiment – one reason for the ongoing credibility gap in neuromarketing is that most industry players have forged ahead without peer-reviewed studies or even publicly accessible data. And, when neuromarketing studies are disclosed, the data is usually anecdotal (see, for example, Neuromarketing at New Scientist).
The disconnect in the Slate piece (and the subsequent PopSci rehash) is between the attention-getting headlines and the content of Wall’s narrowly focused article. The truth is that neuromarketing encompasses far more than dry-contact EEG studies. Biometric measurements, eye-tracking, facial coding (both automated and manual), and other technologies are being used both alone and in combination to gauge customer reactions to products and ads. Wall mentions fMRI as if it is solely the province of academia. In fact, even that technology, which allows 3D imaging of brain activity vs. the surface measurements of EEG, is being used in neuromarketing studies.
Data is emerging that suggests an even brighter future for neuromarketing. Carnegie Mellon researchers recently published work showing that specific emotions could be identified from brain activity.
Commercial acceptance of neuromarketing is growing as well. At Forbes, I wrote about Coke’s adoption of neuromarketing for all projects involving quantitative ad studies.
When I spoke at the Neuromarketing World Forum in Sao Paulo a few months ago, one of the most exciting aspects of the conference was that a variety of neuromarketing companies are starting to make some of their data and results public – not quite the same as being published in Neuroscience, but welcome progress.
Personally, I like an even broader definition of neuromarketing that includes the use of brain and behavior research to improve marketing. That is, my view of neuromarketing encompasses the development of rules and techniques based on how humans process information and behave rather than studies of specific ads or products.
In short, Wall raises some good points, but the headline writers at both Slate and PopSci ignored the details of the article in striving for punchy, click-worthy headlines. Let’s call it Headlinescam…
Thanks for the dose of realism on this topic, Roger. I had the same reaction. Wall’s article was mostly about a very specific topic, dry electrode headsets, which have been subjected to critical reviews before (for example, Thomas Ramsoy’s post on data from the Emotiv headset at http://bit.ly/XbBOfM). But the PopSci article jumped the shark when it reinterpreted Wall as saying “neuromarketing doesn’t produce solid data–and perhaps it never can.” How did dry electrodes become all of neuromarketing?
Both Wall and PopSci also talk about reverse inference, which is being cited more and more these days to beat up on neuromarketing. We discuss this at length in Neuromarketing for Dummies (coming out Aug 5!). The fact that reverse inference does not product results that are logically true by definition does not mean it always produces results that are false. It is a probabilistic inference that can have a greater or lesser likelihood of being true, given the care with which the experiment has been constructed (see Hutzler, “Reverse inference is not a fallacy per se” in Neuroimage, http://bit.ly/URjPF0). It will be fun trying to get journalists to understand this.
Finally, I second your call for a broader definition of neuromarketing. I think it’s very important for neuromarketing to embrace decisions and actions, not just reactions in the brain. This is another theme of Neuromarketing for Dummies. If you can’t tie your measure of attention or emotion in the brain to some decision or action, what is it good for? Some very interesting academic work is starting to appear that takes this position seriously, such as Berkman and Falk, “Beyond Brain Mapping: Using Neural Measures to Predict Real-World Outcomes.” (http://bit.ly/18JoN2z).
Commercial NM vendors need to follow this lead. Neuromarketing buyers may be able to tolerate black box methodologies if they produce useful predictions, just like we don’t begrudge Google its secret Page Rank algorithm because we see the method works — it gets us good search results. But ultimately, neuromarketers aren’t Google, and I and many others believe that greater transparency, not greater secrecy, is what’s going to move our field forward.
As long as both methods and results are shrouded in mystery, articles like these are going to continue to appear. “Trust me, I’m a scientist and you’re not” is not a good marketing strategy.
I’m looking forward to Neuromarketing for Dummies, Steve!
Thanks for this Roger! First I’d like to point out that Emotiv is not a dry headset, it uses saline-soaked pads connected to gold-plated electrodes to conduct current from the scalp. A common misconception that everyone seems to be making.
Further, just because the Emotiv system isn’t as great in some tech specs as the lab-grade devices doesn’t mean the results are any less valid. To offer a comparison, you don’t need 20/20 vision to see what color a wall is! The Emotiv headset reliably detects frequency bands (http://ieeexplore.ieee.org/xpl/articleDetails.jsp?reload=true&arnumber=6273184) and ERPs (http://neurofeedback.visaduma.info/emotivresearch.htm) used in neuromarketing research. We shouldn’t treat Emotiv as a useless “toy” when neuroscience labs around the world are recognizing it for what it is: an inexpensive EEG system with utility in a wide variety of applications. (Also, Ramsoy didn’t critique dry electrodes so much as the built-in detection algorithms used by Emotiv, which everyone knows are just silly 🙂
Thanks for the additional info, Jake. I think there are two separate questions:
1) What has been demonstrated as predictive in peer-reviewed academic research?
2) What has been found to be predictive and reliable by neuromarketing providers, but not published?
There are always going to be those who insist that only the first “counts.” Personally, I hope to see solid academic work published but agree that just because data isn’t public doesn’t make it wrong or fraudulent. Of course, the absence of peer-reviewed research and industry-standard methods puts the onus squarely on the buyer of neuromarketing services to evaluate the credibility and proof-of-performance for their vendors.
Agreed, and I think neuromarketing providers should stick to only what’s been demonstrated as valid in the peer-reviewed research. If we’re talking about being predictive of emotions, peer-reviewed research shows that EEG measures were used to predict emotional states of subjects, with a 95% accuracy at classifying 6 emotional states using a multiclass support vector machine (http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4959627&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D4959627). If we’re talking market-related performance, I would love to see more peer-reviewed research in the space, but Ohme’s papers and this one by Ravaja (http://psycnet.apa.org/journals/npe/6/1/1/) are promising. I find that it’s much easier for people to write off the whole field of neuromarketing (or even the subfield of EEG neuromarketing) than to take the time to evaluate the research and realize there’s something there. Keep fighting the good fight Roger!
Good article Roger! You never let me down.
There is no peer-reviwed evidence for any of these statements – why bother?
The lack of peer-reviewed evidence doesn’t mean that there is no value in neuromarketing research, BMM. It does mean that buyers of those services need to be cautious and take a hard look at the data vendors can show them.
Remember, we’re talking market research – focus groups are not particularly reliable, nor are surveys, but marketers use them all the time. Sometimes the insights may help. If I were a brand or product manager evaluating neuromarketing services, I’d try to structure an A/B or other tests of the recommended actions to ensure the testing was predictive in my environment.
Focus groups are meant to provide “color” and fill in the “whys” behind quantitatives “whats”. The problem I see is the perception that neuromarketing provides concrete 100% reliable answers to what people are thinking. From what I’ve been reading lately, that notion is very much up in the air…
Jeff, I think that neuromarketing can provide additional insights, but I hope no vendor is promising 100% reliability in determining what people are thinking! The CIA would be interested if it could. Actually, I’m sure they are using a few neuromarketing techniques, but not with perfect results.
One other note: results can be quantitative but not accurate. Surveys, etc. are fine for certain kinds of questions, but are unreliable on others.