If you have spent any time in sales training, executive coaching, or the personal-development industry over the last twenty years, you have been told something like this. Watch the prospect’s eyes. If they look up and to the left, they are remembering a visual image. Up and to the right, they are constructing one — possibly a lie. Listen for the predicates they use. If they say “I see what you mean,” they are visual; if they say “that sounds right,” they are auditory; if they say “I have a gut feeling,” they are kinesthetic. Mirror their dominant representational system. Match their breathing rate. Pace, then lead. Anchor the resourceful state. Reframe the objection. Calibrate.

This whole vocabulary belongs to one specific framework: Neuro-Linguistic Programming, or NLP. It was created in the early 1970s by Richard Bandler and John Grinder, two University of California Santa Cruz figures with no formal credentials in psychology or neuroscience. Their core claim was that they had reverse-engineered the structure underlying expert communication and behavior change — that excellent therapists, persuaders, and performers shared an underlying “code” that, once decoded, could be taught to anyone in a weekend seminar.

NLP went on to become one of the most commercially successful psychological-sounding frameworks of the late twentieth century. Tony Robbins built much of his early career on it. Sales training programs from Insurance to enterprise software adopted its techniques. Coaching certifications — “NLP Master Practitioner,” “NLP Trainer,” “NLP Coach” — became a multi-million-dollar industry that still operates today. By any commercial measure, NLP succeeded.

By every scientific measure it failed.

Most of the entries in this hub describe effects that started credible and shrank as the methodology tightened. NLP is unusual because it never had a credible scientific foundation in the first place — and when academic psychology finally got around to testing the specific claims, the rejection was unusually clean. Sharpley’s 1984 and 1987 reviews dismantled the preferred-representational-system hypothesis. Eye-accessing-cue studies failed to find the predicted relationships. Witkowski’s 2010 review of thirty-five years of NLP research concluded the framework was pseudoscience. The Sturt et al. 2012 systematic review in the British Journal of General Practice found no good evidence for NLP-based interventions on any health outcome. There is essentially no controversy in academic psychology about whether NLP’s specific empirical claims have been supported. They have not.

The interesting tension worth understanding is the gap between that scientific verdict and NLP’s continued dominance in sales coaching, leadership development, and self-help. This piece is for strategists who are evaluating an “NLP-certified” coach, considering an NLP-based sales-training program, or trying to figure out whether the framework their company is paying $15,000 per person for is anything more than an expensive vocabulary.

What NLP Originally Proposed

Bandler and Grinder published The Structure of Magic: Volume I in 1975 — a book they marketed as a formal model of therapeutic communication, drawing surface analogies to transformational grammar (then a hot topic in linguistics) and to the work of three then-prominent therapists they claimed to have “modeled”: Fritz Perls (Gestalt therapy), Virginia Satir (family therapy), and Milton H. Erickson (clinical hypnosis). (Bandler & Grinder, 1975, Science and Behavior Books)

The “modeling” claim is doing a lot of work in NLP’s origin story. Bandler and Grinder asserted that they had reverse-engineered the specific linguistic and behavioral patterns these three therapists used to produce change in clients, and that the patterns could be extracted, codified, and taught. The output was a set of techniques that became the original NLP toolkit: the Meta-Model (a set of linguistic challenges to clarify vague client statements), the Milton Model (deliberately vague hypnotic language drawn from Erickson), representational systems (the claim that people preferentially process information visually, auditorily, or kinesthetically), and predicate matching (the prescription to use sensory words matching the listener’s preferred system).

In the late 1970s and early 1980s, the framework expanded rapidly. Frogs into Princes (1979) and Trance-Formations (1981) added more techniques — anchoring (associating a state with a touch or gesture), reframing (changing the meaning context of a behavior), submodalities (the fine-grained features of internal sensory experience), the swish pattern (a visualization sequence claimed to alter habitual responses), the fast phobia cure (a brief visualization protocol claimed to remove phobic reactions in a single session).

Bandler and Grinder split acrimoniously in the early 1980s, leading to a fragmented landscape of competing NLP “schools” that has persisted ever since. There is no single institutional body, no agreed curriculum, no shared standards for certification. A “Certified NLP Master Practitioner” credential from one school may bear little resemblance to the same-named credential from another. This matters for evaluation: there is no single thing called NLP whose claims can be uniformly tested. There is a loose, evolving cluster of practices that share branding, founding myths, and a handful of core techniques.

The core empirical claims that have been tested are the ones most foundational to the framework: that people have a preferred representational system detectable through their language and eye movements, that matching this system produces rapport and influence, and that the specific NLP techniques (anchoring, swish, fast phobia cure, eye-movement integration) produce reliable behavior change beyond what generic communication or relaxation would produce.

Those are the claims the scientific literature actually engaged with. Each of them failed.

The Specific Claims That Have Been Tested

To evaluate NLP fairly, you have to be specific about which claims are on trial. NLP advocates often retreat to unfalsifiable territory when challenged (“NLP is a methodology, not a theory” or “the research tested the wrong version”), so the careful approach is to enumerate the testable claims and look at the data on each one.

The preferred representational system hypothesis (PRS). People supposedly favor one sensory modality — visual, auditory, kinesthetic — for processing information, and this preference is detectable both in the words they choose (visual: “I see,” “look,” “picture”; auditory: “hear,” “sounds,” “rings a bell”; kinesthetic: “feel,” “grasp,” “touch”) and in their eye-movement patterns when accessing different kinds of memory. Matching the listener’s PRS in your own language is supposed to increase rapport and persuasive effectiveness.

Eye-accessing cues (EACs). Bandler and Grinder claimed a specific lateralized map: right-handed people look up-left when remembering visual images, up-right when constructing them, laterally left when remembering sounds, laterally right when constructing them, down-left when conducting internal dialogue, and down-right when accessing kinesthetic feelings. This map became the basis for both NLP communication advice and, infamously, for popular claims that you can tell if someone is lying by tracking where their eyes go.

Predicate matching for rapport. The clinical and persuasive claim: if you match the predicates the other person uses, they will perceive you as more credible, trustworthy, and persuasive. This is the operational core of NLP-based sales training.

Specific therapeutic techniques. Anchoring, the swish pattern, the fast phobia cure, the Visual-Kinesthetic Dissociation procedure for trauma, and eye-movement-based protocols are claimed to produce specific behavior changes — often dramatically, often in a single session.

These are testable claims. They make predictions that can be checked. The literature checked them. Here is what it found.

What Sharpley 1984 and 1987 and Witkowski 2010 Found

Christopher Sharpley’s 1984 review in the Journal of Counseling Psychology was the first major academic synthesis of the NLP literature. He examined fifteen studies that had tested the preferred representational system hypothesis — most published in counseling and educational psychology journals between 1979 and 1983 — and his conclusion was direct: the data did not support the existence of a PRS as Bandler and Grinder had described it. Studies measuring PRS through predicate usage often produced inconsistent classifications. Studies comparing predicate-matched versus mismatched conditions usually found no significant difference in rapport, perceived empathy, or persuasive effectiveness. The few studies showing positive effects suffered from methodological problems (no blinding, demand characteristics, small samples, lack of pre-registration). (Sharpley, 1984, DOI 10.1037/0022-0167.31.2.238)

Sharpley’s 1987 follow-up was even more pointed. By then, several NLP-aligned researchers had attempted to rebut the 1984 review by arguing that the failed studies had not properly applied NLP principles. Sharpley examined the additional studies these critics cited and added them to the analysis. The expanded review reached the same conclusion: the empirical evidence did not support the PRS hypothesis, and the methodological pattern was familiar — NLP advocates would claim a study had not been done correctly whenever it produced negative results, while accepting positive results without applying the same scrutiny. He titled the follow-up “Research findings on neurolinguistic programming: Nonsupportive data or an untestable theory?” — the question itself being his answer. (Sharpley, 1987, DOI 10.1037/0022-0167.34.1.103)

Eye-accessing cues fared no better. A body of studies through the 1980s tested whether eye movements followed the lateralized pattern Bandler and Grinder predicted, whether the pattern was reliable enough to use diagnostically, and whether it correlated with the type of cognitive processing the person was actually doing. The accumulated evidence was overwhelmingly negative. People’s eye movements when remembering or constructing information do not follow the NLP map in any reliable way. The lie-detection claim — that constructed (right-side) eye movements indicate fabrication — has been specifically tested and found not to work, including in a 2012 PLOS ONE study by Wiseman, Watt, ten Brinke, Porter, Couper, and Rankin that explicitly examined the claim and reported null results. (Wiseman et al., 2012, DOI 10.1371/journal.pone.0040259)

Tomasz Witkowski’s 2010 review in the Polish Psychological Bulletin — explicitly framed as a thirty-five-years-after-the-fact accounting — surveyed the entire NLP empirical literature. He examined 33 studies meeting a minimum methodological bar, classifying each as supportive, neutral, or non-supportive of NLP claims. The result: 18.2% supportive, 27.3% uncertain, 54.5% non-supportive. He concluded that NLP, despite its commercial success, had failed to develop an empirical foundation in three and a half decades and represented “a pseudoscientific decoration” of older psychological ideas. (Witkowski, 2010, Polish Psychological Bulletin)

The Witkowski review is important because it disposes of the most common defense NLP advocates raise against earlier critiques — that the research is outdated, the methods have evolved, the early studies tested an immature version of the framework. By 2010, NLP had had three and a half decades to produce supporting evidence under any version of the framework. It had not. The summary of the literature was the same in 2010 as Sharpley’s had been in 1987, just with more studies stacked on top of the same conclusion.

This is unusual. Most contested findings in psychology produce a literature where you can find studies on both sides, where reasonable researchers disagree about which methodological choices are appropriate, where the central effect shrinks but does not vanish under replication pressure. NLP does not produce that pattern. It produces a literature where the methodologically careful studies are overwhelmingly null and the supportive studies cluster among NLP-aligned researchers using protocols that critics consider underpowered, unblinded, or open to demand characteristics.

The 2012 Healthcare Systematic Review

The most rigorous single piece of evidence on NLP’s clinical effectiveness is the 2012 systematic review by Jacqui Sturt and colleagues, published in the British Journal of General Practice. The team — based at the University of Warwick, the University of Birmingham, and other UK academic institutions — set out to evaluate whether NLP-based interventions improved health-related outcomes across any condition.

Their methodology was standard for a Cochrane-style review. They searched MEDLINE, Embase, CINAHL, PsycINFO, AMED, the Cochrane Library, ASSIA, and other databases through November 2010 for randomized controlled trials, controlled trials, before-and-after studies, and interrupted time series that tested NLP interventions against any comparator on any health outcome. They applied standard quality-assessment instruments to each included study. They synthesized the evidence using the SIGN methodology for grading clinical recommendations.

The final dataset comprised 10 studies meeting inclusion criteria. The conditions targeted ranged across anxiety disorders, post-traumatic stress, substance use, weight management, claustrophobia, and morning sickness. The outcomes measured varied accordingly.

The headline finding: “There is little evidence that NLP interventions improve health-related outcomes. This conclusion reflects the limited quantity and quality of NLP research, rather than robust evidence of no effect.” The authors did not say “NLP definitely does not work” — they said the evidence base supporting any positive effect is thin enough that no defensible clinical recommendation can be made. (Sturt et al., 2012, DOI 10.3399/bjgp12X658287)

This matters for the coaching and corporate-training context too, even though the review focused on healthcare. NLP advocates often defend the framework against academic skepticism by saying the research community has not properly tested it. The 2012 Sturt review represents exactly the kind of testing that would have surfaced real effects if they were there. Thirty-five years of clinical application, then a thorough systematic search, then 10 studies that meet basic methodological standards, then a finding of “no good evidence.” This is what it looks like when a framework’s empirical claims do not hold up under scrutiny.

Devilly’s 2005 paper Power therapies and possible threats to the science of psychology and psychiatry placed NLP within a broader cluster of “power therapies” — Thought Field Therapy, Emotional Freedom Techniques, Eye Movement Desensitization and Reprocessing in its more extravagant claimed-mechanism form, and several others — that share a marketing pattern of dramatic claims, weak evidence, and aggressive certification economies. His analysis treated NLP as a representative case in a broader phenomenon: psychological-sounding frameworks that earn money in the absence of evidence by exploiting clinicians’ and consumers’ difficulty in evaluating the quality of psychotherapy research. (Devilly, 2005, DOI 10.1080/j.1440-1614.2005.01601.x)

Where NLP’s Techniques Actually Come From

If NLP’s empirical claims do not stand up, where does the apparent effectiveness reported by practitioners come from? The honest answer is a mix of three things, none of which require any of NLP’s distinctive theoretical apparatus.

Recycled hypnosis. Milton Erickson was a clinical hypnotherapist of genuine skill whose linguistic patterns Bandler and Grinder transcribed and renamed. The “Milton Model” within NLP is essentially a vocabulary for the deliberately ambiguous language Erickson used to facilitate hypnotic induction. To the extent NLP practitioners produce effects similar to those of skilled hypnotherapists, this is because the underlying technique — guided suggestion in a relaxed and attentive state — has a real (if narrower than often claimed) evidence base, and the NLP repackaging changes the vocabulary without much changing the mechanism. Anchoring is essentially Pavlovian conditioning relabeled. The swish pattern is a guided-imagery exercise of a kind cognitive-behavioral therapists were already using.

Standard rapport behaviors. Mirroring postures, matching speech rate roughly, maintaining eye contact, paraphrasing back what someone has said — these are basic interpersonal skills that pre-date NLP by decades and have a substantial evidence base in counseling psychology and social psychology research outside any NLP framing. The non-verbal-synchrony literature shows that some forms of behavioral matching do facilitate liking and cooperation, with moderate effect sizes that vary by context. A good NLP practitioner who mirrors a client’s posture is doing something that works. The work is being done by the matching, not by any NLP-specific theory of representational systems or accessing cues. (Chartrand & Bargh, 1999, DOI 10.1037/0022-3514.76.6.893)

Placebo, expectancy, and therapeutic alliance. A coach or therapist who projects confidence, gives the client a clear interpretive framework for their problem, prescribes a specific procedure to address it, and creates ritualized expectation of change tends to produce some change — across an enormous range of frameworks, from psychoanalysis to crystals to NLP. The literature on common factors in psychotherapy is quite clear that the relationship, the therapist’s expertise as perceived by the client, and the credibility of the framework account for far more outcome variance than the specific techniques. This means a charismatic NLP coach probably does help some clients in some ways; it just does not mean NLP-the-framework is doing the helping.

The combined effect is that an NLP-trained coach with good interpersonal skills, doing essentially what any decent communicator would do, can produce real and visible client improvements that get attributed to “the techniques.” This is the engine that has kept the certification industry running for fifty years. The improvements are real. The theoretical attribution is wrong. And the framework remains untestable in any way the proponents will accept, because every positive case is credited to NLP and every negative case is attributed to the practitioner having applied NLP incorrectly.

Why NLP Persists In Coaching And Sales Training

NLP’s continued dominance in sales coaching, executive coaching, and leadership development is one of the more interesting case studies in how scientifically untestable frameworks survive in business contexts where academic scrutiny does not reach. A few reinforcing dynamics:

Folk credibility. NLP’s vocabulary sounds scientific. The word “neuro” is in the name. The “linguistic” piece references real linguistics. The “programming” metaphor borrows from computer science. To a marketing manager or sales leader without psychology training, “Neuro-Linguistic Programming” sounds like it must have been validated somewhere. The credibility comes from the naming, not from any underlying research.

Lack of falsifiability in the use case. A salesperson trained in NLP who closes a deal credits the techniques. The same salesperson who loses a deal blames the customer’s resistance, not the framework. The same person who closes more deals after the training credits the training, but the counterfactual — what would their close rate have been without it — is essentially never measured. In a business context, there is no replicated controlled experiment. Every individual practitioner observation supports continued use of the framework because the framework cannot be disconfirmed at the individual level.

The certification economy. NLP certification programs generate revenue. Newly certified practitioners need to recoup their investment, which they do by selling NLP-based coaching to clients and by becoming trainers themselves, certifying the next cohort. This is a self-sustaining economic structure that does not require external validation, and it actively resists scientific scrutiny because such scrutiny threatens revenue. The structure mirrors that of multi-level-marketing organizations — value flows toward those who certify others — and is one of the warning signs Lilienfeld and colleagues flagged in their broader work on questionable practices in psychology and education.

Survivorship bias in testimonials. Tony Robbins, who built his early career partly on NLP, is enormously successful. So are some other prominent coaches with NLP backgrounds. The testimonials of successful people are visible; the much larger population of NLP practitioners who did not become Tony Robbins is invisible. The framework gets credit for the visible successes and pays no cost for the invisible failures. This is the same selection mechanism that supports all weak frameworks claiming a celebrity endorsement.

Genuine helpfulness of unrelated practices. As discussed above, the standard rapport behaviors, hypnotic-suggestion techniques, and credibility-and-expectation effects bundled inside NLP do help some clients. The framework gets the credit for benefits the wider tradition of clinical hypnosis, counseling psychology, and standard communication practice would have produced anyway.

None of this means NLP coaches are dishonest people. Most NLP practitioners genuinely believe they have learned something powerful. The point is structural: a system that explains all outcomes (success = NLP works; failure = practitioner did NLP wrong) and that operates outside the institutions where empirical testing happens can persist indefinitely without ever earning scientific support. NLP is one of the cleanest examples of this dynamic in the broader landscape of self-help.

What’s Honest To Say About Communication And Influence Now

Rejecting NLP does not require rejecting interest in communication, persuasion, or influence — there is real evidence-based work in social psychology and communication science that addresses the same questions NLP claims to address. The honest summary of what is and is not known:

Robert Cialdini’s six (now seven) principles of influence — reciprocity, commitment-and-consistency, social proof, liking, authority, scarcity, unity — have a mixed but actually existing evidence base across decades of social psychology research. Some principles replicate strongly (reciprocity in field studies, scarcity in field studies); others have weaker or more conditional support (commitment-and-consistency depends heavily on context; social proof is complicated by reactance effects). Cialdini is not perfect, but he is on a different empirical planet from Bandler and Grinder. (See the Cialdini Influence Principles entry in this hub for the full evaluation.)

Non-verbal mimicry and behavioral synchrony do have some research support, but the effect sizes are modest, the contexts in which mimicry helps versus hurts are not fully understood, and the mechanisms are debated. Chartrand and Bargh’s 1999 “chameleon effect” paper established that low-level mimicry can increase liking under specific conditions, and follow-up work has both replicated and qualified the original findings. The “mirror everything” advice in NLP-derived sales training is an over-extrapolation from a real but more limited phenomenon. (Chartrand & Bargh, 1999, DOI 10.1037/0022-3514.76.6.893)

Cognitive dissonance, foot-in-the-door, and door-in-the-face are commitment-and-compliance paradigms with much stronger empirical foundations than anything in NLP. They are also much more disciplined about specifying their boundary conditions. (See the Cognitive Dissonance, Foot-In-The-Door, and Door-In-The-Face entries for the boundary conditions.)

Active-listening techniques from counseling psychology — paraphrasing, reflective listening, open-ended questioning, summarizing — were systematically taught in Carl Rogers’s client-centered tradition decades before NLP and have a more substantial outcome literature in clinical research. A salesperson or coach learning these techniques without the NLP packaging gets the actual evidence-based skills without the empirical baggage.

Motivational interviewing, developed by William Miller and Stephen Rollnick, is a structured client-centered counseling approach for behavior change that has accumulated meaningful evidence in substance-use, health-behavior, and adherence contexts. It draws on the same broader tradition of communication-and-influence work that NLP borrows from, but is institutionally embedded in clinical research and has accountability for its empirical claims that NLP never developed.

Specific persuasion research in social psychology — Petty and Cacioppo’s elaboration likelihood model, Hovland’s earlier source-credibility work, the framing-effects literature in behavioral economics, the negotiation literature from Harvard’s Program on Negotiation — provides a much richer and more honest scientific basis for thinking about influence than NLP’s repackaged hypnotic vocabulary.

The honest position for a strategist who genuinely cares about evidence-based influence: there is real research, much of it less commercially packaged than NLP, that actually addresses the questions NLP claims to address. The reason to skip NLP is not that influence and communication are unimportant or unstudyable. They are important and they have been studied. The reason to skip NLP is that it is one of the parts of the literature that has been studied and rejected.

What This Means For Hiring NLP-Trained Coaches Or Adopting NLP Sales-Training Programs

The practical decision-relevant summary:

If you are evaluating an NLP-certified coach. The certification itself carries no information about the coach’s effectiveness, because NLP certifications are granted by competing private training organizations with no consistent standards and no requirement to demonstrate evidence-based competence. Effective coaches with NLP credentials are effective because of general skills (active listening, genuine interest in the client, business judgment, communication craft) that do not depend on NLP. Ineffective coaches with NLP credentials are ineffective for ordinary reasons. The credential adds no signal. Evaluate the coach on direct evidence of effectiveness with clients similar to you, not on the framework label.

If your company is considering NLP-based sales training. The framework’s empirical foundation does not support the costs of these programs, which often run $5,000 to $15,000 per participant for multi-day immersions or longer for “master practitioner” tracks. To the extent the training produces real improvement, it does so through standard rapport-and-influence content that less expensive, more evidence-based programs cover. Sales-training programs grounded in research on goal-setting, deliberate practice, structured feedback, and (where applicable) the behavioral-economics literature on choice architecture are likely to produce better return on investment.

If you are personally considering an NLP self-help product. The same caution applies and possibly more strongly. NLP self-help materials sell hard on transformation promises that are not supported by evidence. The actual content tends to be standard self-help advice (visualize the outcome, manage your state, build rapport with people) layered with NLP-specific jargon that adds no value. If you find the framework appealing for its structure or motivational qualities, you can probably get the same benefit from any number of evidence-grounded resources at a fraction of the cost.

If someone uses NLP terminology in a sales presentation to you. This is informative. It tells you the person has been trained in a framework that is widely recognized as pseudoscientific in academic psychology, and that they are willing to use vocabulary from it in a credibility-staking context. This does not mean they are a bad-faith actor — most NLP-trained sales professionals have been told NLP is well-established science and have no reason to doubt it. But it should slightly lower your prior that the sales pitch is calibrated to evidence rather than to influence techniques.

If you are setting evaluation standards for any “behavioral science” or “psychology” training your organization buys. A useful screening question for any vendor: “What systematic reviews or meta-analyses establish the effectiveness of the specific techniques you teach?” Vendors selling evidence-based methods will have answers. Vendors selling NLP will pivot to testimonials, the framework’s history, or claims that it has been “used by” prominent people. The pivot pattern is itself diagnostic.

The single most important takeaway: the certification industry around NLP exists in spite of the evidence, not because of it. A strategist who is asked to spend organizational money on NLP-based programs has a defensible basis for declining and recommending an alternative — and the academic literature, not just personal skepticism, backs the decision.

Sources

  • Bandler, R., & Grinder, J. (1975). The Structure of Magic I: A Book About Language and Therapy. Science and Behavior Books.
  • Sharpley, C. F. (1984). Predicate matching in NLP: A review of research on the preferred representational system. Journal of Counseling Psychology, 31(2), 238–248. DOI: 10.1037/0022-0167.31.2.238
  • Sharpley, C. F. (1987). Research findings on neurolinguistic programming: Nonsupportive data or an untestable theory? Journal of Counseling Psychology, 34(1), 103–107. DOI: 10.1037/0022-0167.34.1.103
  • Witkowski, T. (2010). Thirty-five years of research on neuro-linguistic programming. NLP research data base. State of the art or pseudoscientific decoration? Polish Psychological Bulletin, 41(2), 58–66. Sciendo
  • Sturt, J., Ali, S., Robertson, W., Metcalfe, D., Grove, A., Bourne, C., & Bridle, C. (2012). Neurolinguistic programming: A systematic review of the effects on health outcomes. British Journal of General Practice, 62(604), e757–e764. DOI: 10.3399/bjgp12X658287
  • Devilly, G. J. (2005). Power therapies and possible threats to the science of psychology and psychiatry. Australian and New Zealand Journal of Psychiatry, 39(6), 437–445. DOI: 10.1080/j.1440-1614.2005.01601.x
  • Wiseman, R., Watt, C., ten Brinke, L., Porter, S., Couper, S.-L., & Rankin, C. (2012). The eyes don’t have it: Lie detection and neuro-linguistic programming. PLOS ONE, 7(7), e40259. DOI: 10.1371/journal.pone.0040259
  • Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: The perception-behavior link and social interaction. Journal of Personality and Social Psychology, 76(6), 893–910. DOI: 10.1037/0022-3514.76.6.893
  • Bandler, R., & Grinder, J. (1979). Frogs into Princes: Neuro Linguistic Programming. Real People Press.
  • Grinder, J., & Bandler, R. (1981). Trance-Formations: Neuro-Linguistic Programming and the Structure of Hypnosis. Real People Press.

FAQ

But my NLP coach really helped me — are you saying that did not happen? The improvement is real. The attribution is what is in question. A good NLP coach typically combines genuine listening, structured questioning, expectancy-building, and (sometimes) hypnotic-suggestion techniques in a way that produces meaningful client benefit. The benefit you experienced is consistent with what well-trained coaches in many traditions produce — including coaches with no NLP training at all. What the evidence does not support is the claim that the specific NLP techniques (representational systems, eye-accessing cues, anchoring, swish pattern) are doing the work over and above what any competent coaching relationship would do. Your coach probably helped you; the framework probably did not, in any specific way that distinguishes it from other coaching traditions.

What about Tony Robbins — did he not build his career on NLP? Robbins’s early training and early business included NLP, and some NLP techniques remain in his curriculum. But Robbins also draws on a much wider set of practices, including elements from cognitive-behavioral therapy, motivational psychology, peak-performance literature, and his own significant personal stagecraft. Attributing his success to NLP specifically is the same kind of survivorship-and-attribution error covered above. Many practitioners trained in NLP did not become Tony Robbins. The framework should not be credited with one practitioner’s exceptional outcomes any more than a particular guitar brand should be credited with Jimi Hendrix.

Is mirroring posture or speech rate useful at all? There is moderate evidence that low-level non-verbal mimicry can increase liking and cooperation in some contexts — Chartrand and Bargh’s chameleon-effect paper and follow-ups establish this much. The effect sizes are not large, the boundary conditions are not fully mapped, and there are documented cases where mirroring backfires (when the other person perceives it as manipulative or insincere). The honest summary: subtle, natural-feeling matching of someone’s general energy and pace probably helps slightly in many conversations; deliberate, conspicuous mirroring of specific postures and gestures probably ranges from neutral to actively counterproductive. The NLP-style advice to consciously match the dominant “representational system” through predicate selection has no evidence base of its own.

What evidence-based sales training exists if not NLP? A few useful starting points. Neil Rackham’s SPIN Selling has an actual research foundation — it grew out of a multi-year study of thousands of sales calls and proposed specific question types that correlated with closure rates. Solution-selling and consultative-selling frameworks vary widely in evidence quality but generally derive from observable practices in successful enterprise sales. The behavioral-economics literature on choice architecture (Thaler and Sunstein’s Nudge and the academic work behind it) provides defensible principles for structuring offers and decisions. Motivational interviewing (Miller and Rollnick) has substantial evidence in change-conversation contexts. None of these are perfect, all have weaker evidence than their proponents claim, but each has stronger empirical foundations than NLP.

Did NLP work for therapists at the time it was popular in the 1980s? This is the most charitable framing of NLP’s origin story — that Bandler and Grinder were practicing therapists who codified techniques that produced real clinical outcomes for them. The honest assessment is that they did adopt and rebrand effective elements of Ericksonian hypnosis, Gestalt therapy, and Satir family therapy, and that the underlying techniques in those traditions do have some clinical track record. But the specific theoretical claims they layered on top — the representational systems, the eye-accessing cues, the “structure of magic” framing — were never validated and never differentiated NLP from the source traditions it borrowed from. NLP’s contribution was packaging and marketing, not new clinical content.

Is it ever defensible for a business to invest in NLP-based training? The defensible-investment threshold for any training program is whether it produces enough improvement in specific business outcomes to justify the cost. For NLP-based programs the evidence supporting that improvement is weak enough that the defensible answer is almost always no — there will be a cheaper, more evidence-based alternative for whatever you are trying to teach (rapport, listening, persuasion, change management). Edge cases exist — for example, a team that has already built a working sales-coaching culture around NLP vocabulary may find that switching imposes costs that exceed the benefit. But for a new investment from a standing start, the answer is to spend the budget on training with a better evidence base.

How do I respond to a colleague or client who insists NLP works? Carefully and without contempt. NLP enthusiasts often hold the belief sincerely and have had experiences that feel like confirmation. Lecturing them on the Witkowski review is unlikely to change their mind and may damage the relationship. A more productive approach is to ask what specific outcomes they care about, then look together at what evidence-based approaches address those outcomes — and let the comparison speak for itself. The goal is not to win an argument about NLP; the goal is to spend the organization’s resources on what works. Sometimes that requires not making the argument explicit at all.

If NLP is pseudoscience, why do so many smart people use it? Because the framework is engineered to be unfalsifiable at the individual practitioner level. Smart people who use NLP have personal experience that feels like confirmation — clients improve, colleagues report being persuaded, sales close. They lack a controlled comparison condition, so they cannot see that those outcomes would have happened with any reasonably skilled communicator regardless of framework. They are also embedded in a community that reinforces the framework and provides no incentive to question it. This pattern is not specific to NLP. It is the same dynamic that supports astrology, homeopathy, and graphology among otherwise intelligent practitioners. Smart people without explicit training in evaluating evidence are not better at evaluating evidence than anyone else — sometimes worse, because their confidence in their own judgment is higher. The defense against this in your own life is to maintain epistemic humility about frameworks you have not personally compared against actually-blinded control conditions, and to weight published systematic reviews above your own n=1 impressions when the two disagree.

replication-crisis nlp-pseudoscience sales-coaching pop-psychology evidence-evaluation
Share this article
LinkedIn (opens in new tab) X / Twitter (opens in new tab)
Atticus Li

Experimentation and growth leader. CXL-certified CRO practitioner, Mindworx-certified behavioral economist (1 of ~1,000 worldwide). 200+ A/B tests across energy, SaaS, fintech, e-commerce, and marketplace verticals.

About LinkedIn Newsletter