The Human Delta

An acquisition is a transplant

Transplant surgery had to learn something the hard way. An operation fails for two entirely different reasons, and confusing them is fatal. Either the organ itself was damaged — or the organ was perfectly healthy and the receiving body rejected it. Two different failures, two different tests, two different remedies.

Acquisitions fail in exactly these two ways. Either the team is genuinely not capable of what the plan requires, or the team is capable and the receiving organisation rejects it: different rhythm, different rules, a different idea of what good work looks like. The second case is the more expensive one, because the diligence report said the target was fine — and it was fine, in isolation.

Surgery solved this by measuring compatibility before the operation, not after. Two measurements, taken separately, then read together. The method below does the same thing, and it is organised around those two questions.

Team score minus friction is the deal score. The final step turns it into a coefficient, and into money.

• Is the organ healthy? Can these specific people do what their specific roles require, in a company at this specific stage of its life? Steps 1–4 answer this and produce the team score.
• Will the body accept it? How far apart are the two organisations, how deep and how fast is the integration, under whose governance? Step 5 answers this and produces the friction.

A healthy organ can still be rejected.

Step 1. Team capability, part one: measuring the person

Assessment starts by measuring eight capabilities. What matters most is how they are measured: not by asking people to describe themselves, but by recording what they do.

A conventional questionnaire asks a person to report on their own adaptability, resilience or attention to detail. What comes back is the self-description of a confident adult who has answered that question many times before. It is not dishonesty; self-report is simply weak evidence, and it is easy to shape deliberately.

Instead, the person plays a strategy simulation for 30–60 minutes under steadily rising load, with limited resources and little time. The rules are explained in advance — this is not a test of whether someone can guess the rules, but of how well they apply them as load increases. While they play, the system records thousands of small traces — the decisions themselves, what each one cost, and how they change as the load rises. Those traces become the eight numbers.

• What it is — eight behavioural and cognitive capabilities, derived from decisions made inside a simulation: Mental Efficiency, Learning Agility, Progress Monitoring, Openness to New, Risk Appetite, Result Focus, Agreeableness, Conscientiousness.
• The scale — 0–100 percentiles against a comparison sample of 14,850 people in real jobs from 500 companies, 21 industries, 21 functions and 8 grades — from junior specialists to heads of function. A score of 70 means "higher than 70% of that base", not "70% correct".
• How to work with it — never read a single number as good or bad on its own. A 90 for risk appetite is superb in a founder and a poor match for a chief accountant. Numbers acquire meaning only against a role requirement, which is Step 2. Very high scores are not automatically an advantage either: at that level a capability can work against its owner: a person may chafe at the pace of colleagues, or lose engagement quickly in a role that never calls on that strength.
• What it is for — it is the raw material of everything downstream. Nothing else in the method is measured; this is the only place where real observation enters.

The scores are far harder to flatter: the player never learns which decision reveals what.

Step 2. Team capability, part two: defining the requirement

A measured person means nothing without a stated requirement. And the requirement cannot be one universal ideal, for two reasons.

First, roles differ in kind, not only in quality. The profile that makes an outstanding research scientist would make a poor warehouse manager, and the reverse is equally true. Scoring both against one ideal measures the wrong thing twice.

Second — and this is the part usually missed — the same role requires different things depending on how old the organisation is. A finance director in a two-year-old company improvises without a rulebook; the same title in a forty-year-old institution defends the rulebook against improvisation. Same job title, opposite job. This is why the requirement is defined per role and per lifecycle stage.

• What it is — a desired range on each of the eight capabilities, stated for one role at one stage of corporate life. For example: a software architect in a fast-growth company, Mental Efficiency 75–95.
• The scale — the same 0–100 percentiles as the measurement, expressed as a band rather than a point, because roles tolerate variation.
• Where it comes from — the NeuroFrame Profile Library: 287 professions across 34 occupational areas, taken from the Russian Ministry of Labour occupational register and professional standards, each profiled across the 8 stages of Ichak Adizes' corporate lifecycle and broken down by nine facets of role design — cognitive complexity, cost of error, autonomy, level of management responsibility and others. 2,296 benchmark profiles in total.
• How to work with it — three rules, deliberately asymmetric. Inside the band, the requirement is met and that earns credit. Below the band is costly, though the penalty on any single capability is capped so that one weak scale cannot sink an entire person. Above the band is cheap but small, and a large excess raises a flag: restraining a strong capability costs a person little energy, while sustaining a capability one does not have leads to error and a sustained loss of performance. Excess capability is a retention risk, not a bonus.
• What it is for — it converts eight raw numbers into one meaningful statement: this person, in this seat, in this company, at this moment.

Same job title, opposite job.

Step 3. How much each seat moves the result

A team is not an average of equals. If the chief executive is a poor fit for the plan, that fact should move the result far more than the same shortfall in an analyst. Every deal lawyer already accepts this: it is why a key-man clause names two people and not two hundred. The method simply states it as a number.

• What it is — a weight per person: how much this individual moves the final result.
• The scale — a share of the total, so all weights across the assessed group sum to 1.00. Built from three observable factors: structural authority (decision scope, budget, reporting level), revenue dependency (what share of revenue is attributable to this person), and replaceability (estimated time to find an equivalent).
• How to work with it — defaults come from role seniority and can be overridden where the deal team knows better. A founder holding all the key customer relationships carries far more weight than the org chart suggests. Every override is recorded in the audit file, because it is a judgement, and judgements should be visible.
• What it is for — it makes the team score reflect the concentration of risk, not just its average level. Two teams with identical average capability can carry very different risk depending on where the weakness sits.

It is why a key-man clause names two people and not two hundred.

Step 4. Team capability, result: the team score

Steps 1 to 3 combine into a single number: each person's fit against their own role requirement, blended by their weight.

• What it is — one number describing how well the target's people match what their roles require at this company's stage of life.
• The scale — 50 is the market-average team. A team fully inside the required bands on every capability lands around 60. Serious, broad shortfalls put a team in the low 40s or below.
• How to work with it — read it as the answer to the capability question only. It says nothing yet about your deal: a team scoring 60 can still be a bad purchase, and that is what Step 5 exists to establish.
• What it is for — it isolates capability from context, so the two can be examined separately and addressed separately. If the team score is the problem, the measures are hiring, restructuring roles or walking away. If friction is the problem, the measures are in the deal design — and those are usually far cheaper.

A team scoring 60 can still be a bad purchase.

Step 5. Will the body accept it: friction

Capability is delivered inside conditions. The same team produces different results depending on how different the acquirer is, how much has to change, how quickly, and whether anything holds the key people in place. Six conditions are measured, each stated in points, each bounded. There is nothing else in the model — this is the complete list.

A seventh input, the value thesis, subtracts nothing at all. It only re-weights which capabilities matter: a deal that pays back through new products and markets leans on curiosity, learning speed and calibrated risk, while a deal that pays back through discipline leans on conscientiousness and progress monitoring. A cross-border transaction adds a further overlay for language, regulatory and reporting differences.

Total friction is capped. Reaching that cap is itself a finding: past that point the honest conclusion is not "a risky team" but a badly designed transaction, and the correct response is to restructure it rather than discount it. Two of the conditions also re-weight the capabilities rather than only subtracting — deep integration raises the weight of Openness to New, Agreeableness and Learning Agility, and a change of governance raises the weight of Learning Agility and Progress Monitoring.

• Culture proximity — how far apart the buyer's and the target's operating logic are. The largest single item, because culture distance is one of the most thoroughly documented causes of lost deal value (Mercer, 2018).
• Integration depth — from an autonomous target that keeps running itself, to full absorption where everything changes at once.
• Governance and pace change — a new board cadence, monthly reporting packs and a faster decision clock. Absorbing this is a capability, not a formality.
• Growth engine after close — whether value creation rests on the people already there, or on people yet to be hired. If it rests on the existing team, every gap hits the plan directly.
• Integration timeline — from a runway compressed to the point where nothing can be sequenced, to one long enough to adapt and correct.
• Key-person retention lock — from nothing at all, to multi-year earnouts, retention pools and enforceable non-competes on critical roles. With nothing in place, the general acquisition base rate applies: roughly a third of acquired employees leave within the first year, against 12 per cent of comparable regular hires (Kim, MIT Sloan, 2019).

Past the friction cap, the honest conclusion is not "a risky team" but a badly designed transaction.

Step 6. The result: the line, the coefficient, the money

Team score minus friction gives the deal score. Its distance from 50 is the human delta.

The line is 50. Above it the people add value; below it they cost money. One point is worth roughly up to one per cent of the price.

• What it is — two expressions of the same result. The delta in points, for discussion. The NDC — NeuroDelta Coefficient — for the valuation model.
• The scale — NDC is centred on 1.0 = neutral, deliberately matching the convention of beta, which every investment professional already reads without translation. Above 1.0 is priced risk; below 1.0 a documented premium.
• How to work with it — NDC enters valuation through the door already built for deal-specific risk: the company-specific risk premium in a build-up valuation, or a direct WACC adjustment. (NDC − 1.0) multiplied by the equity risk premium gives the human capital risk premium in basis points. The point-based rule of thumb and the formal route should land close together; if they diverge sharply, neither should be trusted and the inputs need checking.
• What it is for — it puts human capital in the same sentence as every other priced risk, which is the only way it will ever be argued about seriously in an investment committee.

The line is 50. Above it the people add value; below it they cost money.

Now with your own data

Everything above runs on a sample team. The workbench lets you run yours.

Drop in a NeuroFrame results file, or type the people in by hand. What comes back is not a rating. It is a list: which person, which capability, how far below the requirement — and what closing each gap is worth, in points and in money.

Everything stays in your browser; nothing is uploaded anywhere. Headers are recognised in English or Russian: a participant-code column — or, if you work inside your own perimeter, a name column, since nothing leaves the browser — plus the eight scales, with an optional role column.

What comes back is not a rating. It is a list of named, priced gaps.

What this is built on

Almost nothing in the method is invented. Each layer borrows an established body of work; the contribution is the arithmetic that connects them.

• Measuring the person — the five-factor model of personality; reinforcement sensitivity theory (Gray), the balance of approach and inhibition systems that underpins how risk appetite is derived; Eysenck's PEN temperament model; the Cybernetic Big Five; and Evidence-Centred Design, the engineering discipline that turns an event inside a simulation into admissible evidence about a construct.
• Defining the requirement — occupational standards and classifications, giving 287 professions across 34 official areas and with them the role architecture; and Ichak Adizes' corporate lifecycle, eight stages each with its own culture, goals and therefore its own demands on people.
• Measuring the context — Cameron & Quinn's Competing Values Framework and the OCAI: four culture types validated across thousands of organisations, used here to compute the distance between an acquirer and a target rather than to label either of them.
• Pricing the result — Damodaran on the implied equity risk premium and the company-specific risk premium; the build-up method and the CAPM lineage; and the long-standing practice of adding documented, deal-specific risk to a discount rate, the same channel through which country risk, size and illiquidity already flow.

The contribution is not a new theory. It is the arithmetic that connects four established ones.

Calibrating the magnitudes

The point values attached to friction are anchored to published loss data, so that the model's output lands inside the ranges those studies document. No retrospective test against completed transactions has been run yet — see the section on the three kinds of confidence below. The anchors are the four findings in the header of this article: 70–90 per cent of acquisitions failing to create value (Christensen et al., 2011); 43 per cent of deals delayed, terminated or repriced by cultural misalignment, with 67 per cent seeing delayed synergy realisation and 30 per cent never reaching financial targets for culture-related reasons alone (Mercer, 2018, across 1,438 respondents in 54 countries and more than 4,000 deals); 33 per cent of acquired employees leaving within the first year against 12 per cent of comparable regular hires (Kim, MIT Sloan, 2019); and 83 per cent of mergers failing to enhance shareholder value in the early KPMG benchmark of 1999.

This is what "calibrated" means in practice. A serious culture clash is priced at a specific number of points because, through the one-point-equals-roughly-one-per-cent bridge, that puts a serious cultural problem at 8–12 per cent of enterprise value — the middle of the documented range. The figure is then rounded, so a human can hold it in their head and re-derive it.

The point values used for friction are not taken from any single study. They are engineering judgements, calibrated so that the model's output lands inside documented loss ranges, then rounded for memorability. They carry a version number and are re-fitted as validation data accumulates. Anyone who wishes to disagree with a specific value can see what it was anchored to and re-derive the order of magnitude from the same published data; the values themselves are set out in the methodology paper.

Anyone who wishes to disagree with a specific value can see what it was anchored to and re-derive the order of magnitude from the same published data.

Three different kinds of confidence

Not everything here deserves the same level of trust, and pretending otherwise would be the fastest way to lose a serious reader. Here is exactly what rests on what.

• Measured and disclosed — the measurement itself. The assessment has been validated as an instrument and the numbers are stated openly. Comparison sample: 14,850 people in real jobs from 500 companies, 21 industries and 21 functions. Construct validity CFI ≈ 0.96, confirming that the two-domain structure fits the real data. Internal consistency α = 0.69–0.77. Test–retest reliability above 0.83, so results hold on repeat and the method captures stable traits rather than the mood of a Tuesday. Predictive value checked against real KPIs and manager ratings on 3,000+ employees: AUC ≈ 0.77 — the instrument separates stronger from weaker performers well above chance. These are first-party figures, obtained on NeuroFrame's own sample against its own criteria, and they have not been peer-reviewed.
• Established — the component frameworks. The five-factor model, the Adizes lifecycle, the Competing Values Framework, occupational role architecture and discount-rate construction each carry decades of independent literature and use. None of them depends on us being right.
• Calibrated — the magnitudes. The specific point values are engineering judgements anchored to published loss data, not laws of nature. They are stated openly, versioned, and revised as evidence accumulates. The right question is not "are these numbers true?" but "are they better than the alternative?" — and the alternative in current practice is an unwritten impression formed in a meeting room.
• Not yet proven — the composite coefficient. How the parts combine into a single number has not yet been back-tested against completed transactions. It is in Phase 1 of a four-phase validation programme: publish the full method now; back-test against 50+ completed deals, correlating pre-close coefficients with post-close outcomes; run prospective validation on live deals with twelve-month follow-up; publish the empirical study for peer review. We state this in every document, because every standard the deal world now trusts — credit scoring, beta, equity risk premium tables — earned its authority in exactly this order: disclosure first, calibration through data, adoption through demonstrated predictive power.

The alternative in current practice is an unwritten impression formed in a meeting room, which cannot be examined, compared or corrected at all.

Why behaviour rather than self-description

Behavioural measurement avoids the central weakness of self-report: the respondent describes themselves rather than demonstrates behaviour, and such a description is easy to shape deliberately. That is not an abstract preference. Among traditional selection methods, the published meta-analytic record already shows how much of the differences in work performance each approach accounts for, and the spread is wide.

Schmidt and Hunter's 1998 meta-analysis of 85 years of research put work samples, structured interviews and cognitive tests at the top, with roughly a quarter to a third of performance differences accounted for; personality questionnaires around a tenth; years of experience at a few per cent; and years of education close to nothing. Sackett and colleagues' stricter 2022 re-estimation corrected a systematic over-correction for range restriction and pushed several of those figures down — cognitive tests to around a tenth, personality questionnaires lower still.

The most telling entry on that list is the one with no number at all. Several of the instruments most widely used in corporate practice, MBTI and DiSC among them, publish no recognised predictive validity coefficient for job performance: in the meta-analytic tables that row stays empty.

NeuroFrame's own validity figures are stated in the section above and are deliberately not placed on this scale. They were obtained on a different sample, against different criteria, by a different design; different studies are not strictly comparable one-to-one, and a number lifted out of its own study and dropped into someone else's ranking is not evidence, it is decoration. What can be said is narrower and more useful: the instrument is validated against recognised tests and against real KPIs and manager ratings, and the full technical documentation is available to clients under non-disclosure agreement.

The most telling entry on the list is the one with no number at all.

What the method is not

Four limits, stated plainly, because each of them is a place where a method like this is normally oversold.

• Not a judgement on individuals. It prices team-level integration risk for one transaction. A low fit means the role and the stage ask for something a person does not naturally supply — a design problem, not a verdict on a human being. Use for dismissal decisions falls outside what the method supports and is not permitted.
• Not a replacement for diligence. It does not replace interviews, judgement or the rest of due diligence. It replaces the absence of a number where a number was always needed.
• Not a black box. The deal arithmetic — the structure of the calculation and the principle by which every magnitude is calibrated — is published here; the values themselves and their derivation are set out for clients in the methodology paper. A method that cannot be inspected is not a method; it is a brand. The scoring key of the assessment itself stays closed for a different reason: publishing it would turn an implicit behavioural measure into an explicit, coachable one and would invalidate the reference base for everyone measured afterwards.
• Not a guarantee. Matching a requirement raises probability and lowers risk. It does not promise an outcome, and any provider who tells you otherwise is selling something else. The ranges are a reference point, not a verdict: a gap on one or two parameters is something to explore at interview, not grounds for rejection.

A method that cannot be inspected is not a method; it is a brand.

Author, version and standing

Author: Mariia Andrianova.

The complete methodology paper carries every formula, weight and calibration value, the worked examples, and the validation roadmap in full. A live deal screen runs in 15–18 business days, in parallel with financial and legal diligence.

NeuroFrame · The NeuroDelta Method™ · NDC (NeuroDelta Coefficient). Calibration v2.1. Benchmarks from the NeuroFrame Profile Library, 287 roles across 8 Adizes lifecycle stages. The money translation uses the implied equity risk premium of 4.23 per cent (Damodaran, January 2026).

This is a methodology disclosure for professional readers. It is not investment, legal or valuation advice on any specific transaction. Assessment is conducted on coded, de-identified participants; all calculations in the interactive tool run locally in your browser.

Disclosure first, calibration through data, adoption through demonstrated predictive power.

Sources

1. Clayton M. Christensen, Richard Alton, Curtis Rising, Andrew Waldeck (2011) The Big Idea: The New M&A Playbook. Harvard Business Review, March 2011.hbr.org
2. Mercer (2018) Mitigating Culture Risk to Drive Deal Value (M&A Readiness Research). 1,438 respondents across 54 countries, covering more than 4,000 deals over 36 months. Mercer / Marsh & McLennan.corporate.marsh.com
3. Daniel Kim (MIT Sloan), reported in MIT Sloan Ideas Made to Matter (2019) Your acquired hires are leaving. Here's why.. MIT Sloan Ideas Made to Matter, 8 January 2019.mitsloan.mit.edu
4. KPMG (1999) Unlocking Shareholder Value: The Keys to Success. KPMG (archived copy, NYU Stern).pages.stern.nyu.edu
5. Frank L. Schmidt, John E. Hunter (1998) The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124(2), 262–274.doi:10.1037/0033-2909.124.2.262
6. Paul R. Sackett, Charlene Zhang, Christopher M. Berry, Filip Lievens (2022) Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040–2068.doi:10.1037/apl0000994
7. Aswath Damodaran Implied Equity Risk Premium and current market data. NYU Stern School of Business.pages.stern.nyu.edu
8. Ichak Adizes (1988, 2004) Corporate Lifecycles (1988) and Managing Corporate Lifecycles (2004) — the eight-stage model that determines which requirement profile applies to a role. Adizes Institute.adizes.com
9. Kim S. Cameron, Robert E. Quinn (1999) Diagnosing and Changing Organizational Culture, and the Organizational Culture Assessment Instrument (OCAI) based on the Competing Values Framework. OCAI / Jossey-Bass.ocai-official.com
10. Jeffrey A. Gray, Neil McNaughton (2000) The Neuropsychology of Anxiety (2nd ed.); reinforcement sensitivity theory, reviewed in Corr, P. J. (ed.), The Reinforcement Sensitivity Theory of Personality (Cambridge University Press, 2008). Oxford University Press.search.worldcat.org
11. Hans J. Eysenck (1967) The Biological Basis of Personality — the PEN temperament model. Charles C. Thomas, Springfield.search.worldcat.org
12. Colin G. DeYoung (2015) Cybernetic Big Five Theory. Journal of Research in Personality, 56, 33–58.doi:10.1016/j.jrp.2014.07.004
13. Robert J. Mislevy, Linda S. Steinberg, Russell G. Almond (2003) A Brief Introduction to Evidence-Centered Design — the methodology by which an event inside a simulation becomes evidence about a psychological construct. ETS Research Report Series.doi:10.1002/j.2333-8504.2003.tb01908.x
14. NeuroFrame NeuroFrame internal validation and Profile Library (proprietary first-party evidence, not peer-reviewed). Comparison sample of 14,850 people in real jobs from 500 companies, 21 industries and 21 functions; construct validity CFI ≈ 0.96; internal consistency α = 0.69–0.77; test–retest reliability > 0.83; predictive validation against KPIs and manager ratings on 3,000+ employees (AUC ≈ 0.77). The Profile Library holds requirement ranges for 287 professions across 34 occupational areas, each profiled across the eight Adizes lifecycle stages and nine facets of role design — 2,296 benchmark profiles. Full technical documentation is provided to clients under non-disclosure agreement. NeuroFrame.neuroframe.ae
15. Министерство труда и социальной защиты Российской Федерации Реестр профессиональных стандартов и справочник профессий — источник перечня из 287 профессий и 34 профессиональных областей, по которым построен Справочник профилей. Минтруд России.profstandart.rosmintrud.ru

Questions