What is evidence-based management? Evidence-based management is making organisational decisions through the conscientious, explicit and judicious use of the best available evidence from multiple sources, in order to increase the likelihood of a favourable outcome.
The Center for Evidence-Based Management, the leading authority on evidence-based practice in management, identifies four sources of evidence that a sound decision draws on. Scientific findings are the cumulative body of published research relevant to the question. Organisational data are the internal facts, metrics and contextual information available to the organisation. Practitioner expertise is the experience-based judgment of the professionals involved. Stakeholder concerns are the values, interests and perspectives of the people affected by the decision — employees, customers, and the broader constituency. The core discipline is that all four are appraised critically rather than accepted at face value, and that none is treated as sufficient alone. A decision resting only on internal dashboards is data-driven but not evidence-based, because it omits what is already known from research, what experienced practitioners judge, and what those affected value. The framework is set out in Evidence-Based Management: The Basic Principles by Eric Barends, Denise Rousseau and Rob Briner.
What Is Evidence-Based Management?
Evidence-based management means making decisions through the conscientious, explicit and judicious use of the best available evidence — from four sources, not one. A dashboard is not evidence-based management.
Last updated: 2026-07-29
In short
Most organisations that describe themselves as data-driven are using exactly one of the four sources of evidence. Evidence-based management asks for scientific findings, organisational data, practitioner expertise and stakeholder concerns — each appraised critically, none treated as sufficient on its own.
Data-driven is not the same as evidence-based
The distinction matters more than it sounds. An organisation running entirely on internal metrics is using one source and ignoring three: what published research already established, what experienced practitioners judge, and what the people affected actually value.
- •Internal data answers what happened here — not whether it generalises, or why.
- •Research answers what is generally known — often already settled elsewhere, and cheaper to read than to rediscover.
- •Practitioner judgment carries context that no dataset encodes.
- •Stakeholder concerns are evidence, not noise. The values of those affected are a legitimate input, not an obstacle to manage.
- •Critical appraisal is the discipline. Each source is weighed for quality, not simply cited.
The four sources
The CEBMa model is deliberately symmetrical: four sources, each appraised, none sufficient alone.
Scientific findings
The cumulative body of published research relevant to the question — particularly bodies of evidence rather than single studies. The most commonly skipped source, and often the cheapest: a great many organisational questions have already been studied.
Organisational data
Internal facts, metrics and contextual information. The source most organisations over-weight, because it is the one they own. Strong on what happened locally, silent on whether it generalises.
Practitioner expertise
The experience-based judgment of the professionals involved. Genuine evidence, and simultaneously the source most vulnerable to bias — which is why it is appraised rather than deferred to.
Stakeholder concerns
The values, interests and perspectives of those affected — employees, customers, the wider constituency. Treating these as evidence rather than as objections to be handled is the part organisations most often get backwards.
Where the evidence has to survive contact with the argument
Gathering four sources is the easy half. The hard half is that they routinely disagree — and the decision has to reconcile them explicitly rather than by quietly privileging whichever is most convenient.
Evidence attached to the claim it supports
In an argument tree each piece of evidence sits under the specific claim it backs, so it is visible whether a claim rests on a study, a dashboard, an opinion, or nothing at all.
Conflicting sources held side by side
When the research and the internal data disagree, both stay attached to the claim as competing support rather than one being silently dropped.
Stakeholder input as structured argument
Concerns from affected groups enter as arguments with reasons, not as sentiment to be summarised — which is what lets them be weighed rather than counted.
Unsupported claims are conspicuous
A claim with no evidence beneath it looks exactly like what it is. In prose it reads as confident; in a tree it reads as bare.
This is the sound reasoning element of decision quality doing its work: evidence only improves a decision if the argument connecting it to the conclusion holds.
Explore further
Enterprise decision excellence
The umbrella: how evidence, authority, culture and learning combine.
Data-driven decision making
The narrower discipline — one of the four sources, done well.
Decision culture
Whether inconvenient evidence actually reaches the decider.
What is argument mapping?
Laying out claims and the evidence beneath them visually.
The Toulmin model
Claim, grounds and warrant — the anatomy of evidence supporting a conclusion.
Decision intelligence
Combining decision processes with data, analytics and AI.
Frequently asked questions
What is evidence-based management?
Making organisational decisions through the conscientious, explicit and judicious use of the best available evidence from multiple sources, to increase the likelihood of a favourable outcome. The Center for Evidence-Based Management is the standard reference for the definition.
What are the four sources of evidence?
Scientific findings from published research; organisational data and internal context; practitioner expertise and experience-based judgment; and the concerns, values and interests of stakeholders affected by the decision. All four are critically appraised, and none is sufficient alone.
Is data-driven decision making the same as evidence-based management?
No. Data-driven decision making typically uses one of the four sources — organisational data. It is valuable but incomplete: it cannot tell you what research has already established, what experienced practitioners judge, or what the people affected value.
Why does practitioner expertise count as evidence?
Because experienced professionals carry contextual knowledge that no dataset encodes. It is also the source most vulnerable to bias and overconfidence, which is why the discipline is to appraise it critically alongside the others rather than to defer to seniority.
Why are stakeholder concerns treated as evidence?
Because the values and interests of those affected legitimately bear on whether a decision is a good one, not merely on whether it will be accepted. Treating them as obstacles to manage rather than as evidence to weigh is a common and consequential inversion.
What does critical appraisal mean here?
Weighing each source for quality and relevance rather than simply citing it: how strong is this study, how representative is this internal metric, how reliable is this expert judgment in this specific context, how broadly held is this stakeholder concern. Citation is not appraisal.
References & further reading
Barends, E., Rousseau, D. M., & Briner, R. B. (2014). Evidence-Based Management: The Basic Principles. Center for Evidence-Based Management.
The four-source model and the definition of evidence-based practice in management.
View source →Center for Evidence-Based Management (CEBMa) — What is evidence-based management?
CEBMa's own reference definition and resources.
View source →CIPD. Evidence-based practice for effective decision-making (factsheet).
A practitioner-facing summary of the four sources and how to apply them.
View source →Toulmin, S. E. (1958). The Uses of Argument. Cambridge University Press.
Grounds and warrant — the structure connecting evidence to the claim it is offered for.
View source →Put the evidence under the claim it supports
Four sources of evidence only help if it is visible which claim each one backs — and which claims are resting on nothing.
Start free