Issue 266, May 28, 2026
A client I worked with several years ago published trust survey results showing 84 percent of customers reporting high trust in the institution. The figure was prominently displayed in the annual report, repeated in every stakeholder and governance meeting, and used as a justification for the client’s strategy of competing on relationship quality rather than discounts. Two years later, the client’s customer base contracted sharply when they were hacked for the third time in two years. Analysis conducted via interviews revealed that a substantial portion of the customers who had reported high trust on the former survey had completely lost trust in the organization. Not because of the minimal consequences, but because of how the organization handled and communicated each of the situations.
The trust score had not predicted the behavior. The behavior revealed what the trust score had failed to capture. Customers had answered the survey honestly. They simply didn’t know what they would actually do under pressure, because the survey asked about their current sentiment rather than the conditions under which trust would be tested, as it was with the hacks. The client had been measuring something, but it wasn’t measuring trust.
This pattern, in which trust appears strong by survey and collapses under situational pressure or behavioral observation, is one of the most consistent findings in modern measurement research. We often overlook how strongly emotions, thoughts, and beliefs influence us and our organizations. We often forget how important the human factor is in personal and organizational life.
Trust is among the most consequential intangibles in organizational life. It predicts employee retention, customer loyalty, partner reliability, regulatory cooperation, and the operational resilience that determines whether an organization survives disruption. It is also among the most poorly measured intangibles in standard practice, because most organizations rely on instruments that capture sentiment at a moment in time rather than the behavioral patterns that reveal whether trust actually exists. This article examines what trust is, why it resists conventional measurement, and how organizations can operationalize trust measurement through behavioral observation rather than survey self-report.
Why Trust Resists Direct Measurement
The fundamental problem with measuring trust through surveys is that trust is not a state. It is a pattern, one built over time and informed by actions and behaviors (or the lack of either). The distinction matters operationally. A state is something a person possesses or does not possess at a particular moment, and it can, in principle, be measured by asking. A pattern is something that emerges across many interactions over time, and it can only be reliably measured by observing the interactions themselves. When you ask a person whether they trust an organization, you are asking them to summarize a pattern of interactions in a single moment of reflection. Their summary is influenced by recent events, by the framing of the question, by the social desirability of certain answers, and by the absence of any meaningful test of the trust they are reporting. The summary may be sincere. It is not predictive of future behavior, nor of how trust will hold or fail when tested by events.
Frances Frei and Anne Morriss, in their widely cited 2020 Harvard Business Review article and subsequent book, described trust as a triangle composed of three components: authenticity (the perception that a person or organization is being genuine), logic (the perception that their reasoning and judgment are sound), and empathy (the perception that they care about the interests of the person whose trust is at stake). Frei and Morriss’s framework is useful for this exploration because it disaggregates trust into observable surfaces. Each leg of the triangle reveals itself through behavior, not through declaration in response to a survey question. Authenticity is observable in whether the organization’s stated values match its operational behavior over time. Logic is observable in whether its decisions hold up to subsequent scrutiny. Empathy is observable in whether its actions during periods of stress prioritize the interests of the stakeholders whose trust it depends on. Each of these is measurable, but none of them is measurable through a survey item asking whether trust exists and is present.
Roger Mayer, James Davis, and David Schoorman provided an earlier and equally influential model in their 1995 paper in the Academy of Management Review. Their formulation identified three antecedents of trust: ability (the trusted party has the competence to deliver), benevolence (the trusted party intends to act in the trustor’s interest), and integrity (the trusted party adheres to principles the trustor finds acceptable). The Mayer-Davis-Schoorman model has been validated across dozens of empirical studies, and its operational utility lies in the fact that each antecedent generates observable behavioral evidence. Ability is observable in delivery patterns over time. Benevolence is observable in decisions made under pressure, particularly decisions where the trusted party could have prioritized its own interest at the cost of the trustor’s. Integrity is observable in the consistency between stated principles and actual behavior across contexts. Measuring trust, in this framework, becomes the measurement of these three behavioral patterns over time, not the measurement of subjective belief at a single moment.
The 2025 Edelman Trust Barometer, in its twenty-fifth annual edition, documented that global trust in institutions has remained at or near historic lows for the past several years, with particular weakness in employer trust during periods of organizational change. I have covered various editions of the barometer as evidence in my books (The Truth About Transformation and The Truth About Transformation Leading the Age of AI, Uncertainty and Human Complexity) in the context of the human factor and its role in organizational change and transformation.
The Barometer’s methodological evolution over its quarter-century history is itself instructive. The earliest editions relied almost exclusively on direct trust questions. More recent editions have layered behavioral indicators on top of sentiment measures, asking respondents about specific behaviors they have engaged in or would engage in, including job changes, purchasing decisions, and information sharing. The shift reflects a recognition that what people say about trust is a weaker predictor of behavior than what they actually do, and that the most useful trust measurement is the measurement that anchors itself in observable action.
The Three Surfaces of Trust
Trust in an organizational context shows up on three distinct surfaces, each of which requires its own behavioral measurement approach. Internal trust is the trust that exists among employees and between employees and leadership inside the organization. It determines whether information flows honestly, whether bad news reaches the people who need to hear it, and whether discretionary effort is offered or withheld. Customer trust is the trust that exists between the organization and the people who buy or use what it produces. It determines loyalty under pressure, willingness to forgive errors, openness to expanded relationships, and resilience during competitive challenge. Partner trust is the trust that exists between the organization and the external parties on whom it depends, including suppliers, distributors, regulators, and collaborators. It determines the quality of cooperation during ambiguity, the willingness to share commercially sensitive information, and the responsiveness of the partner ecosystem when the organization needs flexibility outside contractual terms.
Each surface of trust reveals itself through different behavioral signals, and a serious approach to trust measurement requires distinct instrumentation for each. Conflating them, as most organizational dashboards do, produces a single trust score that captures none of the three with sufficient clarity to concretely and intelligently inform decisions. The first discipline of trust measurement is to disaggregate trust by surface, and to design behavioral observation appropriate to each.
Internal Trust
Internal trust is observable in seven behavioral signals that consistently predict the health of trust relationships within an organization. The first is the rate at which problems are raised early versus the rate at which they surface as crises. Organizations with strong internal trust have a steady flow of early-warning communication from frontline employees to mid-level leadership and from mid-level leadership to senior leadership. Organizations with weak internal trust have long quiet periods punctuated by surprise crises that, in retrospect, were knowable weeks or months in advance. Measuring this signal requires tracking the lag time between when a problem was first identifiable at the operational level and when it reached the decision makers responsible for addressing it. The lag is a direct behavioral indicator of trust in upward communication.
The second signal is the pattern of response to bad news. Amy Edmondson’s research on psychological safety, particularly her 1999 paper in Administrative Science Quarterly and her 2018 book The Fearless Organization, demonstrates that the most reliable indicator of internal trust is what happens immediately after someone delivers difficult information. Leaders who respond with curiosity, questions, and exploration produce environments in which bad news flows freely. Leaders who respond with defensiveness, blame, or interrogation produce environments in which bad news is filtered, delayed, or hidden. The behavioral observation here is direct. Senior leaders can be evaluated, by their own teams and by independent observation, on the consistency of their response patterns when difficult information is delivered. The pattern is the trust indicator.
The third signal is information flow across boundaries. Organizations with strong internal trust have visible patterns of cross-functional information sharing, voluntary collaboration on projects that span department lines, and willingness to share data and tools across groups. Organizations with weak internal trust have visible patterns of information hoarding, defensive territoriality around domain expertise, and the systematic avoidance of cross-functional accountability. The behavioral observation is the rate and quality of information transfers that occur outside formal reporting structures. Where trust is strong, those transfers happen freely. Where trust is weak, they require formal escalation, political intervention, or executive mandate.
The fourth signal is what happens during mistakes. Edmondson’s foundational research at hospitals revealed a counterintuitive pattern. The hospital units with the highest reported medication error rates were not the units with the most actual errors. They were the units with the highest psychological safety, the units in which staff trusted that admitting an error would not result in punishment. Units with low psychological safety reported fewer errors not because they made fewer errors but because the staff trusted no one with the information that errors had occurred. The behavioral signal of internal trust, therefore, is not the absence of reported errors but the presence of them, combined with the absence of patterns suggesting that error reports are being filtered before they reach awareness.
The fifth, sixth, and seventh signals can be described together. They are the rate of voluntary information disclosure during exit interviews and stay conversations, the rate of voluntary upward feedback to senior leadership outside formal review processes, and the rate of voluntary participation in cross-organizational initiatives that have no direct career incentive. Each of these is a behavioral indicator of internal trust because each requires the employee to risk something, however small, on the assumption that the organization will not punish the disclosure or the engagement. Where the rates are high, internal trust is strong. Where the rates are low, internal trust is weak, and the absence cannot be explained by survey scores that often show the opposite pattern for all of the reasons I discussed earlier.
Customer Trust
Customer trust reveals itself through a different set of behavioral signals, of which four are most informative. The first is the pattern of customer behavior during a competitive challenge. When a competitor introduces a comparable offer at a lower price or with superior features, the proportion of customers who actively evaluate the alternative versus the proportion who do not is a direct behavioral indicator of how deep the trust relationship runs. Customers who do not evaluate the competing product are signaling that the trust relationship is providing value that the price or feature comparison does not capture. Customers who evaluate but stay are signaling that trust is contributing to retention but is not strong enough to prevent active consideration of alternatives. Customers who evaluate and leave are signaling that the trust relationship was a surface phenomenon and did not survive the test.
The second signal is the willingness to forgive errors. Every organization, over a long enough timeframe, will make mistakes that affect customers. The behavioral indicator of customer trust is the pattern of customer response to those mistakes. Customers who give the organization time to remedy a problem, who maintain their relationship through the remedy, and who continue to refer others to the organization during the period of repair are demonstrating that trust survives the experienced error(s). Customers who exit immediately, who escalate publicly without giving the organization an opportunity to respond, or who silently withhold further engagement are demonstrating that the trust was thinner than the survey scores suggested. The forgiveness pattern is a more reliable trust measurement than any reported satisfaction figure because it captures behavior under the kind of pressure that survey questions cannot replicate.
The third signal is the depth of relational engagement over time. Customers with strong trust relationships expand the surface of their engagement with the organization. They participate in customer councils. They provide feedback when feedback is requested. They share information about their needs that goes beyond what is required to consummate the transaction. They refer others. They tolerate the friction of beta participation or pilot involvement in new offerings. The behavioral indicator is the rate at which customer relationships expand beyond the transactional minimum, and the persistence (maintenance and increase) of that expanded engagement over time.
The fourth signal is the pattern of voluntary advocacy. The Net Promoter Score, in its standard form, asks customers whether they would recommend the organization. The behavioral version of the same question is to track whether customers actually do recommend, measured through documented referrals, social media mentions, public reviews, and word-of-mouth attributions in new customer onboarding. The gap between the willingness-to-recommend score and the actual referral rate is a direct measurement of how much of the reported trust is a sincere intention and how much is a survey-friendly aspiration. Organizations that track only the survey-based version of advocacy are systematically overestimating their customer trust. Organizations that track both and examine the gap have a far more accurate and realistic picture of where they stand.
Partner Trust
Partner trust operates on yet a third set of behavioral signals, of which three are most useful for ongoing sustained measurement. The first is the quality of communication during ambiguity. Partners with strong trust relationships raise questions early, share their own constraints openly, and engage in collaborative problem solving when situations arise that the contract did not anticipate. Partners with weak trust relationships default to contractual interpretation, position themselves for negotiation advantage, and hold back information that might be used against their interests. The behavioral indicator is the pattern of communication when contractual terms do not directly govern the situation. Trust shows itself in the willingness to operate outside the contract and its possible ambiguity rather than retreat to its narrowest interpretation.
The second signal is the rate of bilateral flexibility. Trust relationships between organizations and their suppliers, distributors, or strategic partners produce a pattern of give-and-take in which both sides accommodate the other’s operational needs without requiring contractual renegotiation. The behavioral indicator is the frequency and balance of these accommodations. A partner relationship in which one side consistently accommodates while the other consistently extracts is not a trust relationship. It is a power relationship masquerading as a partnership, and it will produce predictable behavior the moment the power asymmetry shifts. Measuring bilateral flexibility requires tracking the pattern of accommodation requests, responses, and reciprocity over time.
The third signal is the willingness of partners to share commercially sensitive information. Strong partner trust shows up in the willingness to disclose pipeline information, cost structures, customer feedback, and operational constraints that would carry real consequences if disclosed to a competitor or used against the partner in future negotiations. The behavioral indicator is the rate at which such information moves between the organizations, the timeliness of the disclosure, and the use to which the receiving party puts the information once it has been shared. Where partner trust is strong, sensitive information flows freely and is used in ways that preserve the trust. Where partner trust is weak, the flow is restricted, delayed, or filtered through legal and commercial review that adds friction at every step.
Practical Application: A Quarterly Trust Audit
The operational application of behavioral trust measurement is a quarterly trust audit, structured around the three surfaces and the specific behavioral signals appropriate to each. The audit is not a survey. It is a structured observation exercise that combines behavioral data already present in the organization’s operating systems with targeted observation that captures what those systems don’t. Internal trust signals can be observed through analysis of communication patterns, issue escalation timing, cross-functional participation rates, error reporting trends, and exit interview content. Customer trust signals can be observed through retention behavior during competitive events, post-error retention rates, customer engagement depth metrics, and the gap between intended and actual referral activity. Partner trust signals can be observed through communication pattern analysis, accommodation reciprocity tracking, and the volume and nature of information shared outside contractual requirements.
The audit produces a profile rather than a score. The profile shows where trust is strong, where it is eroding, and where the gap between sentiment-based trust measures and behavioral trust measures is widest. The gap is often the most important finding because it identifies the surfaces of trust where the organization is most at risk of behavioral surprise, the kind of surprise that the client experienced when its 84 percent trust score didn’t predict the exodus. Organizations that conduct the audit consistently over several quarters can develop the ability to track trust dynamics in ways that survey instruments can’t replicate, and they build the kind of early warning capability that single-instrument trust measurement systematically misses.
Building this capability takes time, disciplined commitment, and attention. It requires the willingness to invest in observational infrastructure that produces less precise numbers than a survey but far more accurate intelligence about the relational foundation on which all organizational performance depends. The 2024 Chartered Institute of Personnel and Development (CIPD) evidence review on trust and psychological safety, drawing on more than two decades of empirical research, concluded that behavioral observation over time provides a substantially more reliable indicator of organizational trust than any survey-based measure, including instruments designed by leading psychometricians. The methodological consensus from research is clear. The operational adoption, in most organizations, is still incomplete. The bridge between the two is the willingness to treat trust as a behavioral asset that can be measured, rather than a sentimental quality that can only be felt.
Why This All Matters
We measure revenue with precision because the systems that produce and record revenue are visible, and the consequences of getting it wrong are immediate. We measure trust with almost nothing at all because the systems that produce trust are invisible, the consequences are delayed, and the survey instruments we use produce numbers that feel reassuring even when they are wrong. The organizations that change this pattern will not be the ones that find a way to reduce trust to a single number. They will be the ones that develop the discipline of behavioral observation across the three surfaces of trust, that hold the gap between sentiment and behavior in continuous view, and that treat trust measurement as a strategic capability rather than a public relations exercise. The question for every leader and their leadership team is not whether trust matters. It is whether the organization is willing to measure trust honestly, by what people actually do, rather than reassuringly, by what they say on a structured form designed to make the organization feel good about the relationships it claims to have built.
Further Reading
The themes explored in this article connect to several other pieces in the Measuring What Matters series and the broader Ideas and Innovations archive.
The Value of Intangibles (Measuring What Matters Series) examines why 92 percent of organizational value resides in intangible assets, with trust as one of the most consequential of those assets.
How Do You Trust? (Issue 42) offers an early exploration of trust as an organizational asset and the mechanisms through which trust is built, damaged, and rebuilt over time.
The Leading Indicators of Resistance (Measuring What Matters Series) describes behavioral signals that reveal compliance without conviction, with direct parallels to the gap between reported trust and behavioral trust.
Taking the Pulse of Your Employees (Newsletter) examines practical approaches to understanding employee sentiment beyond annual surveys, including behavioral indicators that complement direct sentiment measurement.
Structural Silence (Issue 256) explores why organizations train people not to speak, and how the resulting silence corrodes internal trust in ways that survey instruments rarely capture.
The Organizational Memory Problem (Measuring What Matters Series) describes how institutional knowledge erodes in environments where internal trust is weak, with implications for the connection between trust measurement and learning measurement.
Sources
Mayer, R.C., Davis, J.H., & Schoorman, F.D. (1995). “An Integrative Model of Organizational Trust.” Academy of Management Review, 20(3), 709–734.
Edmondson, A.C. (1999). “Psychological Safety and Learning Behavior in Work Teams.” Administrative Science Quarterly, 44(2), 350–383.
Edmondson, A.C. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Wiley.
Frei, F.X. & Morriss, A. (2020). “Begin with Trust.” Harvard Business Review, 98(3), 112–121.
Frei, F.X. & Morriss, A. (2020). Unleashed: The Unapologetic Leader’s Guide to Empowering Everyone Around You. Harvard Business Review Press.
Fukuyama, F. (1995). Trust: The Social Virtues and the Creation of Prosperity. Free Press.
Edelman (2025). 2025 Edelman Trust Barometer: Global Report.
CIPD (2024). Trust and Psychological Safety: An Evidence Review. Chartered Institute of Personnel and Development.
Covey, S.M.R. (2006). The Speed of Trust: The One Thing That Changes Everything. Free Press.
Maister, D.H., Green, C.H., & Galford, R.M. (2000). The Trusted Advisor. Free Press.
This is such a great read and resource! I’m currently transitioning to consulting so I look forward to digging into the provided resources and your book to help me refine my knowledge.