Task-Based Trust, Not Relational Trust, Is What Predicts Team Performance

Key Takeaways 

• Two kinds of trust get treated as one in most team effectiveness work: relational trust (bonding, warmth) and task-based trust (reliability, follow-through). Only one predicts performance [1]. 

  
  

• Teams that get on well personally can still miss the deadlines they've promised each other. Warmth and reliability are different things entirely. 

  
  

• Trust is contextual, permission granted for one thing under one set of conditions, not automatically for another. Relational and task-based trust are two different permissions, not two strengths of one quality. 

  
  

• Lencioni's own vulnerability-based trust, admitting mistakes and asking for help, is a task-based habit. It gets diluted into a bonding exercise more often than it gets tested. 

  
  

• Task-based trust is built by diagnosing what the work actually demands and composing the team accordingly, then reinforcing it through clear decision rights, visible accountability, and coaching inside real meetings, not offsites. 

  
  

• When the roster can't be changed, composition matters less than structure: Google's own research found team dependability tracked with clarity and explicit expertise, not with who was in the room [4][5]. 

  
  

• 75% of cross-functional teams are dysfunctional [3]. 

  
  

• Building relational trust without task-based trust behind it doesn't fix a team. It makes the dysfunction more comfortable to sit inside. 

  
  

Two members of an executive team can trust each other completely and still let the same deadline slip for the third quarter running. They back each other in the boardroom, defend each other publicly, grab a drink together after a hard week. None of that tells you whether either one will actually follow through on what they said they'd do. 

Most team effectiveness work quietly conflates two things that don't behave the same way. 75% of cross-functional teams are dysfunctional [3], and a good share of that dysfunction survives inside teams that would describe themselves, honestly, as close. 

The Trust That Feels Good and the Trust That Actually Works 

Colin Fisher, an organisational psychologist at UCL, puts the distinction plainly: relational trust is ‘I trust you with my deepest, darkest secrets, or I trust that you have my best interests at heart.’ Task-based trust is different: ‘I trust that you're going to deliver your part of the work, at a certain standard, to a certain quality, at a certain time.’ The two don't reliably travel together. You can trust a close friend or family member completely with the first and still have no idea whether you'd trust them to run a thorough stakeholder analysis. You can trust a colleague to deliver against a joint objective without ever confiding in them personally, and for a team's purposes, that's the trust that actually matters. 

A team can be strong on the first and weak on the second, and this is exactly where most team-building work goes wrong. An offsite that builds warmth without ever testing reliability produces a team that likes each other more and still can't predict each other's follow-through. The two are genuinely separate variables, and treating them as one is how a team ends up feeling closer while performing exactly the same. 

For a closer look at what effective team-building work typically covers and how to get it right, see Esendia's Top Team Effectiveness Handbook.

What Lencioni Actually Meant by Vulnerability 

No one brought the role of trust in teamwork to wider attention more than Patrick Lencioni. His seminal The Five Dysfunctions of a Team set out five failure points stacked on top of each other, each one building on the one below it: absence of trust, fear of conflict, lack of commitment, avoidance of accountability, and inattention to results. Trust sits at the base of that model, which is exactly why it's worth being precise about what Lencioni actually meant by it. 

The trust he places at the foundation is what he calls vulnerability-based trust, the willingness to admit mistakes and ask for help without fear of losing face. In practice, it's often diluted into a bonding exercise: a trust fall, a round of personal disclosures, an afternoon that feels warm and produces very little that transfers back to how the team actually works. 

Read correctly, it isn't a relational habit at all. A team member who says ‘I'm behind on this’ or ‘I don't know how to solve this’ out loud, in front of the group, is demonstrating reliability under pressure: the willingness to surface a problem before it becomes everyone else's problem too. That's task-based trust wearing the language of vulnerability. Teams that mistake it for a bonding exercise get the warmth without ever running the actual test. 

How Task-Based Trust Actually Gets Built 

If it isn't built through bonding, what does build it? Fisher's own answer starts earlier than most teams think to look: before assessing whether a team can be trusted to deliver, ask what the work in front of it genuinely demands, specific skills, specific knowledge, and check honestly whether the team was actually composed with that demand in mind. Most leadership teams weren't, because that's not how they get composed in the first place. A leadership team is usually built from the organisational chart: whoever holds the top role in each function is in the room, by virtue of their position, not because anyone checked whether that specific mix of people covers what the team's actual mandate requires. No amount of trust-building addresses a gap that was there from the composition stage. 

That's a real constraint, and most leadership teams are stuck with it. The reassuring finding here comes from Google's own research into what separates its best teams, which found that composition mattered less than how the team actually worked together [4]. Dependability, their term for essentially this same trust, tracked far more closely with structure and clarity, decision rights and goals stated explicitly enough that everyone in the room would describe them the same way, than with who specifically was in the room. Google's teams built this through a short, structured check-in against exactly this question, discussed openly, rather than leaving it as an unspoken feeling nobody named directly. 

A second lever works even when the roster is completely fixed: making expertise explicit rather than assumed. Teams naturally develop a shared sense of who knows what, but left undesigned, that sense defaults to whoever is most senior or most vocal rather than whoever actually holds the relevant judgement. Naming, out loud, who specifically owns which kind of call, and routing decisions there deliberately rather than to habit, is what researchers studying this call task credibility: literally, how much a team trusts a given member's knowledge in a given domain [5]. It's built through visibility, not through changing who's on the team. 

Day to day, the same logic holds at a smaller scale. Task-based trust builds through accountability mechanisms that make follow-through visible to the team rather than private between two people, and through coaching that happens inside real working meetings, on the actual decision in front of the team, rather than in a separate session removed from the work itself. None of this is a bonding exercise. It's structural, and it produces trust as a byproduct of the team functioning well, not as a goal pursued directly. 

Why Relational Trust Still Matters, Just Not for This 

None of this makes relational trust worthless. Rachel Botsman's work on trust makes a related point worth borrowing carefully [2]: trust isn't a single feeling but a form of permission, and that permission is contextual, trusted by whom, to do what, under what conditions. Someone can be trusted completely in one context and not at all in another, without either judgement being wrong. That's consistent with Fisher's split, even though Botsman's work is about trust and risk as leadership lenses generally, not team performance specifically, and shouldn't be read as making the same claim. 

Applied here, the point is narrow but useful: relational trust isn't a lesser version of task-based trust, it's permission granted in a different domain entirely. A team doesn't need to choose between them. It needs to stop assuming that earning one automatically earns the other. 

For a broader look at what makes teams effective in the first place, see The Complete Guide to Team Effectiveness: Building High-Performing Teams That Actually Work.

What the Top Team Gets Wrong, the Rest of the Organisation Inherits 

None of this stays contained at the top. A leadership team that hasn't built task-based trust tends to hold decisions closer, because members genuinely aren't sure whether a colleague will follow through without checking. That caution cascades: the layer below watches the top team second-guess itself and stops taking initiative for exactly the same reason. Fixing trust at the top without addressing what's already cascaded downward leaves the real cost of the problem untouched. 

This is the same structural point from a different angle: unclear decision rights make task-based trust almost impossible to build, because nobody can be reliably accountable for a call that was never clearly theirs to make. No amount of relationship-building fixes a team whose actual problem is a broken decision-rights map. 

	Relational trust	Task-based trust
What it feels like	Warmth, personal disclosure, being liked	Confidence someone will do what they said, and flag it early if they can't
How it's usually built	Offsites, social time, shared personal stories	Tested follow-through, visible accountability, structural clarity on who owns what
What it predicts	How much the team likes working together	Whether the team actually delivers

“As a team, we get on well, which is great, but as an effective leadership team, we also need to be strategic and hold each other to account. Ines challenged our thinking. With the psychometrics, she helped us accelerate our understanding of our blind spots as much as our strengths.” Katherine Allison, Chief People Officer, Moore Barlow 

What to Actually Test, Not Just Discuss 

The honest question for any leadership team isn't whether people get on. It's whether the team could predict, with any confidence, which of its members would flag a problem before it became visible to everyone else. If that answer is uncertain, the fix isn't a better offsite. It's a proper diagnostic that separates what the team feels from what the team can actually rely on. We're covering exactly this distinction, and what to do about it, in our Top Team Effectiveness webinar on 9 September. 

Frequently Asked Questions 

Can a team have strong relational trust and weak task-based trust at the same time? 

Yes, and this is the most common pattern in teams that describe themselves as close but underperforming. The two are separate variables measuring different things, not two strengths of the same quality. 

Does building task-based trust mean giving up on relational trust entirely? 

No. Relational trust is genuinely valuable, and it plays a real role as the precondition for a colleague's reliability being read fairly in the first place. It just isn't the variable that predicts performance on its own, and task-based trust is the one that needs deliberate, tested attention rather than assumed to follow from the other. 

Is vulnerability-based trust the same thing as relational trust? 

It's often treated that way and shouldn't be. Admitting a mistake or asking for help out loud is a demonstration of reliability under pressure, a task-based behaviour, not a bonding exercise, even though it can feel emotionally similar to one. 

What if I've inherited a team and can't change who's on it? 

That's the normal situation, not the exception, and it's not the obstacle it seems. Google's own research found team dependability tracked with structure and clarity, not with composition, and a second lever, making expertise explicit rather than assumed, works with any fixed roster. Neither requires changing who's in the room. 

Is an offsite ever the right intervention? 

It can be, if it's built around testing and building task-based reliability, not just shared time. An offsite that only builds warmth without addressing accountability or decision clarity tends to raise how much a team likes each other without changing what it delivers. 

References 

[1] Fisher, C. (2025) The Collective Edge: Unlocking the Secret Power of Groups. Discussed on Nudge (host: Phil Agnew), 'Is This Famous Team-Building Model Wrong?', 22 September 2025. Available at: https://metacast.app/podcast/nudge/Tri4jzdZ/is-this-famous-team-building-model-wrong/L05wU9Jp 

[2] Botsman, R. (2026) Keynote address on trust. Workhuman Forum London, Hilton Park Lane, London, 20 May 2026. 

[3] Haas, M. and Mortensen, M. (2016) The Secrets of Great Teamwork. Harvard Business Review. Available at: https://hbr.org/2015/06/75-of-cross-functional-teams-are-dysfunctional 

[4] Rozovsky, J. (2015) The Five Keys to a Successful Google Team. re:Work Blog, Google. Available at: https://rework.withgoogle.com/en/subjects/team-effectiveness 

[5] Lewis, K. (2003) Measuring Transactive Memory Systems in the Field: Scale Development and Validation. Journal of Applied Psychology, 88(4), 587-604.