Most advice about trust asks you to be something — more open, more authentic, more consistent. That is hard to act on, because you cannot check tomorrow morning whether you were more authentic. Behaviour you can check.

What are the twenty behaviours that strengthen trust?

They are twenty observable actions, grouped under the five Trust Questions. Each one is something another person could see you do or not do.

Do they believe what I say? — Clarity & Honesty

  1. Say what is true without using vague or confusing language.
  2. Share relevant information openly, not selectively.
  3. Communicate consistently with different people on the same issue.

Do they believe I’ll do what I said? — Delivery & Consistency

  1. Set and confirm clear expectations before work begins.
  2. Give updates on progress and raise issues early when there are risks or delays.
  3. Pause before responding when under pressure.
  4. Apply the same standards in tone, decisions and communication across people and situations.

Do they believe I can handle what’s needed? — Capability & Realism

  1. Only agree to what you can realistically deliver and control.
  2. Prepare before key interactions and decisions.
  3. Ask for support when needed.
  4. Say clearly when something cannot be delivered.

Do they feel I really understand them? — Understanding & Connection

  1. Listen and let others finish before responding.
  2. Ask questions and check your understanding before replying.
  3. Acknowledge how others feel before responding to the task.
  4. Show openness to others’ ideas, even when you disagree.

Do they believe I act with their interests in mind? — Fairness & Consideration

  1. Address issues directly with the person involved.
  2. Involve people in decisions that affect them.
  3. Explain the reasoning behind decisions that impact others.
  4. Give credit to others for their contribution.
  5. Consider how your actions may affect others before acting.

Why these behaviours and not others?

Because each one is a mechanism, not a virtue. Look at what a behaviour actually does and the list stops looking arbitrary.

None of these behaviours works because it is a good thing to do. Each one closes a gap between what someone expected of you and what they got.

Behaviours 8 to 11 stop the gap opening. You only commit to what you can control, you prepare, you ask for help before you need it, and you say plainly when something can’t be done. Behaviours 4 to 7 manage the gap once work is under way — you set the expectation, you update it as things change, and you don’t become a different person under pressure. Behaviours 1 to 3 apply the same discipline to information rather than commitments. Saying the same thing to three different people is not a courtesy. It’s the same alignment problem, in a different medium.

That is twelve of the twenty doing one job: keeping what you said and what happened close enough together that nobody has to guess.

Behaviours 16 to 20 do something different. They govern how a decision reaches the people it lands on — whether you take it to them directly, whether they had a say, whether you explained your reasoning, whether you noticed who did the work.

Behaviours 12 to 15 are the ones people expect to see on a list like this, and they work for a reason that is easy to miss. Our definition of trust is someone’s confidence in you with something they care about. You cannot act on what someone cares about until you know what it is, and the only way to find out is to listen properly, check what you heard, and stay open when they say something you did not expect. Listening isn’t a nice quality. It’s how you get the information the other four groups depend on.

Which of these does research actually support?

The alignment argument is the best supported part of the list.

Tony Simons described behavioural integrity as the alignment a person perceives between what you say and what you do. It is not something you can audit from your own side, because it is their judgement about your pattern, not your view of your own intentions. That is the mechanism behind behaviours 1 to 7, and it is why behaviour 3 matters as much as it does — a different message to a different audience is a word-and-deed gap that someone will eventually see.

Consistency is also the strongest single influence anyone has measured. A 2026 meta-analysis pooled 404 effect sizes from more than eight thousand people and found that how consistently a partner reciprocated mattered more than anything else tested. Those are repeated laboratory games rather than working relationships, so the direction travels further than the rate does.

For behaviour 18, there is direct support. A longitudinal study by Jason Colquitt and Jessica Rodell found that the explanations people receive predict how much they trust later — not just how they feel at the time.

Behaviour 5 has an unusual piece of evidence behind it. Maurice Schweitzer and colleagues compared trust damaged by unreliability with trust damaged by deception. Unreliability recovered as behaviour improved. Deception did not recover at all, even when the person’s later behaviour was good. That is the strongest argument on this list for raising a problem early — not because early news is pleasant, but because the concealment costs more than the failure it hides.

And the reason early updates are rare is well documented. Sidney Rosen and Abraham Tesser named the MUM effect in 1970: people are reluctant to pass on bad news, even when they had nothing to do with it. Behaviour 5 asks you to work against a documented instinct, which is why it needs practice rather than agreement.

Where does the evidence stop?

At behaviours 12, 13, 14, 15 and 19.

We could not find direct evidence that listening, checking understanding, acknowledging how someone feels, staying open to their ideas or giving credit cause another person to trust you more. The studies we would want don’t appear to exist, and we would rather tell you that than dress up something adjacent.

What we have instead is an argument. Behaviours 12 to 15 follow from the definition — you can’t act on what someone cares about without finding out what it is. Behaviour 19 is our own thinking, and it comes from watching the same thing go wrong repeatedly: the credit itself matters less than whether people believe you notice.

Treat those five as our position rather than a finding, and test them against your own experience. If they don’t hold there, they don’t hold.

Which behaviour should you start with?

The one the other person would notice.

Working through twenty behaviours is not the point, and doing all of them slightly is worse than doing one of them properly. Pick the person whose confidence matters for something specific, work out which of the five questions is actually in doubt, and take the behaviour that answers it.

Trust is built with specific people, in specific situations. A behaviour practised generally, on nobody in particular, is a habit without an audience.

What this list is not

It is not a personality test, and it is not a score. The behaviours describe what a person could observe, which means they say nothing about who you are and everything about what someone has experienced from you.

It is also not a ranking. Behaviour 1 is not more important than behaviour 20 — the numbers are an index, nothing more. Which behaviour matters most depends entirely on which of the five questions the other person is quietly asking.

Each of these has an opposite. Those are the behaviours that erode trust, and they pair one to one with the twenty here.