Care to Deliver

Most organizations try to deliver first and figure out the human conditions later. I’ve run it the other way for fifteen years, across engineering and leadership, and I’ve measured the trade more than once: it holds.

Care to Deliver isn’t a process. It’s a sequence of conditions, and the order matters — skip a link and the chain breaks somewhere downstream, usually where you can’t see it yet.

Sharing is the hinge

Everything upstream of trust is cheap to say and hard to do. Sharing is where it gets real, because sharing isn’t only information — it’s context, uncertainty, mistakes, credit, and responsibility. That requires vulnerability, and sustained vulnerability is what becomes trust.

So a team that hoards knowledge, moves hard conversations into private channels, or stops asking questions in the open isn’t failing at communication. It’s protecting itself. The behavior is the symptom; the missing condition is safety.

Care isn’t a trade

The obvious objection is that this sounds transactional: I care, therefore you owe me something back.

It isn’t, and the distinction does real work. People often do respond to being treated well, and that response is what lets the chain spread past one person instead of stalling there. But the moment you’re caring in order to get it, care has become a technique — and people read that faster than you’d like.

So the model doesn’t say “I care so you’ll care back.” It says: I show up in a way that makes better participation possible, and what anyone does with that is theirs.

Care creates the conditions. Reciprocity amplifies them.

The practical consequence is that the method still holds when nobody reciprocates. Some people won’t. Some environments punish openness outright. That isn’t the method failing.

Operating principles

Each of these corrects something the framework gets mistaken for.

Care is not niceness.

Care should make honest disagreement, clear boundaries and real correction possible — and the relationship survive all three.

Rules out Artificial comfort · Avoiding the hard conversation · Quietly lowering the bar

Trust is not blind.

Earned and verified. Delegate and verify — one move, not two.

Commitment is not consensus.

After open discussion a leader may still decide, and the team commits. Without that, a team that takes safety seriously can’t decide at all.

Keep apart Voicing disagreement and blocking it · Committing and being persuaded

Activity is not outcome.

Measure what was achieved, not what was done.

Information is not private power.

Hoarding knowledge buys status; sharing spreads capability. It’s the one people defend hardest, usually without saying so.

Hoarding manufactures Scarcity · Dependency · Fear

Safety sits next to accountability, not instead of it

Psychological safety is a felt condition, not a policy. You can’t announce it. You can only produce the conditions where people ask questions, admit mistakes and challenge ideas without calculating the cost first.

But safety alone isn’t the goal:

Safety without accountability turns into complacency. Accountability without safety turns into fear.

Strong teams hold both at once — the harder thing, and the one most teams get wrong in one direction or the other.

Accountability has an order

Care doesn’t mean protecting people from the consequences of their mistakes. It means creating the conditions to face the error together, resolve what’s urgent, and then change the system that allowed it.

The order is the part that does the work:

1. In the moment it doesn’t matter who — it matters what.

   What broke, and what it takes to fix it. A leader whose first question after a failure names a person has already made every future disclosure expensive.

2. Afterwards, who matters.

   The person needs the feedback, and withholding it isn’t kindness. So does the boundary — help given once isn’t help guaranteed.

   Name when afterwards is at the moment you defer it, or the deferral is avoidance and looks like avoidance to everyone watching. A retrospective on the calendar guarantees a meeting; it never guarantees the thing gets named in it.

3. Then the system changes.

   Process, policy, or resourcing, and resourcing can mean people. A correction that stops at the individual leaves the conditions that produced the failure exactly where they were.

Collapse that order and it breaks in both directions. Ask who first and people learn to conceal. Never ask who and nobody learns anything. Stop before the system and the team quietly absorbs the same failure for years.

Teams don’t fail suddenly — they stagnate

They slow down, and from the inside it still looks functional.

Signals Blockers stay unresolved · Channels go quiet · Meetings turn into reporting rituals · Experimentation stops

The fix is movement, deliberately introduced. Fishermen used to put a shark in the tank during transport: without it the fish stop moving and arrive unhealthy. On a team, the shark is a person who asks the uncomfortable question, challenges the assumption, pushes for clarity. Not conflict for its own sake — movement.

Tension only works inside safety. Outside of it, it’s just pressure.

What it produced

Outcomes that look improbable from the outside usually weren’t forced directly. They came from a system healthy enough to produce them.

As COO of a ~70-person consultancy it moved three numbers in six months — no new metrics, no mandates, trust rebuilt first and delivery allowed to follow (full case study →).

That last pairing is the whole argument. Utilization near 99% on its own is what a squeezed workforce looks like; an eNPS of 95 on its own is what a comfortable company that doesn’t deliver looks like. Getting both at the same time is not a trade you can negotiate — it’s what happens when the conditions are right. A year later leadership asked me to treat people as cost lines instead of the mechanism that was working. I left. That was a decision to protect the method, not to abandon it.