Relational public services good, transactional bad? That’s not how real systems work. And it’s not how change happens. My ladder of degrees of relationality might help

Relational public services good, transactional bad? That’s not how real systems work. And it’s not how change happens. My ladder of degrees of relationality might help: https://www.linkedin.com/posts/antlerboy_relational-public-services-good-transactional-activity-7434526591252082688-DA_l

Relational public services good, transactional bad?

That’s not how real systems work. And it’s not how change happens.

I’ve sketched a way to think about this: not a maturity model, a ladder of degrees of relationality. It’s diagnostic, and a way to plan movement. If we see relationality as a set of modes that coexist in one place, at one time, the question is: what are we designing to increase?

A quick run up the ladder.

Most public services aren’t ‘non-relational’. They’re relational in this way: power sits with the system, and the citizen has to adapt.

-2 ‘the boot’ (done to). Coercion. Sometimes necessary. Always dangerous.

-1 ‘school dinners’ (done for). Indifferent provision. “We met the spec.”

0 ‘IVR palisade’ (done for). Scripted bureaucracy. Citizens translate life into categories.

Level 1 is where ‘relationality’ starts.

1 basic relationality. Listen at the point of demand. Design against demand. Reserve some discretion.

Then the gradients that people often blur:

2 Good Help. Same contact, different purpose. Build agency, not dependency.

3 coordinated provision. Orchestrate across providers around one person’s situation over time. Vital. But it can become heroic navigation through a broken maze.

4 tending the garden. Fix the maze: rules, measures, money, data, commissioning, eligibility. Mark Smith nailed the trap: ‘If you start with services as your focus for change, you end up with services.’

5 place operating system. Make whole-place coordination the default, not the exception. Shared authority at the edge. Shared learning loops. Shared measures that don’t punish relational work.

6 citizen space. Not ‘more integration’. A shift of centre of gravity: citizens in association are primary. Institutions become guests and allies

Two practical tests if you’re thinking about this:

Where does authority sit at the boundary? With scripts, or with judgement? And what happens when we automate that boundary?

What is the citizen’s role in design, delivery, and management? Object? Input? Participant? Owner?

And two caveats

You need multiple degrees at once. Standardise when you can. Personalise when you must. Be deliberate, not ideological.

And every degree must be designed to progress, or they will retrogress. Measures drive behaviour drive purpose, so design to learn. or example, there’s always wiggle room. Street-level bureaucrats always have discretion. But automation is limiting this and moving it into software and data models.

If you’re a manager thinking ‘if I can’t scale or target it, what can I do with it?’: scale is not only replication. You can scale rules, measures, permission, capability, and learning. You can scale the conditions for good discretion.

If you’re working in this space, where do you see your system sitting most days? What would you change in this model?

See Google doc https://docs.google.com/document/d/1FewYqWndCJbjGYOlkWFZGBHxbs3AM4lA/edit?usp=sharing&ouid=115526108239573817578&rtpof=true&sd=true for comments