I’ve put this post together using the information I have picked up doing contract work with Engage Process and getting to speak to their 90+ UK Local Gov customers on how they are delivering successful transformation and continuous improvement. These are some observations and musings around the decentralisation of process mapping.
Here is the uncomfortable maths
Lets assume that good central improvement team of four analysts will properly map and improve somewhere between 20 and 40 processes a year. A district council runs several hundred. A unitary runs well over a thousand. A new unitary formed from five predecessors inherits five versions of most of them.
At that rate, a council would work through its own process estate roughly once every thirty years. By which point the first ones mapped would be wrong.
Centralised process mapping is not a capability, it is a queue. And the queue is permanently longer than the year.
The alternative is to stop treating mapping as something a small team does to services and start treating it as something service teams do themselves, supported by a centre that sets the standards and holds the quality bar. That is decentralisation, and modern tooling has finally made it realistic.
Why this failed before
It is worth being honest about why councils tried this before and gave up:
- Visio and BPMN require notation training
- The artefacts they produce are readable by analysts and nobody else.
- The files end up scattered across shared drives, network folders and someone’s laptop.
- Version control is a filename with “v3 FINAL (2)” in it.
So decentralisation produced fragmentation, inconsistency and a lot of maps that nobody trusted. Councils recentralised and the queue came back.
However, the functionality development in process management platforms are now built for workshopping as opposed to the analyst’s desk and complete changes the discussion:
Process Mapping workshops can be carried out live, with subject matter experts involved and this means that the process is discussed, with comments noted and workarounds revealed.
Platform complexity is low enough to train at scale which means that training 200 “citizen mappers” is more than achievable.
Ideas, bottlenecks and staff sentiment are captured at the point of mapping which means improvement suggestions get recorded and voted on while people are still engaged, instead of evaporating between the workshop and the report.
Everything lives in one repository which prevents decentralisation becoming fragmentation. Map locally, store centrally, with consistent structure, named owners and review dates.
Time, cost and volume analysis is built in which means a service manager can build their own business case rather than waiting in a second queue behind finance.
Maps can be published as live process guidance which means they no longer are “project artefacts” but become operational documentation, induction material and knowledge management that people can actually use.
So what does this means to councils today?
Let’s consider 2 scenarios: Councils that aren’t going through LGR and those that are.
If you are not going through LGR
The case here is about capacity, resilience and the “hidden”costs that never appear on a savings plan.
You get coverage instead of a queue. Once mapping capability sits with services, throughput stops being a function of how many analysts you can fund and becomes a function of how many staff you can train. That is a different curve entirely.
You get the real process and not the described one. Interview-led mapping captures the process as managers understand it. Direct capture by the people doing the work surfaces the exceptions, the workarounds, the spreadsheet that lives outside the case management system and the three times a year it all goes wrong. That gap is almost always where the cost is hiding.
Improvements stick. Changes designed by the team that runs the process survive contact with business as usual. Changes handed over at the end of a project revert the moment the analyst moves on. Decentralisation is a change management strategy wearing a process mapping costume.
Your central team moves up the value chain. Released from mapping everything, a small central function does the work only it can do: setting conventions, assuring quality, spotting patterns across services and taking on the genuinely complex end-to-end redesigns. It also starts to see the same failure demand appearing in four different places, which is the insight no single service can generate on its own.
You get resilience. Undocumented processes sit in the heads of long-serving officers. Every retirement is an unmanaged risk; distributed mapping, published as live guidance, turns that into documented and transferable practice.
You get audit and assurance evidence as a by-product. Internal audit, ISO work, information governance reviews and inspection preparation all ask variations of “show us how this actually works.” A maintained process library answers that question in an afternoon rather than a fortnight.
And you can build the automation case properly. Every council has an automation ambition and a list of candidate processes assembled largely on instinct. Volume, cycle time and hand-off data across a wide estate of mapped processes turns that list into a ranked pipeline. You cannot automate what you have not mapped and you should not automate a bad process at speed.
And for those going through LGR
Everything above still applies. But reorganisation adds a constraint that makes decentralisation less of a preference and more of a mathematical necessity.
The timescale does not permit anything else. You need to understand how a number of sovereign councils each handle garden waste, licensing, debt recovery and homelessness applications, and you need it in time to inform a day-one target operating model. A central team of four cannot do that. Parallel mapping across all predecessor organisations, into a shared repository with common conventions is the only approach that fits inside the window.
Comparability is the whole point. Multiple councils mapping independently in multiple formats gives you multiple PDFs and an argument. Multiple councils mapping to the same conventions into the same library gives you a genuine evidence base for convergence decisions. You can see, side by side, that one predecessor does this in six steps and another does it in nineteen, and have a realistic conversation about which one is right rather than which one is loudest.
Safe and legal on day one becomes demonstrable. Continuity of service is the minimum bar for vesting day. Knowing what each predecessor actually does, in detail, is how you evidence that the bar has been cleared and how you spot the process that only works because of an arrangement in a council that is about to cease to exist.
Convergence gets sequenced by evidence. Not everything can harmonise on day one. Volume, cost and risk data across mapped predecessor processes tells you what must converge immediately, what can run in parallel for a year, and what can wait. That is a far better basis for sequencing than service-by-service negotiation.
Knowledge stops walking out of the door. Reorganisation will inevitably lead to staff attrition and it drives it hardest among experienced staff with options. Every one of them is carrying undocumented process knowledge. Distributed mapping captures it while those people are still in post, which is a narrow window that closes quietly.
Staff get a stake in the new organisation. Mapping workshops are one of the very few LGR activities where frontline staff are asked what they know and what they would change, rather than told what has been decided. The engagement value is real, and in a period defined by uncertainty, it is not a small thing.
And the library becomes a founding asset. The new authority starts life with a documented, owned, published process estate, rather than a blank repository and a plan to get round to it. Given everything else a new unitary has to do in year one, that is a meaningful head start.
THE HONEST CAVEATS
Decentralisation without governance produces inconsistent quality, duplicated effort and mapping for its own sake.
Three things have to be true:
A centre of excellence, not a mapping factory. Someone has to own naming conventions, levelling standards, the definition of “done” and a MVP quality control. MVP matters – insisting on perfect rebuilds the queue you just dismantled.
Protected time. If mapping is what staff do after the day job, it will not happen. This needs to be scheduled, sponsored and visibly backed at SLT level.
Licensing that does not punish participation. Per-seat pricing quietly recentralises everything because the rational response to per-seat costs is to limit seats. Check the commercial model before designing the operating model.
So, in summary:
The choice is not between a central team and a distributed one. It is between mapping a small fraction of your processes properly and mapping most of them well enough to act on.
For a stable council, that is the difference between an improvement programme and an improvement capability.
For a reorganising one, it is the difference between designing the new authority on evidence and designing it on assertion.
Neither is a tooling decision. But without tooling built for workshops rather than analysts, neither is achievable.