A catalogue of things that are stopping change. Part III, governance.

7 min readApr 7, 2020

It’s been a good few weeks since we promised to publish this blog, since then an awful lot has happened. Like our colleagues across government we’ve been focusing on working with our teams to support the response.

The reality of responding to a situation like this is that things are changing day by day and hour by hour. We need to be responsive to the changing context, do the right thing for users, especially the front line and coordinate with the rest of the government. Long standing governance mechanisms like monthly board meetings, detailed risk logs and gantt charts fail to facilitate the responsiveness that we so crucially need. This certainly isn’t a time to dispense with governance but never has it been so important to implement it in a way that supports our work.

In this post, we are taking the term ‘governance’ to refer to the whole range of activities including the way people outside a team oversee, monitor, challenge, understand, support, direct, and control our work.

First, a little history

Back in 2013 Mike Bracken spoke about GDS supporting good governance through providing clear principles and shared understanding. GDS went to some lengths to reduce the burdensome nature of the way the government’s technology estate was managed. The image gives an indication of how complex the structure of governing boards was, and shows the many possible points where delivering value could be blocked. GDS realised it would have to close these boards and find new ways to govern its technology if it was to ship the exemplar services and clear the path for others.

We think GDS were onto something with their governance-as-a-service approach. It’s evident to anyone doing delivery, however, that departments have a long way to go in terms of how they govern their services. We have grouped the shortcomings into four groups:

Tribal divides
Organisational design
Maintaining the status quo
Funding

Tribal divides

The difference between happy clappy digital ‘show the thing’ governance culture and the stodgy, heavier handed traditional governance is stark. Neither approach is perfect, but people tend to fall clearly on one or the other side of the cultural divide. This often results in a stubborn tribal dogma on all sides, rather than the kind of open communication that’s needed to reach a pragmatic middle ground.

Digital “show the thing” governance

The lovely Helen and Matt presenting a roadmap

Good for:

showing what has actually been done
demonstrating progress in a regular, tangible and accessible way
building empathy with users
people who don’t like to or have time to read things
encouraging feedback and conversations between service teams and stakeholders

Bad for:

working for senior stakeholders
effective communication of more complex services, particularly over a longer period of time
decision making
people who are not bought into digital ways of working

Traditional governance

Good for:

keeping an audit trail, being consistent and detailed
governance through ‘correspondence’
people who like to read long-form text (a lot of senior civil servants)

Bad for:

making sure decisions are made at the right level, empowering teams and making sure that people who understand the problem choose the best course of action
board theatre, the verbal jousting between people with different views who think that sounding good is more important than user outcomes
An exhaustive risk-and-fear register, RAG ratings and 18 month plans can give a false but tempting sense of security, an illusion of safety, certainty and order in the complex, uncertain world of service delivery
Writing things down can also be a way to cover our backs rather than solving problems
Monthly board meetings or design authorities provide infrequent ‘big-bang’ style governance which is costly in terms of both time, money and, often, team morale

It’s good to write things down, track and store, but admit that’s what it’s for: Audit. Not the active management of services or policy. Conversely, digital and agile are not excuses for not writing anything down or not managing work properly. We have a responsibility to deliver value and to demonstrate we are doing so in a controlled and continuous way.

Organisational design

The reporting lines and funding streams at the core of the workings of our organisations supports tribalism and hierarchy, often at the expense of allowing groups of people with the right range of skills and expertise to solve problems as a team.

To add to this there’s often a spread of accountability throughout the organisational hierarchy. That generally means there’s a high number of ‘hurdles’ that our teams have to jump before major decisions are made.

Decision-makers are generally those with senior positions in the organisational hierarchy. There are those (of both junior and senior persuasions) who believe seniority is all about being able to make the best decisions, personally. This can be dangerous thinking as good decision making has to be informed by the facts, and senior people do not have the time to appreciate all of these personally. It’s those who are in the detail: researchers, analysts, people who are best placed to understand users, constraints and context who have the collective knowledge to make the best decisions.

Maintaining the status quo

This is one that applies equally across the professions, but I’m going to pick on the technology crowd to illustrate this point.

The technology that we use to power government services has evolved hugely over the last 20 years. For example, setting up a server used to involve (and to a certain extent still does involve) obtaining physical assets that needed to be secured, cooled and connected to other “tin and wires”. The modern equivalent of this is possible without leaving your desk since it’s possible to spin up an S3 bucket from a browser in a matter of minutes. Tin and wires has been replaced by user interfaces and code.

Michael Brunton-Spall put this really neatly in one of his recent Cyberweekly newsletters:

“The failure here, as always, is the failure of applying a previous generational set of tools and processes to a new generation of capabilities. The processes that you use to manage a physical data centre are optimised over years of being applied to physical infrastructure and are therefore inherently flawed in attempting to manage cloud workloads.”

The risk profile that may have led the IT professionals of 30 years ago to keep their servers somewhere close have changed entirely. With the birth of things like secure cloud environments and the ability to store data in multiple world regions to increase resilience comes the need to take an entirely different view on where the real risks lie. However organisational instinct can often be to hang onto what it knows, in IT as well as other areas.

Advances in technology, applications and ways of working will invariably seep into government, whether organisations have actively embraced them or not. Shadow IT, the use of unauthorised SaaS platforms and general pressures to use new approaches and technology is rife. This situation will continue to be so until people get the tools they need, expect and work for them, and that will never be static.

Funding

The way that the Treasury asks departments to bid for and justify expenditure is linked to the political cycle. It drives big, up-front programmes which ask us to describe exactly what we’re going to do and what the benefits will be, well before we do it. For well understood processes, such as house building, that could be a perfectly reasonable ask. However if we are doing something new, for example attempting to create behaviour change, the truth of it is we generally have very little idea how we’ll do it until we start to test in the real world. Statistical modelling can be a valuable predictive tool in many situations but accounting for how a a new service will change human behaviour is not possible to predict with statistics alone.

None of this negates the need to justify how taxpayer’s money is being spent. How can we fund teams of people working together to solve the most valuable problems than writing benefit cases before we’ve run a single test?

We believe

It’s useful to thinking of governance as a service that supports and challenges teams to make choices that are appropriate in their context
We need governance that both allows us to ‘see the thing’ for real and consistently and effectively manages the complexity and cost of services
The hallmarks of good governance are inclusivity, transparency, collaboration, regularity and responsiveness
Governance isn’t higher than delivery

Good governance accounts for and is carefully adapted for its context. It doesn’t deny advances in favour of maintaining the status quo; it evolves with technological advances and new ways of working. It allows teams to operate in ways that work for them and supports, indeed insists, on the regular release of value and short feedback loops. Only by putting in place governance that meets this can we be responsive to user needs.

We’re already seeing some of the barriers above being broken down as we’ve organised to respond to COVID-19. We’re learning more about each other, breaking down departmental silos and organising to respond quickly to whatever might come next. Despite the inevitable difficulties that lie ahead, we’re hopeful that this can make a lasting improvement to how we work together, deliver and govern work.

Blog posts in this series: