public digitalThe public digital logo

Relational contracting: test and learn across organisational boundaries

This article is included in our book, Adopting the practice of test and learn. It was written by Liam Sloan.

You can download a PDF version of this book, or visit this page to request a printed copy.

Since 2024, Public Digital and the Government Outcomes Lab at the University of Oxford’s Blavatnik School of Government have been working with both UK public sector commissioners and service providers from the charitable and private sector to explore how a more relational approach to public sector contracting can better support iterative service delivery.

Over the last 25 years, continuous delivery with fast feedback loops and rapid iteration has become mainstream. Digital products as diverse as Duolingo to the NHS App change daily to better meet user needs.

Today, test and learn is also being used within deeply interpersonal human services. Public Digital supports multidisciplinary teams which use continuous experimentation and iteration to improve healthcare, social care and employment provision.

But the state also contracts out a large number of these types of public services to third party organisations to deliver. Here, the professional disciplines involved in procuring and contracting outside providers create particularly high barriers to adopting test and learn.

The perils of contracting out

At its best, contracting out important areas of public service provision to charities or businesses can improve quality and responsiveness, introduce innovation, and improve efficiency.

At its worst, high profile scandals and provider failures have brought many to question whether the state should rely on outsourcing to deliver public services at all.

Both stereotypes have some truth to them. In my experience working in this arena over the last 15 years, the difference between success and failure often comes down to the relationship between commissioners and providers. Are both organisations aligned to achieve the same outcomes? How openly are they working with each other? How are they learning together? How are they able to adapt service provision to respond to that learning?

In other words: does the contracting approach encourage test and learn across organisational boundaries, or is it the worst of ‘waterfall’, assuming that a detailed specification defined upfront will be right and that any deviation from it is a breach of what has been agreed?

The more you look at this problem, the more you see that the norms of buying in the public sector - procurement, legal and contract management processes - are stacked against adopting outcomes-focussed, relational, test and learn ways of working. And often this isn’t about what the rules and regulations actually allow, but about the way they are interpreted and applied by corporate functions that remain stubbornly disconnected from actual service delivery.

What’s really the biggest risk?

In its essence, a contract is a mutual agreement describing how a relationship will work, what happens if something changes, and how risks are managed.

Traditional transactional legal approaches to contracting involve trying to nail down exactly how a service will be delivered, replete with KPIs and a payment mechanism to enforce this, and attempting to name every potential risk upfront and what will happen if it occurs. The academic name for this is “presentiation”: an attempt to bring possible future occurrences into the present through a legal document.

This approach assumes we know upfront what will work. In this world, the biggest risks are perceived to be providers trying to cheat the system or take advantage of public bodies, or proper processes not being followed.

But anyone with experience of working on frontline public service delivery knows that attempting to achieve perfect foresight is futile. Whether it's the economy, service user demographics, the context and environment for the service, or simple human behaviour - everything will change to frustrate the best laid plans.

That means that a well-run service should constantly test, learn and adapt to continue to achieve its outcome. You must optimise for learning and adaptation, not efficiency.

In reality, the biggest risk in contracting out is that a service simply fails to achieve the desired outcome because it has been set up wrong and can’t adapt to change, resulting in wasted public money.

Collaborating in hushed tones

When contracting out service provision, the only way to unlock adaptive approaches is through a truly relational partnership between commissioners and providers based on learning and collaborating together to achieve shared outcomes. This requires working openly, putting in the effort to build trust, and expecting to adapt when assumptions are proven wrong or conditions change.

We know from our research that some commissioners try to work this way with their providers, but have to do so in “hushed tones,” working informally outside of what the contract actually says or out of sight of procurement colleagues.

The irony is that the rules do allow you to work in this way intentionally. Dialogue and co-design are explicitly championed in the UK’s new procurement regulations, and formal procedures have existed for years to collaboratively innovate and co-create new services together iteratively.

The barrier to this way of working is usually a lack of knowledge, ingrained professional practices, and aversion to the risk of doing things differently.

At the same time, simply asking public bodies and providers to “be more relational” feels reductive. There remain important considerations about fair procurement processes, safeguarding public money, and protecting against corruption which must be held onto.

How to work relationally, formally

Resolving these tensions requires a way of working which is both explicitly relational, but formally codified in a contract; one which recognises the anthropological yin of contracts alongside their lawyerly yang.

A commitment to dialogue, openness and learning needs to run through the process: from shaping a service, through procurement and contract management. But at its heart should be a collaboratively co-designed charter, describing how the partners will work together, to sit within or alongside the formal legal contract.

Many multidisciplinary teams already have such a charter in place. It becomes even more important when a team stretches across multiple organisations.

This should be a formal (but not necessarily legally binding) document acting as a constitution for the work - less concerned with specifying precise deliverables, and more concerned with structuring how partners work together. It should be written to be accessible: short enough to be pinned up on the wall so that everyone can see it and memorise it. It should specify:

  • What is the shared outcome we are all working towards?

  • How are we going to work together? What shared principles will underpin this project?

  • What governance processes will we put in place so that, as partners, we can hold each other to account for working like this?

The cultural acts of co-creating such a charter, of centering it in the service’s day-to-day identity, and referring to it repeatedly to remind everyone of their obligations, are powerful social mechanisms to build and nurture relational working. They emphasise the importance of mutual agreement and shared accountability for success. Both principles (such as openness, sharing learning, expecting to pivot, and tolerating failure in the learning process) and processes (such as regular retrospectives, open book accounting, data-sharing and joint reporting) can be brought into the light and championed as the fundamental behaviours that underpin successful service delivery.

This will never replace the legal clauses needed in a broader public sector contract that provide strict performance levers and safeguards, such as defining liabilities if things go wrong, and specifying terms around Intellectual Property. But alongside these, the charter will not only hold partners accountable for outputs and outcomes, but also how they collaborate to pursue them.

An approach like this could be transformational in improving people’s understanding of what good contracted service delivery should feel like. We are about to start working with commissioners across the public sector to test this approach and create online resources to help put this into practice. You can follow our progress here: BARCODE.

With thanks to Connie van Zanten and Cate McLaurin (Public Digital) and Dr Felix-Anselm van Lier and Michael Gibson (Government Outcomes Lab, Blavatnik School of Government, University of Oxford)

Previous:

Next:

About Liam Sloan

Liam is Public Digital’s Commercial Director, and has over fifteen years of experience in business development and commercial roles across public sector outsourcing, healthcare and education. He is a Visiting Fellow of Practice at the Blavatnik School of Government at the University of Oxford, bringing together research and practice around relational approaches in public sector contracting at the Government Outcomes Lab.