Design principles for good procurement
Since the Digital Services 2 procurement process began, I’ve been trying to get some idea of the way a new Digital Services framework should work.
Throughout the debate about DS2, one of the recurring themes for me has been that the framework, as designed, doesn’t follow the principles that GDS had laid out for effective governance. At root, procurement processes are really there as a form of governance over purchasing.
So why not follow principles in this context that work in others?
Don’t slow down delivery
Procurement needs to be quick. A process to make sure procurement happens legally is necessary, but it needs to be minimal. A large part of good governance is about removing blockers to progress, and procurement is a biggie.
We need procurement frameworks that are designed to work quickly.
Decisions when they’re needed, at the right level
The right people to make decisions about purchases are the people who are delivering. Too often, procurement teams have too much influence in the process: insisting, for example, that buyers follow LETF instead of MEAT when evaluating responses, or requiring reams of documentation to be produced that add little value.
Procurement should be simple enough that delivery teams can do it themselves, with minimal assistance. If the process is too complicated for delivery teams to follow, then it’s too complicated to be fit for purpose.
We need to empower delivery teams to make decisions about who they work with and how they select them.
Do it with the right people
By and large, talented teams deliver good work, no matter the process that they use. And no matter how good the process, a poor team won’t succeed. We need to recognise that unless buyers know what they’re doing, no procurement process will lead to good outcomes.
With that in mind: there’s no point in designing a procurement process to protect under-skilled buyers from themselves. So we have to design a framework for intelligent buyers that assumes projects are being delivered by the right people. And to the extent that that’s not the case, we need a parallel mechanism for buyers who need help to get it.
We need to recognise that not all teams have the people they need to have in order to know how to procure the right thing, and that collaboration, coaching and advice need to be facilitated for those who need it.
Go see for yourself
Throughout the life of this framework, it’s vital that we measure how effectively user needs are being met. We need to measure other things too (like spend) but meeting user needs is the first and most important thing. And it’s very hard to do this without actually looking at what is being done.
The only way we can meaningfully measure the effectiveness of this framework – or take action when things are done badly – is to understand life at the coal face. We need a government team visiting delivery teams and writing up independent case studies. We need suppliers to be more vocal about the work they’re doing and a culture of openness about costs and outcomes. We need buyers to share their successes and failures.
We need the framework team to meaningfully understand the experience of the buyers and suppliers who are actually delivering, and to judge the success of the framework by measurable standards of user satisfaction.
Only do it if it adds value
Too much procurement practice, as currently constituted, is designed to avoid legal and commercial risk. Some of this is sensible and necessary, but one of GDS’s founding ideas is that government can’t outsource risk. But still, too much of the process consists of futile attempts to reduce the risk of failure and legal challenge.
We need procurement to start with a minimum viable process. Anything that we add after that should be adding value by genuinely helping to ensure that projects meet user needs.
Trust and verify
GDS have set out a compelling model for the delivery of a new breed of digital services, and they have a reasonable interest in ensuring that that model – broadly speaking – is followed. But there’s no way we can do that with heavy-weight up-front process. If government wants to use procurement as a policy lever to influence delivery, this needs to be upended: less process at the start, and more hands-on attention during delivery.
The current selection process on Digital Services is almost entirely up-front: it requires suppliers to demonstrate how they implement the activities the service manual describes. But this can’t work. Ultimately, the way a selection questionnaire is completed won’t have much or any bearing on the way projects are actually delivered.
Anyone who’s been unlucky enough to have to contend with rubbish procurements before will tell the same story: everyone treats procurement as the strange game that it really is. You write what is required. You follow the process necessary to things done, and when you’re there, you get on with doing the work. If you have integrity, you do your best to do the right thing. It couldn’t matter less that those things aren’t always what got talked about during procurement.
We need to set sensible policy, trust suppliers to offer the right services through frameworks, and then verify that they do. Just as we need people on the ground to help buyers to procure the right things, we need people on the ground to find companies selling inappropriate services and remove them.
Up to now, I don’t think I’ve ever been at an event where delivery teams, procurement and suppliers have come together and discussed all this as a group. I think that should change. We should try it, at least once: a procurement workshop for agile folk trying to get things done. If you’re up for it, pop me an email or a tweet. If there’s demand, I’ll make a plan.