Below is a video and summary of a talk by Tom Loosemore at the Camp Digital conference, 2016.
Design is inherently political, and we must not ignore this
Ask yourself: who is the “we” that gets to build the future?
If you don’t understand how something works, you are a consumer, not a citizen. Don’t be fooled by ‘magic’.
Richard Pope – “Software is politics, now” – it shapes power dynamics.
GDS came up with the design principles so that people would have a new language to use to change reality.
The advantages of working in the open
Child benefit tax calculator. They made a mistake, so someone suggested a fix on github which has now been incorporated.
Ministry of Justice – problems with a form used by divorcing couples. Proprietary software. Took months to fix.
The change on github took 3 days. Massive difference.
What is digital, and what is our job in a digital world?
Definition of digital: “Applying the culture, practices, processes & technologies of the Internet-era to respond to people’s raised expectations.”
It’s not about technology – it’s about taking advantage of technology to redesign services and organisations to meet changed expectations.
Focus on delivery
Martha Lane Fox’s 4-page report gave just enough cover to start delivering. No need for a big strategy.
“The strategy is delivery” – key phrase at GDS.
Internal metric: write 100x more lines of code than lines of business cases justifying code.
Guy Moorhouse designed icons for GDS. But then he tested and found out that they didn’t help people, so he removed them and blogged about why.
Building the political case for change
GDS alpha was done openly. This was to create buzz outside the system to convince ministers that it was a good idea. This helped overcome reluctance from senior civil servants.
Do something valuable -> build political capital through an early win -> get rid of the ‘no’ people (spending all of the political capital)
Old approaches to service delivery are flawed
When Tom Loosemore started at the DWP in 2013, he asked ‘so, what have you been doing with all this time and money?’ For 3 years of work, they showed a 600 page policy design manual.
The DWP senior leadership thought of Universal Credit as a policy. But they hadn’t designed anything – they’d written a document. It had thousands of untested assumptions about people’s behavior.
“a document full of false certainty”
When Tom arrived, the DWP processes were as follows (with each step done by a different team):
- Invent policy
- Guess requirements
- Procure IT system
- Inflict on users
- Operate (aka ‘stasis’)
This is the wrong way to deliver services.
You must observe real user behaviour
People don’t know what they need. You have to observe real people in the real world
“observer their actual behavior. Surveys are useless. Actually focus groups are useless.”
“Watch what they do, don’t listen to what they say”
“False certainty if our mortal foe”
“Start humble, stay humble”
Start small, build a shared vision and empower the team
Start really small. Iterate based on how people actually use the service.
Craft a vision that everyone can use to steer every decision. Use simple language.
Empower people to make decisions based on this vision without having to run it up the hierarchy.
And because you have governance check-ins every 2 weeks through a show-and-tell (demo), things won’t go out of control.
Build an empowered multi-disciplinary team
The multidisciplinary team worked together in a room.
To enter the room, you had to be fully empowered by your bit of DWP or HMRC or LA to make decisions in the room. No one senior. It was surprising how easy it was for the organisation to identify who needed to be in the room.
Video of user testing convinced IDS to make a change to the benefits policy immediately.
Start multi-disciplinary; stay multidisciplinary.
Don’t just remove these people once you’ve ‘launched’
Obtain a mix of mindsets: Pioneers, Settlers, Town Planners.
“User research is a team sport”
Continually assess your knowledge and your readiness
Each sprint, they asked themselves: What have we proved? Do we understand user needs better? Have we designed the service to scale massively? Do we know how to operate?
“If you can’t release software every day in an emergency you’ll never be secure, because a new threat will emerge and if you can’t respond like that *clicks fingers*, your organisation is inherently insecure”
Governance
“Governance was very simple: Ministers come to the show and tell, we’ll show you what we’ve made, we’ll show you what we’ve learnt, and what we’re going to do next, and we’ll talk about risks and issues if you want. But the real governance is seeing the thing being made and seeing the evidence and user research that it’s likely to have the intent that the minister wanted. Every week. And give credit to ministers, they turned up.”
“If your senior management doesn’t show up to show and tells, look them in the eye and tell them that they are failing at governance. Use that word.”
“Show the thing” – a thing you can use, not a thing you can see.
If you’re sending screengrabs, you aren’t showing the thing, you’re showing pictures of the thing.