Casting out the devil’s advocate

3 years ago, when I was a design lead at NHS Digital, I was asked to go to London for a meeting about Wuhan flu.

The meeting was with some NHS colleagues, and people from the Department for Work and Pensions, and the Department for Health and Social Care. There was also a doctor there, who gave us quite a chilling vision of what was going to happen later in 2020. People would need to stay at home. There would be a COVID wave in the spring, followed by a second wave in the winter. I learned new terms, such as ‘contact tracing’. It was surreal, because after that, the rest of the meeting was about designing a service, so we had to break out the sticky notes, and kind of act like normal.

This meeting would result in the swift launch of Get an isolation note — the first service launched in response to the looming COVID–19 pandemic.

The intent was to allow people in full-time employment to stay at home for 14 days, without contacting their GP for a “fit note” (which you need for more than 7 days off sick). This was to achieve two goals: to stop infectious people going out in public, and to stop GPs being overwhelmed by requests for fit notes.

If you had symptoms, you needed to isolate. It would be hard to prove you had symptoms in any ‘official’ way. But we still needed to ensure people could isolate, and give something to their employers to say so. We designed a simple service whereby, if you declared certain things (symptoms, the date they started, etc), the service would email you a PDF declaring you had to isolate.

(Bear in mind this was all happening in early March, before the country went into lockdown, and way before tests were available to the public).

What if someone games the system?

There was some anxiety that the system made it too easy to get a note. I remember in one meeting, someone, playing devil’s advocate, said “why don’t we just put a button on the homepage, saying click here to get 14 free days’ holiday?”

While he had a point, it also demonstrates how premature worrying about abuse of a system or service, can restrict the development of something that helps people in need.

Every system has inefficiencies, and we have a tendency to dedicate inordinate effort into eliminating waste or abuse, at the expense of pursuing the system’s intent. Doing this almost always makes the experience worse for the intended user.

In the case of Get an isolation note, there was no real way around it. Putting any kind of friction or checks into the service would increase the risk of people who were experiencing symptoms, of going back to work. Worrying about abuse or inefficiency would make the service worse. And yet there are devil’s advocates at every turn.

Optimism ≠ naivety

When designing a service, you need to approach it with optimism. You need to think about what is the best way to solve a problem for someone. Optimism is not the same as naivety. There is a time for discussions around the practicalities of delivering a service to the right people, but premature cynicism can stifle a good outcome.

Take disabled parking spaces in supermarkets. Definitely a provision that makes things easier for disabled people to go about their day. However, I’m sure at some point, everyone’s seen some selfish, self-entitled person in a hurry, park their BMW in such a space because it’s just easier for them.

So it’s an inefficient service. Some people clearly abuse it from time to time. But by doing the work to eliminate that abuse, what would you need to do? Introduce checks? Demand a blue badge? Have security guards? Cameras that scan your number plate, and check it’s a car registered to a disabled driver?

Not only would these things be expensive, they would also make the provision of these spaces significantly worse for the people who need them. What if a disabled person had forgotten their blue badge? What if they have an invisible disability, or one that does not come with an “official” diagnosis? What if they’re not online and can’t register their car with whatever service would need to exist, for the number plate scanner to reference?

All you would be doing is introducing friction and stress for people who need it, to eliminate abuse by a much smaller number of people. But the devil’s advocate’s work would be done.

I would like to see more stemming of devil’s advocacy in the early days of service design. If you are in an early workshop for something, and someone pipes up with a “that’s all well and good, but what if someone tries to…”, find a way to defer those conversations until another time, when it is more useful to stress-test ideas for inefficiencies.

Design with optimism.