Faced with the challenge of reducing business costs by 30% and saving 90 million Danish Kroner, the Danish State Administration chose F2 as the platform to support their ambitious business transformation project.
The State Administration has over 200,000 cases a year, dealing with family law (adoptions, divorces, child custody) and a number of other complex matters. At any one time there are more than 30,000 open cases.
Services such as divorce are chargeable to citizens, so there is a constant pressure to provide value for money.
By taking a checklist based approach and adding self-service, the State Administration were able to take out more than 50% of the cost for administrating divorces.
Doing the hard work to make things simple
At cBrain we’ve been spending some time looking at how to support mission critical processes from a standard platform.
The ambition was to provide something that directly meets what users need, but without locking people into fixed processes and creating a new legacy system.
For this to happen we had to start with a big assumption: our users are smart enough to know what they need to get done. Our job is just to make it as easy as possible.
By working with users we came to two realizations.
Swimlanes don’t work
The first was that traditional process modelling using swimlane diagrams didn’t really help. People found it very hard to get anything other than a superficial understanding of the process. Most people looking at a swimlane would say it looks more or less OK. But when walking through actual cases they quickly find variants and conditions and lots of reasons why the swim lane doesn’t represent what actually happens in practice.
And if we don’t support what people actually need to do then we are short-changing users.
Only automate what should be automated
The second realization was that we shouldn’t try to automate tasks where dealing person to person is more beneficial. While some government transactions are purely data driven, many are dealing with complex personal circumstances, different motivations and social pressures. Often speaking with a person is the best way to clarify what the real need is.
So we needed to find a way to automate the trivial work while at the same time supporting tasks where the knowledge and experience of practitioners is critical to getting the right results for the citizen.
We also wanted to look at the data around outcomes: people make mistakes, but we wanted to understand when and why they did so.
The Checklist Effect
This is where checklists come in. There is a growing body of evidence that shows checklists can have impressive results in getting the right outcomes for service users.
“Checklists allow complex pathways of care to function with high reliability by giving users the opportunity to pause and take stock of their actions before proceeding to the next step.” World Health Organization (WHO)
Led by Professor Atul Gawande, the WHO introduced a Surgical Safety Checklist which had a significant effect in reducing patient mortality during surgery. Professor Gawande in turn was inspired by the checklists used by airline technicians to improve aircraft safety.
And it’s not just for surgeons. In one of our early implementations when doing user research at Ministry of Interior we got the comment: “If you really want to know how we work, we just use this” said the user, and pulled a checklist out from their drawer.
Putting it into standard software
And that user comment led to our checklist approach to process modelling being born.
Using lists as a way of organizing work and ensuring quality has the advantage of being easily understandable, and a neat way of simplifying complex actions. Even when intelligent people know what to do they can help ensure the right outcome by introducing moments of review to ensure that common causes of error are addressed.
And putting a checklist into software instead of on paper allows organizations to build a dataset showing how the organization actually works.
We decided to build a process editor on top of the core F2 platform that would allow users to add tasks and structure them in phases. We could create different sets of tasks for different processes, grouping them in a ‘Task Guide’
It quickly became apparent that for mission specific tasks you would need to extend the data model with additional data, so we provided a way to specify the extra fields required for each case type.
Added to that we introduced forms to capture the additional data - either by self-service, integration or internal user input.
Simpler, Cheaper, Faster
Using the checklist approach, and introducing self-service for the divorce process the State Administration was able to reduce internal administration costs by approximately 50%, which in turn led to the charge for citizens being reduced from 900 Danish Kroner to 420 Kroner.
If both parties agree, then divorces can happen quickly and effectively. However if there are disagreements the checklist allows for variances and complex cases where there are children involved, financial disagreements, and where face to face discussion or legal representation is required.
This means that complex processes that require administrator involvement could be supported alongside simple processes that can largely be automated - in one instance a process that normally took 5 days just to receive an acknowledgement of request has been shortened to taking only 5 minutes end to end with no back-office intervention.
Interestingly the question now for the State Administration is whether they have become too efficient, and whether they should introduce a pause for reflection in the divorce process.
Which is not something you can say about many government organizations…