In my previous few posts, I shared the discovery of the Fractal (Repeating Pattern) Phenomenon and the development of Systemic (Pattern) Thinking in my Drim (Dramatic Improvement) Journey.
In the next few posts, I’ll show you some Goal|Problem|Solution (GPS) Patterns and an example Systemic GPS Exercise, before resuming my Drim Journey story.
This post presents the GPS pattern for complex decision-making within an insurance claims context – but key elements of it apply to any complex decision-making process.
Brief situation description
My client was the manager of a Claims department in a personal (life and health) insurance company. There was a big backlog of claims. Even without the backlog, claim resolution times were long.
With 55,000 medical conditions and a nearly 250 page Claims Manual, it took a new claims specialist a number of months to gain the expertise necessary to assess a claim reliably, on his or her own.
The cost of claims mistakes are high:
- Deny a claim that should have been paid and the consequences are dire, as there is a free insurance industry ombudsman complaints service and the penalties are big.
- Pay a claim that should have been denied and the unnecessary payments could run to many, many thousands of dollars.
Natural and vocational claims people have a higher than normal level of empathy for people so, although some legitimate claims were probably being denied, it was much more common for claims that should have been denied to be approved, costing the company (and ultimately all customers) significantly.
Another concern was the unnecessary stress on customers caused by the delay and uncertainty – as well as the risk of medical conditions deteriorating in the absence of early intervention.
The GPS (Goal|Problem|Solution) Pattern
We started with a simpler GPS, but here’s what ended up with:
- Goal
Pay deserving claims out – and deny undeserving claims – quickly, efficiently and painlessly. - Problem
It’s impossible to know, ahead of time, what information will be required to assess a claim reliably. - Solution
Use reverse-engineered asymmetrical decision-trees and profile-matching tables to deny, approve and refer claims to seniors and the Claims Committee.
My client liked the early version of this GPS (which only identified the decision-tree element of the solution), but his management team were skeptical, thinking that the decision-tree would be so big that it wouldn’t fit on one whole wall of their open-plan office space.
Unpacking the GPS
The first breakthrough: The 80:20 Pattern
I worked with a senior claims specialist and within a few days we’d realised that nearly two-thirds of the claims were straightforward and did not require medical tests – and a similar proportion of the remainder required a combination of only a small number of standard medical tests.
We’re all familiar with the 80:20 concept – but are often blind to how it applies in our reality.
In this situation, we realised that improving junior claims specialist decision-making was sufficient to deal with the backlog and ongoing workload. We could progress to senior claims specialists later.
The second breakthrough: The Reverse-Engineering Pattern
Within a further few sessions we’d whittled the primary decision-tree (identifying the 62% that didn’t need any tests for a decision) to 22 questions, by realising that there are only three possible end-states for a junior claims specialist:
- Accept the claim.
- Decline the claim.
- Refer the claim to a more expert claims specialist, who will either
- Accept the claim.
- Decline the claim.
- Require medical tests or
- Refer the case to and even more accomplished claims specialist or the Claims Committee.
Reverse-Engineering – working backwards from the end-state we’re seeking – is a very powerful breakthrough pattern and applies to nearly everything: yet we seldom remember to apply it to our world.
Working backwards from end-states, by asking questions like, “In what circumstances would we accept a claim?” made it way quicker and easier than working forwards through 250 pages of practice manual.
The third breakthrough: The Asymmetry Pattern
As we enhanced and simplified the decision-tree (enhancing the GPS as we went) it dawned on us that the decision wasn’t symmetrical – in fact, it was very asymmetrical. It doesn’t matter how many reasons there are for accepting a claim: even one reason for declining it (like premiums not being up to date; excluded conditions and; insufficient wait-time having passed) results in a decline decision.
This pattern of asymmetry applies to a whole raft of reality and dramatically reduces workload, cognitive load and wasted time and energy. Our natural assumption is that everything is symmetrical – when it very often isn’t. (Interestingly, we also have a tendency to assume asymmetry when things really are symmetrical: but that’s another subject!)
Asymmetry enabled us to focus on the “Decline Claim” end-state and the final result was five simple questions (because there are five reasons for an automatic decline).
This made the claims decision very simple, in hindsight, and resulted in a decision-tree diagram which could comfortably fit on an A5 sheet of paper.
‘Not bad, when compared with an entire office wall!
Implementation
The next challenge we faced – once my client had satisfied himself that the decision-tree was robust and watertight – was convincing the senior claims specialists.
(It’s common, in expert functions which rely heavily on domain-specific knowledge and experience, for the smartest, most loyal and committed of experts to resist simple breakthrough solutions that they don’t come up with themselves. So we knew we had our work cut out for us.)
The GPS contained the solution to this problem as well, of course: we needed to develop a simple, reverse-engineered asymmetrical decision-tree to bring the senior claims specialists to adopt and enhance the Claims Decision Tree.
My client had the brainwave of setting the team a 20-minute challenge to break the decision-tree: prove that it was unreliable, by finding examples, from their experience, that it would fail on.
After 40 minutes of suggesting more and more challenging corner-cases – without a single failure – the team agreed that it was robust enough to use (with a level of supervision at first, followed by a period of decision-approval and then an auditing period).
We printed out the A5 sheets and gave everyone basic training on how to use them – and built a simple Decision Process in Excel that generated scripts based on the answers to each of the five questions. It was so simple and straightforward that novices could use it right away.
Impact
- Claims backlog was totally removed within a few weeks.
- The three-week average claims decision-process for the 62% straightforward cases was reduced to 10 minutes on the phone.
- Instead of it taking months for new junior claims specialists to be able to make reliable decisions on their own, they could do so in their first week.
- The junior claims specialists loved being able to actually help people directly and immediately – and loved knowing exactly what they were doing and being confident that they would get it right, first-time.
- The throughput and satisfaction of expert claims specialists improved dramatically as well, because of the almost 100% reduction in the need for them to be involved in the straightforward cases.
- Customer complaints stopped almost completely – and unsolicited compliments grew dramatically.
- We didn’t get actual data on this, but knew that for a some health conditions, earlier claim resolution results in earlier, quicker, less expensive, painful and more certain and effective recovery.
Further application
Although we knew that we could apply the same concept to the remaining two thirds of the more complex claims decisions, my client was promoted to a more senior role in another part of the business and I did further work with him in his new role. His successor couldn’t grasp the claims solution and gradually abandoned it over the next few months, and performance and service levels quickly dropped to previous levels.
My client also introduced me to the underwriting side of the business – and we found that, with minor enhancements (like the addition of profile matching tables) the GPS applied to Underwriting as well. But that’s another story.
We only realised later that profile-matching tables could be applied to the next layer of Claims expertise (and the Claims Committee) as well – and make the bulk of that workload accessible to more junior claims specialists, in time. ‘Didn’t get to prove that.
We adapted and applied the Claims GPS pattern to other functions within insurance, however – and other financial sector functions and their clients and customers – over the years, and realised that nearly all of it applies to all forms of decision-making, risk protection and investment.
What’s next
As I said, before, I’ll provide another couple of GPS Pattern examples and then take you through a pattern-thinking exercise end-to-end, before resuming the saga of my Drim Journey. ‘Can’t wait to share some of my recent discoveries and developments!
You must log in to post a comment.