It’s late, very late. The office is eerily empty, and while it’s bitterly cold outside, the Operations Leader of this Claims Department really feels the heat. She did not create the problems, but after only two months in the job she is very much aware of their magnitude. Claims are taking longer and longer to process, quality problems are emerging at a faster pace, the complaints lines can barely handle the volume, and customers are leaving in droves. To make matters worse, as the pressure increases key staff members are calling in sick or resigning. How did this once proud department end up in such an appalling condition?
Well, a large part of the explanation lies in metrics. The problem is that management are flying blind. Firstly, they cannot actually tell from their systems which files are genuinely late, meaning they are beyond Service Level Targets (SLTs). This is because in order to meet targets, files are regularly ‘re-started’, meaning the clock for processing the files is reset to zero. This means the elapsed time data has no integrity whatsoever, and teams cannot effectively prioritise files to process.
Furthermore, they cannot tell from the system who is underloaded or overloaded. This slows down the rebalancing of workloads. Similarly, leaders do not know how long it takes staff to complete files, so they do not really know who is fast or slow. If they knew this they could praise, reward, and emulate the highest performers. Instead what seems to be happening is that everyone is dropping down to the lowest common denominators of behaviour and performance — after all, what’s the point in trying if the management can’t or don’t observe that you have done a good job?
Ideally management would have historical data on expected volumes and causes of volume variation, but such wishes are very much pie in the sky. Any analysis needs to be preceded by a gargantuan effort to extract relevant data from often stone-age systems. So the efforts on forecasting are proving slow, and the attempts to estimate the required staff to meet demand laughable. One of the problems is simply that the past data no longer fits the current circumstance. Let’s face it — this is a crisis.
What’s more, it is clear from inspecting only a handful of files that the input quality from the Claims Call Centre is appalling, which may be largely due to an archaic means of data capture. Yet there is no systematic way to record the errors and create a feedback loop. Indeed the only loop that seems to apply is a vicious spiral: files are late, customers are complaining, the pressure gets ramped up, people make mistakes, and the system slows down further.
Fortunately our hero has a good idea of where to start: metrics. It’s not the complete solution, but tomorrow she will switch on a simple work tracking tool. It does not interface with any core system, but it simply records what needs to be measured at a very basic level. She is capturing the volume, the ‘customer clock’ or the real time it has taken to process a file from a customer perspective, the errors identified, the work-loading on each team and staff member, and the file completion rates right down to an individual staff member. It’s far from perfect, but at least she now has a good place to start.
She will be able to initiate both short-term and medium-term actions, such as moving resources around to meet hot spots in demand, or learning from high-performing team members. Through understanding work completion rates she can start to plan for staff numbers, and by understanding errors, her team can prioritise problems for remediation.
This situation is far from isolated. Even in apparently sophisticated enterprises leaders are often provided with appalling measurement systems. If organisations do nothing else to improve performance, they should implement some tactical data gathering and metric calculation processes. That will give management some chance to act on facts and figures rather than guesstimates. Gut feel is great, but I find my gut works a lot better once I have ingested some quality data.
