Making Workflow Work
By Matt Morgan, Head of Process and Metrics Excellence, Bridgewater Associates and Darren Wright, Consulting Principal – Team Leader, Leonardo Consulting
It is our experience that simply putting existing processes into a workflow system is unlikely to produce significant benefit. Realising the potential of workflow systems means paying attention to the holistic management of a company’s business processes through a structured management approach before applying technology. Organizations must understand and improve their processes if they are to make workflow work.
Matt has written elsewhere of the approach he has taken at a multi-billion dollar financial services company to building a transformative BPM (Business Process Management) capability, summarised as Define-Measure-Improve-Systemize. In this approach, processes are understood, process performance measurements created and tracked, and optimizations made on the basis of the evidence before systemizing the process in a workflow or iBPMS system.
The Define-Measure-Improve-Systemize framework enables companies to take advantage of workflow in the following ways:
• Defining the process and understanding its contribution to customer value means that the most important steps can be automated to reap maximum benefits rather than picking automation targets at random
• Once defined and improved, iBPMS or workflow systems “bake in” the process, improving control and consistency, and reducing unnecessary variation
• Telemetry is built-in to iBPMS systems, meaning that once the measurements that matter have been defined, collecting and analyzing data becomes a matter of reporting from a system rather than extensive from-scratch project
• Codifying business rules as part of the definition and improvement of the process means that iBPMS systems can automate decision points, reducing effort, hand-offs, and error.
This matches Leonardo Consulting’s experience in implementing workflow systems. We see that successful workflow implementations occur only when companies understand the end-to-end pathways that deliver value to their customers.
Organizations must understand and improve their processes if they are to make workflow work
Such advice seems obvious, but modern workflow platforms can tempt companies to build islands of automation. The result is often a disappointment, as the newly automated processes simply accelerate existing flaws and disconnects, and savings get eroded as manual work simply gets pushed down the line and customer experience suffers.
On the basis of this approach, we can offer some examples from the field.
Design workflow with the end-to-end in mind
A strong concept of the end-to-end flow as documented in a high-level process architecture matters, even if not all parts of the process are automated. This helps ensure the customer experience is improved and hand-offs between teams and systems minimized.
A positive example is a large petroleum distribution company that used a change of workflow technology as an opportunity to re-engineer its fuel load process. As a result, an automated fuel load notification can signal a driver to collect fuel, automatically generate paperwork and record load volumes and end events. Despite the underlying technical complexity (over 4 integration and 15 interfaces), the overall process is seamless.
Measurement is key
Knowing what you’re automating does not just know how the process fits together, it also knows what to measure to make sure you’re truly improving the process. As an example, a major civil infrastructure company re-engineered its works management process to use a sophisticated workflow approach. A central part of the workflow is to provide detailed analytics on how each part of the process is performing, so managers can readily identify bottlenecks in real time and apply corrective actions. In this case, the workflow system provides the telemetry for a well-understood and well-defined process.
Improve before and while you automate
Looking at the effectiveness of manual processes, getting to root causes of any issues and designing improvements is critical to getting the best value out of workflow systems. In many cases, this is not simply taking existing forms and putting them into iBPMS screens.
As an example, a large southern hemisphere insurer currently processes around 2 million claims per year. Prior to automation, all of these claims were assessed by staff, occupying a team of up to 70. The insurer commissioned a statistical study of its claims and realised that a great majority of claims with a particular density of certain keywords were much more likely to be rejected, where others were highly likely to be granted. Instead of building a workflow solution that replicated the existing process of claim submission > read claim > make decision > have decision reviewed by experts, it fundamentally re-engineered the process. Now, a rules-based iBPMS solution analyses the claim language, applies the statistical model for rejection, and workflows those claims likely to be rejected for human review while tagging the others for approval.
Modern iBPMS platforms have the potential to radically improve business performance. Pay attention to the holistic process through a disciplined process management approach, and that potential can be realised.