A flat file extract from a sample case in the customer support system was used to generate the data for the VNA. The diagram below shows the progression of each change of status of the trouble ticket. A “status” was used to define the nodes, so a status of “review” correlates to a “reviewer” role. A “change” is represented by the link signifying a “hand off” to another status. This resulted in a role-based network view of the work process.

The accumulating status changes become more complex in each box, showing the progression of the network pattern over the 54-day life cycle of the case.

It is easy to see that a problem network pattern was developing with a long “kite tail” pattern visible by day 10, which is not a healthy collaboration pattern. This pattern continued to evolve into a pattern where two poorly connected groups were working on the same issue without resolution. Even when the alert system kicked in 22 days later at day 33 the problem continued. Without the capability to see this hidden network pattern, the problem persisted beyond day 10 - for another 44 days - until the final resolution at day 54. VNA visuals and analytics would have helped the group spot and fix the problem much earlier.

This diagram was developed for illustrative purposes by an analyst:

When customer support management saw this perspective of the customer experience, they admitted that it was a chilling experience to actually "see" how their support organization was behaving - and decided to evaluate a larger number of cases to explore whether this was typical. Fortunately this case was not typical, however what became clear during the investigation was that it was an excellent example for demonstrating the low level of knowledge sharing happening within the organization and the inefficiency of the relevant processes.

Results achieved:

The support group identified and validated network-centric combinations of KPIs relevant to monitoring and predicting the customer support experience. They identified the combination of KPIs, patterns, and thresholds best suited for both understanding why cases failed to meet SLAs and predicting which specific cases would fail to meet their SLA in the future. These new indicators in combination with the VNA visualizations could predict problematic cases up to 90% earlier than the existing escalation system.

This example is from customer support at a different company.

Building full capability for predictive analytics would require automating and integrating value network indicators with the existing escalation process to continuously improve monitoring, governance, and optimization of the workflow.

The two images below illustrate how working with network patterns can improve performance.