“Key Performance Indicators” (KPIs) is a term commonly heard within the Service Assurance domain. KPIs are used by most Service Providers use to align their internal departmental goals and metrics with the experience delivered to their customers or subscriber base.
That said, there is much debate as to how effective KPIs are, even those that are published by various standards and regulatory bodies. Moreover, it’s hard to judge how efficient such an approach is and whether or not the results are actually representative of the experience the majority of the customer base receives.
Take, for example, a simple KPI such as Post Dial Delay (PDD). PDD represents the time it takes for someone to dial a number and then hear either a ringing or busy tone.
The usual method for a Service Provider to measure PDD is to deploy a KPI with a predetermined set of performance thresholds, based on time or duration. The problem with such an approach is that it only takes one or two instances where such thresholds are breached to generate subsequent alarms or alerts. And in some cases, each breach is actually associated with a small subset of subscribers.
A more effective way of implementing such a KPI is to set a conformance target.
Using the same PDD example, Service Providers can set a performance target of, say, one second, for all subscribers, all calls. But in terms of the thresholds set, they can base this on the number of subscribers who experience such a delay. For 95% of all calls, for all subscribers, if the PDD is one second or less, this is OK. The Service Provider is only alerted if the conformance (%) falls below this 95% target.
The net effect is that implementing a set of KPIs becomes more representative of the experience the majority of the customer base receives. Additionally, it results in fewer alarms or alerts that need to be handled by Network Operations or Engineering.