Late last week we borrowed from Ben Evan’s three classes of search/discovery to look into what the future holds for OSS.
- There is giving you what you already know you want (Amazon, Google)
- There is working out what you want (Amazon and Google’s aspiration)
- And then there is suggesting what you might want (Heywood Hill).
Benedict Evans here.
In the last in this series, we look past predictive analytics in OSS to:
CLASS 3 (Precognitive OSS)
Whereas predictive analytics use big historical data sets to identify patterns that repeat (eg degradations in network health or customer experiences likely to lead to churn), precognitive OSS will have the ability to analyse available data and predict events that have no direct precedent.
What about these:
- From aptitudes identified during activities performed previously, what opportunities and training should be given to an operator (ie taking skills-based routing from a historical perspective to a looking-into-the-future view)
- Given that faults and follow-up remediation activities have always been done a certain way, what variants of processes / activities could be tested to try to identify a better way? If that test does lead to greater efficiency, then trickle that into the future mode of processing
- Rather than building a root-cause remediation engine based on historical event-streams, understanding a combination of network configurations, performance metrics and network health to identify the early formation of a fault that hasn’t occurred previously, then identifying a set of actions to resolve it before it becomes service effecting
- Having a decision support tool that can intervene into operator actions that may have unexpectedly detrimental effects on the network that is about to undergo a change (eg an outage to part of the network due to planned [or unplanned] events)
- Using road traffic data and other changing conditions (eg weather, missed jobs, etc) to predict (in near real-time) the best allocation of field staff (ie redirect better-placed, but suitably resourced technicians to a job to cope with ever-changing field conditions)
- Using outside data sources (eg social media), not just internal sources to work out what an operator might want to know next (eg predictive tools have identified a customer is about to churn, but precognitive tools can which one of their churn-response triggers is most likely to be effective in response)
- When planning network expansions or augmentations, consider not just the catchment areas for potential new customers, but also their predilection for certain types of services