Telecom Process Business Intelligence – Doing Things Right? Or the Right Things?
by John Myers
Originally published January 13, 2009
Whenever I hear about concepts like Generally Accepted Accounting Principles (GAAP) and International Organization for Standardization (ISO) 9000 , I always think about the famous, or infamous, Peter Drucker quote:
GAAP and ISO 9000 focus on a standards-based approach to accounting and quality standards. To a certain extent (and by intent), they take the human element out of a particular activity and focus more on following the “right” process than on doing what is “right.” For certain activities, this is a good thing. GAAP accounting has proven itself to be a pretty reliable concept that can be used well or poorly. Likewise, ISO 9000 processes can be reliable in certain situations, but can lead to unintended consequences (“bob, why are we making lead life vests?” “because it says to do so in the ISO book…”).
Everything is a “Nail” to the “Process Hammer”
Despite any potential downsides, there has been an increased focus on removing the human element from business processes via best practices and process automation. This is due to the fact that most humans may, and often do, have the opportunity to make an occasional mistake. “Swivel chair” integration is probably the best example of this concept. The opportunity for someone to “fat finger” a value being passed from one application to another can be the bane of a business process reliant on accurate data (and i'm not aware of automated processes that run well on inaccurate data…).
In these cases, I wholeheartedly endorse the concept of removing as much of the human element as possible from the process execution. However, as these processes become more complex, there are decision points that need to be included in these processes:
As the volume of events/activities associated with these processes increases, it becomes harder and harder for humans to make those decisions; and there are pressures, just as with “integrating” the process, to remove the human element or standardize the decision points of the process. This is where analytical models and predictive analytics can provide great value in terms of the concept of operational business intelligence and process efficiency. But we need to be mindful of how those analytical models are applied.
Following the Fall Line
Many of these automated decisions, from analytical models and predictive analytics, are the result of the application of best practices. Here is an interesting definition of best practices:
Fortunately or unfortunately, most business processes suffer from the concept of “doing things right” or taking the path of least resistance. In skiing, there is the concept of the “fall line” or the most direct route down the mountain. This is the path that water will take. In business processes, managers will apply best practices “simply because” they are best practices and thus be “absolved” of making decisions about the application.
Often, these are predictive models designed for other applications or business models, and they provide less than ideal results when applied to a different situation. I would recommend that in processes, in new processes in particular, a human analyst be employed to take action at the decision points within a given process. In this regard, the ability to empower employees with that responsibility and the insight that they have as it relates to the many variables associated with a decision are important.
In my opinion, Albert Einstein put it best:
Automated decisions, from analytical models to predictive analytics, are not magic. They are based on the information they have – whether it is complete or incomplete, accurate or inaccurate. That is why it is so important – particularly at the early stages of process definition, implementation or reengineering – that there be a certain level of human analysis and associated result sets to make them effective and impactful on the business. Tim Wormus, Analytics Evangelist at TIBCO, Spotfire, recently said the following:
In telecommunications, there are several ways to apply these concepts. One way is to allow customer service representatives to have some leeway in handling both customer service and technical support situations. Rather than just following a decision tree-based script, they should understand exactly how many options a customer should be entitled to based on history and account value. While this is something that a robust analytics application can provide in time, senior customer service representatives should contribute to the development of business rules and result paths similar to the recommendation(s) of Wormus.
In addition, the inter-workings of the telecommunications organization can be aided by allowing for human intervention. While optimization models can provide “optimal” solutions for rote optimization decisions, there is often an environmental component to network architectures that only a network operations expert can “feel” and thus provide adjustments to specific rote definitions. Again, in time, the optimization model can support the environmental components that make a difference between a textbook solution and a real-world one. However, until those models can be updated to include all of the variables associated with the real-world solution, network operation team members should be included and their input valued to make the models better and to develop buy in with a system that may potentially take away some of the lower-level aspects of their jobs.
In summary, I believe that technological advances, associated with automated processes and the ability to manage those processes, have become great productivity boosters in telecommunications. The ability to manage billions of events per day and petabytes of information would not be possible without those enhancements.
However, we should be mindful when attempting to manage the decision points of these processes. The words of John F. Kennedy may inspire those responsible for those processes:
Tasks should not be undertaken because they are easy or efficient. They should be undertaken because they are the right tasks to perform and the right direction to take.
Note: For more information on this topic, check out this recent BeyePERSPECTIVE roundtable
discussion on process intelligence.
Recent articles by John Myers
Copyright 2004 — 2019. Powell Media, LLC. All rights reserved.
BeyeNETWORK™ is a trademark of Powell Media, LLC