by Phyllis Drucker
Date Published September 13, 2019 - Last Updated December 17, 2019

big da·ta

noun | computing

Extremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations, especially relating to human behavior and interactions. (Oxford English Dictionary)

We live in a scary world. The concept of “big data” or extremely large databases of information about everything and our use of social media adding to such data bases has removed any sense of privacy—pushing ads for trips we’ve discussed in our living rooms; tailoring Facebook content based on Amazon searches; making it clear our politics, views, and lifestyles are publicly known and available. 1984 is here. As scary as this may be in the big picture, IT can also make use of the big data available in service management platforms to increase effectiveness in daily operations.

Getting from Knowledge to Wisdom

Anyone who has taken an ITIL® Foundation class might remember the concept of the SKMS or Service Knowledge Management System. The idea was to create a database containing information of everything IT needed at its fingertips to operate daily, like the configuration management database (CMDB) and combining operational data from incidents, problems, changes, service requests, procurement, contracts, etc., so that IT could realize “wisdom” from the data it collected. The concept was that once IT collected enough data, it could use it to make appropriate decisions and demonstrate a level of operational skill akin to wisdom.

At the time this concept was being delivered to Foundation students, the reality was that there was no programmatic way to generate this wisdom in the service management tools that were being used. The game has changed, and the tools have caught up, with several of them now providing predictive analytics capabilities built into their platforms in either (or both) of two ways:

  • Providing analytical engines administrators can tap into to create predictive capabilities for operations performed regularly in the organization
  • Using IBM Watson-style capabilities built into the tool already to deliver extended functionality when using the platform

The second option has become common and may in fact be built on a platform that enables expansion by enabling users to build their own analytics capabilities. These features may not all come easily to many service management platform users, but when they are built into the platform already, they provide an easily adopted advantage, once users know what they can provide.

Join Phyllis at Service Management World to learn how service management platforms are using predictive analytics to enhance operations.
Learn more!

An Inventory of Predictive Analytics Use in Service Management Platforms

As mentioned, predictive analytics enable IT personnel to use information stored in their tools to appear “wise.” The platforms do this by using pre-programmed algorithms to mine the data for particular purposes.

Predictive analytics enable IT personnel to use information stored in their tools to appear “wise.”
Tweet: Predictive analytics enable IT personnel to use information stored in their tools to appear “wise.” @msitsm @ThinkHDI #ITSM #servicedesk #AI

For organizations getting started using these capabilities, it’s good to understand how tool manufacturers are using them. Let’s tackle them using common ITSM practices:

  • Event Management: We all know the drill, operations departments and incident management teams will tell you that it’s difficult to build the CMDB well enough to understand the impact of an alert. What if predictive analytics could search the CMDB, event, incident, and problem records to interpret the alert and provide an intelligent description of its potential impact? It can, and faster than any human! Using artificial intelligence and knowing the affected configuration item, predictive analytics can mine previous data to provide an alert like “ci-name is down, 80% of the time when an issue is logged for this CI it means that website users won’t be able to pay for orders, interrupting their ability to purchase from our website.” Having this level of information available enables the organization to know immediately that this is an all-hands-on-deck major incident, affecting a customer-facing website. No operator had to troll through CMDB data or guess what could be impacted. More importantly, this would be known seconds after the alert is generated.
  • Incident Management: Service desks, particularly at a large organizations that might have a fair number of analysts taking calls, often see another common challenge. Imagine if, when logging an incident, an alert popped up on the incident saying “five calls have been logged with similar descriptions in the last 10 minutes. When this happens, it commonly indicates a larger systemic issue.” Again, in this scenario, time is saved on two fronts. First, the agent knows there’s no reason to try to fix the caller’s potential PC or access issue, rather there’s a more major incident brewing. Second, it identifies the major incident more quickly than standard call-processing and incident management will, again initiating the major incident and enabling restoration of service more quickly.
  • Change Management: Ever wish you knew which changes are likely to cause a customer impact during deployment? Predictive analytics can indicate likelihood of success as a “change success rate” each time a change is logged against a CI, by searching the history of all changes made and looking at how many were deployed successfully.
  • Security Operations: Like event monitoring, when a vulnerability is reported in the NIST database and found by security scanning tools, it’s tough to weed through hundreds of records to find the most critical vulnerabilities to address. Here, too, predictive analytics can be used to produce an impact rating for the organization, by understanding how the CI is used, the criticality of the services running on it, and whether data it uses could be affected. Ability to use artificial intelligence to mine enough data to rate the impact thus enables the organization to address the most critical vulnerabilities first.

These examples demonstrate the value of predictive analytics but getting started can be a challenge.

Three Things to Know Before You Start

It’s helpful to understand some of the challenges an organization will face when implementing predictive analytics and how to get started. Consider the following three areas:

  1. Your data will impact results: Don’t even think about how good life can be with predictive analytics if the CMDB isn’t built and at least mature enough for the data mined by the system to be useful. Discovery and service mapping both support decision-making by the predictive analytics application within the platform. In the change management example, change closure codes that indicate success of the change are key. They need to be descriptive enough to know where an unsuccessful change required more time, had to be rolled back, or caused an outage, for example.
  2. Success requires maintenance: As good as these systems are, they do need “training” and upkeep over time, analysis of the results being found, and tweaking to improve them, much like the use of synonym dictionaries to improve knowledge results. Plan for this.
  3. They require data: Organizations that have just implemented a new tool will not be able to leverage predictive analytics out of the gate. It takes 35,000 to 50,000 records to be successful, so it will take time before these features can be of help. It is important, however, to know that they will be used, ensuring that the required data is collected cleanly and thought is given to resolution, completion codes, etc., to aid the results of predictive analytics later on.

This is not a turn it on and go area, so anyone getting involved with artificial intelligence and predictive analytics needs to consider both the groundwork needed and challenges they might face.

Phyllis Drucker is an ITIL® certified consultant and information leader at Linium, a Ness Digital Engineering Company. Phyllis has more than 20 years of experience in the disciplines and frameworks of IT service management, as both a practitioner and consultant. She has served HDI since 1997 and itSMF USA since 2004 in a variety of capacities including speaker, writer, local group leader, board member, and operations director. Since 1997, Phyllis has helped to advance the profession of ITSM leaders and practitioners worldwide by providing her experience and insight on a wide variety of ITSM topics through presentations, whitepapers, and articles and now her new book on the service request catalog, Online Service Management: Creating a Successful Service Request Catalogue (International Best Practice). Follow Phyllis on Twitter @msitsm.
Tag(s): supportworld, service management, ITSM, metrics and measurements


More from Phyllis Drucker