IT Operations Management (ITOM)

How one Level 3 Customer Support Desk effortlessly lowered MTTR by over 90% with Ops Analytics

How one Level 3 Customer Support Desk effortlessly lowered MTTR by over 90% with Ops Analytics


Guest post by Noam Zilberman, Strategist, Head of Engineering Excellence


As a large, technologically-driven enterprise, HP is a perfect testing location for its own advanced software.

Like any modern enterprise, HP is constantly striving to make its own Concurrent Product Engineering (CPE) desk more efficient by implementing tools that allow Level 3 and 4 engineers to isolate the root cause more quickly, as well as reducing the number of tickets that are escalated to those highly valuable personnel.

Operations Analytics Pilot

In February, one of the HP Customer Support teams launched a pilot program with Operations Analytics (Ops Analytics) to evaluate how it could troubleshoot log analysis, the quality of that operations analysis and its ease of use.


The results astounded even the battle-hardened CPE team.


Coming to HP Discover 2015 Las Vegas? Register to attend:

Session B1038: “Reduce meantime to repair and improve troubleshooting with HP Operations Analytics”

June 4, 10:30 – 11:30 a.m.


Difficult-to-solve customer support problems are escalated to the L3 team. These cases are complex and may require code changes requiring expert knowledge. Typically in major incidents, the Operations support team may involve various subject matter experts — Applications, Networks, Security, Servers, Storage, Database and so on — in order to isolate the root cause, often undertaking manual analysis and correlation of data (Figure 1).


troubleshooting without Operations Analytics.png

Fig. 1: Troubleshooting without Operations Analytics


For the pilot, about 30 customer support team professionals are using Ops Analytics, supporting the BSM Platform, the Operations Agent CPE and the ALM tools team. HP Operations Analytics was tested to assess the speed of pinpointing the root cause of the problem, the accuracy of that analysis, and its ease of use.


The difference was stark: a 90+ percent reduction in MTTR. What the Customer Support team normally required 40 minutes to 2 hours of focused attention by a CPE engineer to manually isolate the root cause of cases, Ops Analytics was able to use machine-powered log analytics and pinpoint the root cause in seconds. Even in a few incidents that required upwards of two days (and sometimes 4 to manually resolve, the Support team found the root cause with OpsAnalytics with no effort.


application ecosystem.png


With Operations Analytics, all the relevant, timely and correlated data was collected into a single dashboard with clear visual analytics and instantly available historical views. The result was that Operations Support required a much less time to identify root causes, with fewer people involved.

Learn more at HP Discover 2015 Las Vegas


Want to hear first-hand how to lower meantime to repair (MTTR) and improve troubleshooting in your organization? Join us to learn more about this customer case, in which HP Operations Analytics used machine-powered log analytics to decrease MTTR by over 90%.


Register to attend Session B1038, “Reduce meantime to repair and improve troubleshooting with HP Operations Analytics”, June 4, 10:30–11:30 a.m.


In this joint session, you’ll also learn how the HP IT organization troubleshoots the complex multimedia application Microsoft Lync, including voice, video, web and chat, for increased communication and collaboration for more than 250,000 employees. HP IT will demonstrate how it improved Lync performance across the enterprise using HP Operations Analytics.

  • operational intelligence
About the Author


Super Contributor.

very interesting use case, showcasing clear valueprop.


Looking forward to the "how to" details, especially given that in such CPE cases you are unlikely to have a continuous steam of collection of log files and other relevant data from customer environment. Having the ability to feed relevant bunch of data asynchronously and getting analysis out of OpsA would be of great value and open up many more analytics use cases like this one.