Trend analysis allows Problem Management to detect errors in the Infrastructure / Application which give rise to recurring problems or incidents. Trend Analysis is performed on incidents to check for some distinct patterns (e.g. Large number of incidents occurring on a certain area, large number of incidents logged against one specific application etc.) which indicate the presence of a potential underlying problem.

The Problem Manager should analyze:

• Frequency of incidents/problems following change requests
• Frequency of incidents/problems around critical business e.g. financial period / year ends, high demand periods
• Service level breaches
• Incident volumes by application/technology
• Incident volumes by closure / resolution category
• Recurrent issues
• Known errors
• Volume of major problems.

Additionally, the Problem Manager must also monitor asset trends for performance, to check for underlying problem or a need to revisit thresholds. Failure mode analysis is a technique to identify recurring/potential failure points in an application/system and determine the impact of each failure. These patterns of events can be defined and programmed into analytical and correlation tools for future recognition.

While carrying out trend analysis, the Problem Manager must consider:

• Consistent data quality ( assuming same or equivalent methods are used for all monitoring)
• Time frame and number of samples, to ensure statistical soundness
• Seasonality of data (parameters that vary during specific times of month or year)
• Type of data distribution
• Similar variance across data sets. minimal missing observations and outliers.

The results of trend analysis can be either specific or general problems that require further investigation. The Problem Manager produces a Trend Analysis Report containing analysis results and the action plans.