Configure Indicators and Indicator Sources

Indicators are the heart of Performance Analytics. An indicator (also called a KPI) is a measurement you track over time. Each collected value is a score, and scores accumulate into a time series that you analyze for trends. This is the single most important distinction in PA versus Reporting: PA stores a history of scores at regular intervals, whereas a report shows the current state of the data.

Indicator types

Know how each type gets its scores.

• Automated indicator: Collects scores from an indicator source in scheduled data collection jobs. This is the most common type. You point it at an indicator source, choose an aggregate (Count, Count Distinct, Sum, Average, Minimum, Maximum, or a script), and a score is stored for each collection period. Use automated indicators whenever the data lives in an instance table.
• Manual indicator: Has no indicator source. Scores are entered by hand in the scoresheet (or imported), which suits data that does not exist in an instance table, for example results from an external survey or a budget figure.
• Formula indicator: Calculates its score from other indicators (and constants) with a formula, referencing each indicator in double square brackets, for example [[Reopened]] / [[Resolved]] * 100. Formula indicators have no indicator source and are not collected: their scores are calculated when a user views the indicator. By default, if any contributing indicator has no score for a period, the formula returns no score; Allow formula component to be NULL changes that. An indicator that is used in a formula indicator cannot be deleted until the formula indicators are dealt with.
• External indicator: An advanced option for scores that come from an external data source.

A useful mnemonic: automated = from a source via a job; manual = typed or imported, no source; formula = math on other indicators, calculated when viewed.

Indicator sources

An indicator source is a data set of filtered records from one table or database view. It specifies:

• The facts table (for example incident).
• Conditions that filter the records, normally including a date-related condition such as [Opened][on][Today] for a daily indicator.
• The Valid for frequency (daily, weekly, monthly and so on). Indicators based on the source use it as their Frequency, which sets the length of each score collection period.

Indicator sources must exist before you can create an automated indicator. Many indicators can share one indicator source, and that is best practice: data collection queries the database once per indicator source, so all indicators on a source get data from the same point in time. Each indicator can add its own Additional conditions on top of the source conditions.

Indicator records and behavior

Key fields on an automated indicator include its aggregate, unit, direction (Minimize, Maximize, or None: whether lower or higher is better, which drives the score color, the change arrow, and the gap to target), precision, Collect records (stores snapshots of the record sys_ids so users can drill down), and Value when nil (the value inserted as the score when no value is collected; it is not used to calculate formulas or aggregates such as an average). The docs advise avoiding the Average aggregate where possible and dividing a Sum indicator by a Count indicator in a formula instead. Indicators can be organized with indicator groups. Targets, thresholds, breakdowns, and visualizations all attach to indicators.