Menu

Operational Analytics and Root-Cause Investigation

Lesson 5: Operational Analytics and Root-Cause Investigation

What is operational analytics?

Operational analytics examines the processes that deliver a product or service.

Area

Example metrics

FulfilmentOrder processing time, pick accuracy
DeliveryOn-time delivery %, median and P90 delivery time
Customer supportFirst response time, resolution time, backlog
QualityDefect rate, rework rate
StaffingUtilisation, cases per agent
CostCost per order, cost per ticket

Measure the process, not just the outcome

Metric

Why it helps

MedianTypical experience, not distorted by extremes
P90 / P95Experience of the slowest 10% / 5% of cases
Backlog sizeWork waiting to be done
ThroughputCases completed per period
Rework rateShare of cases reopened or redone
SLA attainmentShare of cases meeting the service target

An average can hide a small group of extremely delayed cases. Use median and percentiles alongside it.

Example

Support resolution time increased. Break it down:

Dimension

Possible finding

Issue typeBilling issues take 3× longer than login issues
TeamOne team is short-staffed
Time of dayEvening tickets wait longer
ComplexityMore complex tickets arrived this month

The increase may come from a shift in the mix of cases (more billing issues) rather than slower work in every category.

Root-cause thinking

Technique

How it works

5 WhysAsk "why?" repeatedly to move from symptom toward cause
Fishbone (Ishikawa) diagramGroup possible causes: people, process, tools, materials, environment, measurement
SegmentationFind where the problem is concentrated
Before/after comparisonCheck what changed around the time the problem started
Pareto analysisFind the few causes behind most of the problem (the 80/20 pattern)

Steps:

  1. List plausible explanations
  2. Check evidence for each one
  3. Compare before and after periods
  4. Investigate process changes (new software, policy, staffing)
  5. Distinguish a contributing factor from a proven cause

Key takeaways

  • Operational metrics should lead to a testable investigation or action.
  • Don't blame a team based on one metric without considering workload, case mix, and system constraints.

Try it: Choose an operational problem, such as late deliveries. List four metrics and three possible causes you would investigate before proposing a fix.


Data Analytics Tutorial



Learn via Video Course

Data Analysis and Processing with Python using Pandas - Data Science Course Logo

Data Analysis and Processing with Python using Pandas - Data Science Course

5 hrs



Get in Touch For Details