← People analytics

People analytics · technique

Root-cause analysis: investigate the system

Root-cause analysis is often used as shorthand for investigating why a problem occurred.

Statistical and evaluation practice; ADM HR worked application · 2026 teaching guide · Original editorial explanation
Root-cause analysis: investigate the system — simplified model sketchReconstruct events · Explore conditions · Test explanations · Improve & review. Editorial interpretation, after Statistical and evaluation practice; ADM HR worked application (2026 teaching guide).Reconstructevents01Exploreconditions02Testexplanations03Improve &review04ADM HR · ORIGINAL INTERPRETATIONStatistical and evaluation practice; ADM HR worked application · 2026 teaching guide
Original simplified sketch after Statistical and evaluation practice; ADM HR worked application (2026 teaching guide). Read the explanation for assumptions and limitations. Download SVG ↗

The core idea

Root-cause analysis is often used as shorthand for investigating why a problem occurred. In complex work there may be several interacting contributing conditions rather than one root cause. NHS system-based learning guidance provides a useful example of examining work, context and interactions; our HR application is an editorial adaptation, not a clinical investigation protocol. Asking why repeatedly can help inquiry, but the answers need evidence.

Source and attribution [1]

Using it in practice

Define the event or recurring pattern and reconstruct what happened using records and accounts from those involved. Ask what information, resources and constraints were present at the time. Distinguish hindsight from what people could reasonably know. Test competing explanations and look at normal successful work as well as failures. Choose actions that change the conditions, then evaluate whether the problem becomes less likely.

An example, not a reported case

Worked example · illustrative

A fictional employer repeatedly loses starters before their first paid shift. One explanation is poor commitment. A review instead finds delayed access approval, inconsistent contact details and uncertainty about available work. HR checks these observations across cases and with new starters. A pilot clarifies ownership of the handover and sends a confirmed first-shift plan. The evaluation tracks delays and actual starts rather than counting reminder emails.

What to watch for

A neat five-whys chain can reflect the investigator’s preferred story. Avoid ending at human error as if that explains the conditions that shaped action. Learning work does not replace appropriate accountability or formal processes. The purpose and handling of interviews should be clear, particularly where people may fear blame.

Test rather than assume

For each proposed cause, ask what evidence would be expected if it were true and what could contradict it. Look for cases where the same condition existed without the problem. This can reveal missing factors or an overly simple explanation.

Design a stronger action

A reminder or retraining may be insufficient if the system still makes the wrong action easy. Consider clearer ownership, better information and redesigned handovers. Assign an owner and review whether the action changed the relevant condition, not merely whether it was completed.

Describe the problem without embedding a cause

Use a statement such as 'new starters at these sites leave within ninety days more often than the comparison group', rather than 'poor managers drive people out'. Define the period, population and evidence. Keep the first statement open to several explanations. A cause embedded in the problem description can steer interviews and data selection toward confirming it, making the investigation appear more conclusive than it is.

Map the process around the problem

Trace recruitment promises, first-day readiness, training, scheduling, supervision and the exit decision. Ask where information or support breaks down. Use employee accounts and operational records to test the map. Methods such as repeated 'why' questions can generate hypotheses, but a plausible chain of answers is not proof. Complex workforce outcomes may have several interacting causes rather than one root that can be removed neatly.

Test the strongest explanations

For an illustrative scheduling hypothesis, compare the timing and predictability of shifts across relevant groups and ask whether leavers described the same issue. Look for cases that do not fit the explanation. If a site has unpredictable shifts but good retention, investigate what differs. This helps avoid a story built only from confirming examples. Keep confidence levels visible and distinguish a documented process failure from an inferred reason for individual behaviour.

Choose a reversible improvement and learn

Pilot a change linked to the evidence, with a clear outcome and a way to detect unintended effects. If clearer shift information is introduced, check whether employees receive it, whether it is reliable and whether the expected outcome changes over a suitable period. Do not call the problem solved merely because an action was completed. Retain alternative explanations and revise the analysis if the intervention does not work as expected.

Avoid ending the inquiry at a convenient human explanation

An error labelled carelessness may involve confusing instructions, fatigue, missing information or incentives that reward speed over checking. Gather evidence from the work and test more than one explanation. Repeatedly asking why can structure inquiry, but it can also produce a plausible story unsupported by data. Distinguish a contributing condition from a demonstrated cause and avoid assuming every problem has one root.

Test the proposed remedy against the mechanism

If handover information is missing, a generic communication workshop may not address the fact that shifts have no overlap. Trial a specific change and examine whether the missing information and subsequent errors reduce. Record unintended effects, such as extra unpaid work or duplication. Include employees who perform the task in designing the test. A root-cause exercise is complete only when it informs a useful action and review, not when a diagram contains an authoritative-looking final box.

Take it into your next conversation

Three useful questions.

  1. What question can this method answer, and what can it not establish?
  2. Are the comparison, observation period and assumptions defensible?
  3. What decision follows, and how will its consequences be reviewed?

Go to the evidence

Sources & attribution

Further reading: Nadeem Khan and Dave Millner (2023), Introduction to People Analytics: A practical guide to data-driven HR, 2nd edition, PDF pp. 227, 244 and 294–295. Edition & reading notes ↗

[1] NHS England: system-based learning response toolkit ↗

The core idea is an original summary of the cited work. Application notes, examples and sketches are our interpretations, not quotations or reproductions of the authors’ figures. Publisher records may require access to read the full original work.

Explore People Analytics →

Application guidance expanded on 26 September 2026. These practical suggestions and invented examples are editorial explanations; further-reading sources are not the original authors of the historical models unless expressly identified. Original overview review dates are retained.

Published 2026-09-20 · Reviewed 2026-09-22. Editorial approach

Keep exploring

A little more context.

People analytics · practical playbook01

From a business question to a measured outcome

A complete worked route through casual-worker retention, from definitions to a fair pilot and an honest conclusion.

ADM HR editorial workflow6 min
People analytics · models & decisions02

Logic models for HR interventions

Separate resources, activities and deliverables from the changes an initiative is meant to produce.

Programme-evaluation tradition6 min
People analytics · technique03

Cohort analysis: compare like starting points

A cohort is a group defined by a shared starting event or period, such as joining in a particular quarter.

Statistical and evaluation practice5 min