Case Room · Case lab

Diagnose before you decide

Five stages, one case. Find the real problem behind the symptoms, weigh the evidence for each cause, choose a strategy that acts on the strongest, plan it against the client's deadline, and judge the results when they come in. Each stage opens when the one before is done.

Back to 2.3 Leadership and management

Uses the Halden Distribution case file (fictional; simulated data). Afterwards, try the Decision Room on the same case, which asks how to lead.

Interaction

Halden Distribution: problem, cause, strategy, implementation, evaluation

Source: Arjun Agrawal framework

You are advising Amara Osei, Halden's new site manager, at the end of July, once her staff survey is in.

Since a warehouse management system arrived last December, on-time dispatch has fallen from 98.1% to 91.4%, picking errors have more than doubled and over half of the night shift has left within the year. Lumen Home, a third of the warehouse's work, wants dispatch above 97% by the first week of November. Before Amara chooses how to lead, she wants the problem diagnosed. Work through the five stages in order; each one starts from the last.

  1. 01 Problem Now
  2. 02 Cause Locked
  3. 03 Strategy Locked
  4. 04 Implementation Locked
  5. 05 Evaluation Locked

Stage 1 of 5.

The evidence 6 exhibits from the Halden Distribution case file. Fictional; simulated data.

The full case, with its narrative and stakeholders: Halden Distribution.

Exhibit 1

Site performance over the last year

Source: Simulated data
Month On-time dispatch (%) Picking errors per 1,000 orders Absence rate (%)
August 202598.12.13.1
October 202597.62.33.4
December 202596.03.44.2
February 202694.24.04.9
April 202692.74.65.3
June 202691.45.25.8

The warehouse management system went live in December 2025.

Source Simulated teaching data, from the site's monthly reports.
Exhibit 2

People, by shift, over the last twelve months

Source: Simulated data
Measure Day shift Night shift Whole site
Employees now8258140
Average number of employees in the year8154135
Employees who left in the year182846
Labour turnover (%)22.251.934.1
Average length of service (years)7.42.95.5
Team leaders' average experience as team leaders (years)9.10.6
Training on the new system (hours per person)62

Labour turnover is the number of employees who left divided by the average number of employees, multiplied by 100. Two years ago it was 18.9% for the whole site.

Source Simulated teaching data, from the site's personnel records.
Exhibit 3

Staff survey, July 2026 (112 of 140 responded)

Source: Simulated data
Statement Day shift agreeing (%) Night shift agreeing (%)
I know what is expected of me8154
I am consulted about changes that affect my work3821
I have had the training I need to use the new system7233
My team leader helps me solve problems7448
I would recommend Halden as a place to work5829

66 day-shift and 46 night-shift employees responded. Across both shifts, the average share agreeing with the five statements is 53%, the engagement score used in the Decision Room.

Source Simulated teaching data.
Exhibit 4

The client's deadline

Source: Simulated data

Our customers were promised next-day delivery. Since the spring, nearly one order in eleven from your site has missed its truck. We need on-time dispatch back above 97% by the first week of November, when our peak season starts. If it is not, we will move our contract when the current year ends.

Claire Hendry, operations director, Lumen Home
Source Fictional email, written for this case.
Exhibit 5

Why night-shift leavers left (exit interviews with 24 of the 28)

Source: Simulated data
Nobody explained the new scanners properly
11
Targets changed without warning
7
Better pay elsewhere
4
The shift pattern
2
Source Simulated teaching data, from exit interview notes. Each leaver gave one main reason.
Exhibit 6

How the previous manager led

Source: Simulated data

Graham knew everyone's kids' names. He decided everything himself, but you always felt looked after. When the scanners came in, he just told us they were coming.

Dev Kaur, day-shift team leader, fourteen years at Halden
Source Fictional interview, written for this case.
  1. 01 Problem

    What is the problem?

    A diagnosis starts by separating what can be seen (symptoms) from what has to be solved (the problem) and from what will happen if it is not (consequences). A good problem statement is specific, rests on evidence and does not contain its own solution.

    Read: Exhibit 1 Exhibit 2 Exhibit 4

    Is each of these a symptom, the problem, or a consequence?

    1. On-time dispatch has fallen from 98.1% to 91.4% since the new system went live.

      Is each of these a symptom, the problem, or a consequence?
    2. Picking errors have risen from 2.1 to 5.2 per 1,000 orders.

      Is each of these a symptom, the problem, or a consequence?
    3. Over half of the night shift left within the year (labour turnover of 51.9%).

      Is each of these a symptom, the problem, or a consequence?
    4. The night shift was given the new system with a third of the day shift's training and almost no say, and cannot yet work with it.

      Is each of these a symptom, the problem, or a consequence?
    5. Lumen Home will move its contract if dispatch is not above 97% by November.

      Is each of these a symptom, the problem, or a consequence?

    Choose an answer for each item, then open its answer to compare. 5 to do. Choose your answers, then check each one or check them all.

    1. Which is the best statement of the problem?

      Which is the best statement of the problem?

    Choose an answer for each item, then open its answer to compare. 1 to do. Choose your answers, then check each one or check them all.

    Who, what, since when, and the evidence. No solution yet. Kept on this device for your memo.
    Check every answer in this stage to continue.
  2. 02 Cause

    What is causing it?

    Several causes can be suggested for any problem. Weigh each by its evidence before choosing what to act on. A cause with strong evidence and a large effect is worth more than three with a little of each.

    Read: Exhibit 2 Exhibit 3 Exhibit 5 Exhibit 6

    How strong is the evidence for each possible cause?

    1. Too little training on the new system

      How strong is the evidence for each possible cause?
    2. Changes made without consulting the people who do the work

      How strong is the evidence for each possible cause?
    3. Inexperienced night-shift team leaders

      How strong is the evidence for each possible cause?
    4. Pay is lower than elsewhere

      How strong is the evidence for each possible cause?
    5. The scanners or the software are faulty

      How strong is the evidence for each possible cause?
    6. The night shift does not care about the work

      How strong is the evidence for each possible cause?

    Choose an answer for each item, then open its answer to compare. 6 to do. Choose your answers, then check each one or check them all.

    Name the evidence, and say how much of the problem it explains. Kept on this device for your memo.
    Check every answer in this stage to continue.
  3. 03 Strategy

    What should Halden do?

    A strategy should act on the strongest causes, fit the time and money available, and not create a new problem. Judge each option against all three.

    Read: Exhibit 2 Exhibit 4 Exhibit 5

    1. Which strategy should Amara recommend to the owners?

      Which strategy should Amara recommend to the owners?

    Choose an answer for each item, then open its answer to compare. 1 to do. Choose your answers, then check each one or check them all.

    What must be true for it to work? Kept on this device for your memo.
    Check every answer in this stage to continue.
  4. 04 Implementation

    Plan it in the time available

    A plan is a sequence: some activities cannot start until others finish. The longest chain of dependent activities sets the earliest finish. The plan starts in the first week of August; Lumen Home's deadline, the first week of November, is 14 weeks away.

    Must each of these finish before the pilot in two zones can start?

    1. Listening sessions with the night shift

      Must each of these finish before the pilot in two zones can start?
    2. Coaching for the two night team leaders

      Must each of these finish before the pilot in two zones can start?
    3. Redesigning the night pick routes with a group of pickers

      Must each of these finish before the pilot in two zones can start?
    4. Retraining the night shift on the scanners and the new routes

      Must each of these finish before the pilot in two zones can start?
    5. Day-shift buddies for night pickers

      Must each of these finish before the pilot in two zones can start?

    Choose an answer for each item, then open its answer to compare. 5 to do. Choose your answers, then check each one or check them all.

    The plan, week by week

    The plan takes 13 weeks: 1 week inside the deadline.

    The plan, week by weekThe plan takes 13 weeks; the deadline is the end of week 14. Critical path: Listen to the night shift, then Redesign night pick routes, then Retrain the night shift, then Pilot in two zones, then Roll out and review weekly, then Peak readiness check. Coach the night team leaders has 3 weeks of float. Day-shift buddies has 1 week of float.123456789101112131415WeekListen to the night shiftCoach the night team leadersRedesign night pick routesRetrain the night shiftDay-shift buddiesPilot in two zonesRoll out and review weeklyPeak readiness checkLumen Home's deadline
    Plan Black bars are the critical path; blue bars have float, shown as a dotted line to the latest finish. The dashed red line is the deadline. This is a Gantt chart: Gantt charts and critical path analysis, with its float, are HL toolkit content (guide, p. 14).
    The plan as a table
    Activity Weeks Starts in week Ends with week Total float Path
    Listen to the night shift1110Critical
    Coach the night team leaders6163Has float
    Redesign night pick routes3240Critical
    Retrain the night shift3570Critical
    Day-shift buddies48111Has float
    Pilot in two zones2890Critical
    Roll out and review weekly310120Critical
    Peak readiness check113130Critical
    Which activity has no room to slip? Kept on this device for your memo.
    Check every answer in this stage to continue.
  5. 05 Evaluation

    Is it working, and how would you know?

    Decide how success will be measured before the results arrive, then judge the results against the objective, the trend and the time left, not one number on one day.

    Read: Exhibit 1 Exhibit 4

    Is each of these a good measure of whether the plan is working?

    1. On-time dispatch each week, by shift

      Is each of these a good measure of whether the plan is working?
    2. Picking errors per 1,000 orders, by shift

      Is each of these a good measure of whether the plan is working?
    3. Night-shift leavers, counted over each rolling three months

      Is each of these a good measure of whether the plan is working?
    4. Hours of training delivered

      Is each of these a good measure of whether the plan is working?
    5. Halden's profit for the year

      Is each of these a good measure of whether the plan is working?

    Choose an answer for each item, then open its answer to compare. 5 to do. Choose your answers, then check each one or check them all.

    Check-in
    Simulated check-in, week 10 of the plan
    Source: Simulated data
    Measure Before the plan Week 10
    On-time dispatch (%)91.495.6
    Picking errors per 1,000 orders5.23.3
    Night-shift leavers in the last ten weeks3
    Night shift agreeing they have had the training they need (%)3371
    Night shift agreeing they are consulted about changes (%)2158
    Simulated for this lab, to practise judging results. Week 10 is four weeks before Lumen Home's deadline. Three leavers in ten weeks from about 56 night-shift employees is an annual rate of about 28%, against 51.9% a year ago.
    1. What is the most defensible judgement at week 10?

      What is the most defensible judgement at week 10?

    Choose an answer for each item, then open its answer to compare. 1 to do. Choose your answers, then check each one or check them all.

    Go back down the chain. Which cause might not have been treated? Kept on this device for your memo.
    Check every answer in this stage to continue.
Diagnosis memo

To the owners of Halden Distribution

Your sentences from each stage, beside the platform's reasoning. Compare them, then improve yours.

  1. Problem
    Yours

    Not written yet.

    The platform's reasoning

    Since the system went live in December, the night shift, trained for two hours against the day shift's six and consulted less, has not been able to work with it: dispatch and accuracy have fallen and over half the shift has left. Lumen Home's deadline makes this urgent, but it is the consequence, not the problem.

  2. Cause
    Yours

    Not written yet.

    The platform's reasoning

    Two causes have strong evidence and explain most of the problem: too little training on the new system and changes made without the night shift's involvement. Inexperienced team leaders made both worse. Pay matters to a few leavers; nothing suggests faulty scanners or a lack of effort.

  3. Strategy
    Yours

    Not written yet.

    The platform's reasoning

    Retrain the night shift on the system and redesign night picking with the people who do it, supported by day-shift buddies and coaching for the night team leaders. It acts on the causes with the strongest evidence, costs far less than a bonus or a new system, and fits the three months left, provided the training is finished early enough to leave time for a pilot.

  4. Implementation
    Yours

    Not written yet.

    The platform's reasoning

    Listening, redesign, retraining, the pilot, the roll-out and the readiness check form the critical path: thirteen weeks, one week inside the deadline. Coaching and buddying have float. The risk is retraining: every extra week it takes moves the finish by a week, so at four weeks the plan only just meets the deadline and at five it misses it. Train in small groups from the first week the routes are ready, and keep the coaching going alongside.

  5. Evaluation
    Yours

    Not written yet.

    The platform's reasoning

    At week 10 the plan is working but has not yet succeeded. Dispatch is 95.6% and rising by about 0.4 points a week, errors are down by more than a third, and the night shift reports the training and consultation it lacked. Keep going, review weekly by shift, and if dispatch stalls, go back to the causes: check the team leaders' coaching and whether the redesigned routes are being used.