Education, Training and Learning Development Case Studies
Professional5 Case StudiesUpdated Sep 2026
These five case studies follow working teams that had to repair something they depended on: accountability, a strained sponsor relationship, a recurring workflow, a policy nobody wanted, and a funding choice made with thin data. Each is set in a fictional organization and told from the seat of the person asked to fix it. All five use the same lens of roles, service levels, workflow data, stakeholder expectations, and delivery targets, so readers can see how one set of levers behaves under different pressure. Test your own plan before reading the results.
Case Study 1
A learning unit missing targets without clear owners
Brightwater Health Learning Unit, the internal education group for a regional home health network of 2,400 staff
The situation
Brightwater's learning unit has six instructional designers, two facilitators, and one coordinator who handles scheduling. Over two quarters it missed its delivery target for mandatory clinical refreshers, finishing 71 percent of scheduled modules on time against a goal of 90 percent. Requests arrived through three inboxes and a shared spreadsheet, and nobody could say who owned a module once it moved from design to pilot. Ayesha, the unit manager, inherited the team after a reorganization and found that two designers were each waiting on the other for sign-off on the same content. Facilitators complained that scripts arrived late. Regional nurse educators assumed the unit would fix whatever they flagged.
The challenge
Ayesha needed to decide whether the problem was people, process, or the target itself. Two designers were strong performers by every individual measure, and the facilitators were not slow. The unit's commitments to three hospital partners had been written before the reorganization, so nobody knew which service levels still applied. She needed an accountability plan that the team could run without her approving every handoff, and that the regional educators would accept as a real change rather than another meeting.
Analysis
Start with roles, because the symptoms all point to ownership. When two people can each block a module, the schedule is not really owned by anyone. Ayesha pulled the handoff log for the previous two quarters and counted how many modules had passed through more than three hands before pilot. The answer was most of them. Practitioners in Education, Training and Learning Development tend to treat that kind of count as the first piece of evidence, since it separates a capacity problem from a coordination problem. The service levels were the second source of confusion. The three hospital agreements specified different turnaround times, and the unit had been quietly averaging them.
The delivery target itself needed rework before anyone could be held to it. The 90 percent goal had been set from a single regional average and did not account for the two modules that depend on external clinical review, which routinely took longer. Ayesha split the target into internal and partner-facing measures, then tied each to a named owner. Stakeholder expectations mattered here as well. The educators were not asking for more hours of work. They wanted to know, before a deadline slipped, who would tell them. A plan that gave them a single point of contact per module answered that concern more directly than any dashboard would have.
Actions taken
- Assign one accountable owner to every module from request to pilot, and publish the name in the shared tracker before any new work begins.
- Replace the three inboxes with one intake form that records the partner, the agreed turnaround, and the reviewer, so every request enters with its service level attached.
- Hold a weekly fifteen-minute review of modules with more than two open handoffs, with Ayesha chairing only until the team can run it without her.
- Agree a two-tier target with the three hospital partners, separating internal turnaround from externally paced clinical review, and report both each month.
Results
By the end of the next quarter, on-time completion rose from 71 percent to 86 percent. It remains below the original goal, though the partners signed the new measure. Two of the three hospital partners reported fewer escalations, and the regional educators stopped copying Ayesha on routine status messages. The average module passed through fewer hands before pilot. One designer asked to lead the intake review, which Ayesha read as the clearest sign that ownership had moved. The unit still missed its internal turnaround goal in one month, and the team treated that as a question about reviewer capacity.
Questions for discussion
- Which of the three partner service levels should the unit treat as binding when they conflict, and who should make that call?
- How would you test whether the single-owner rule is shifting work onto one person rather than clarifying it?
- What evidence would convince you that reviewer capacity, not the designers, was the real constraint?
Case Study 2
A sponsor escalates over a reporting gap nobody owned
Northvale Learning Group, the internal training function for a regional freight carrier of 3,100 employees
The situation
Teresa runs operations for the carrier's central warehouse, and she had sponsored the new forklift safety and dispatch sequencing program. Her complaint arrived in the middle of the rollout: she said the learning group had promised weekly progress reports and then delivered a single summary after five weeks. Only two of her eleven sites had completed the modules, and nobody had told her why. Her email went to the vice president of operations and copied the group's director, which is how the matter became an escalation rather than a scheduling question.
The challenge
Hamish, the group's director, had three days to respond before the vice president's review. The complaint mixed a legitimate delivery failure with a misunderstanding about what had been promised. Nobody had written down the reporting cadence, and the group's dashboard showed completion by module rather than by site, so Teresa was reading numbers that did not answer her question. Hamish needed to calm the immediate concern, restore confidence in the group, and leave behind commitments that could be measured.
Analysis
The first task was to separate what Teresa expected from what the group had actually agreed. Hamish pulled the original scope memo and found that weekly reporting appeared only in the kickoff slides, not in the statement of work. The more useful finding was that the reporting cadence had never been assigned to a named role, so it lapsed when the coordinator moved to another project. Expectations that live only in slides tend to decay quietly, and the escalation was the first time anyone noticed the decay. Roles had to come before any apology, since an apology without a named owner would only restart the cycle.
The workflow data told a sharper story than the dashboard did. Completion by module hid a site pattern. Once Hamish recut the records by location, the two sites that had finished were the ones where a shift lead had taken on local coordination. The other nine had no one who owned the schedule at floor level. The service levels had to reflect that reality. The group offered a written update every Friday, with completion shown by site and a named coordinator at each, and a target of full completion at all eleven sites within sixty days. Teresa could check those figures herself, which mattered more than any reassurance.
- Clarify expectations Restate cadence, roles, and targets in one written scope note the sponsor signs.
- Resolve immediate gaps Assign coordinators to stalled sites and protect time for the missed modules.
- Report on schedule Send a Friday site table with completion, owners, and any recovery dates.
- Review and adjust Hold the Friday call, check the table first, and reset targets only with agreement.
Actions taken
- Call Teresa the same day to acknowledge the missed reporting, confirm the two completed sites, and name the person who will send the next update.
- Write a one-page scope note that restates the reporting cadence, the site coordinator roles, and the completion target, and send it for her signature before the review.
- Assign a trained shift lead as coordinator at each of the nine stalled sites within the first week, and give each one a fixed block of protected time for the modules.
- Hold a fifteen-minute call every Friday with Teresa and the two sites already finished, reviewing the site table before any general discussion.
Results
The Friday updates went out for the next six weeks without a gap. The matter moved from a formal escalation to a standing item on the operations agenda within the month. By the end of the sixty-day window, eight of the eleven sites had completed the modules and the remaining three were on a written recovery plan with dates. The group missed the full target, and it said so in the update rather than rounding up. Teresa's trust came back slowly, but the second half of the rollout required fewer follow-up emails than the first.
Questions for discussion
- Where should the line sit between a commitment the group can keep and one that mainly reassures the sponsor?
- How would you decide whether site coordinator roles should be formally assigned or left to volunteers?
- What would you have needed to see in the first week to know the reporting gap was structural rather than a one-off?
Case Study 3
Fixing a training record workflow that generated rework
Ridgeline Savings Bank, a regional lender with 1,800 employees whose learning administration team keeps the compliance training record
The situation
Each quarter the bank must show regulators and internal auditors that every employee in a regulated role completed required compliance training. The administration team of four receives enrollment requests from branch managers, books seats with the vendor coordinator, collects completion data from the vendor's portal, re-keys it into the records system, and then sends attestation forms back to managers. Over the last two quarters about one in five records had to be corrected after the fact. Renata, the team's administration lead, had been told that the bottleneck was the vendor, though she suspected the handoffs inside the bank were the larger source of delay.
The challenge
Renata wanted the workflow mapped end to end before proposing a change, because the vendor's portal was fixed and the bank's own steps were not. Three options were on the table from different managers: add two staff, move bookings to a self-service form, or require the vendor to send records in the bank's format. Each option promised a different service level, and each would have shifted work onto a different group. She needed to know which change would cut the rework rate, not just speed up the queue.
Analysis
Mapping the handoffs showed nine touch points between request and signed attestation, and four of them were the bank's own. The rework clustered at two of those points. The first was the re-keying step, where a coordinator typed vendor completion dates into a field that expected a different date format, so entries were rejected and retyped. The second was the attestation return, where managers received forms without the module-level detail they needed to sign, and sent them back. Nobody counted those returns as failures, so the pattern stayed invisible. Roles were also blurred: the booking coordinator and the records clerk both believed the other was checking enrollment eligibility.
Judged by rework rather than queue time, the ranking changed. Adding two staff would shorten the queue but leave the date-format fault in place, so rework would continue at higher volume. The self-service form removed the booking handoff, which the data showed was not where most errors started. Requiring the vendor to adopt the bank's format addressed the root fault, but the vendor's timeline was outside Renata's control and the contract said nothing about format. Renata proposed a staged path: fix the field mapping now, then press the vendor for format compliance with a dated request, while the bank's service levels for managers were restated to match.
Actions taken
- Replace the free-text date field with a validated date picker in the records system, and test it against the last two quarters of rejected entries before release.
- Add a rework counter to the monthly report, so every re-keyed or returned record is logged with its reason and the handoff where it occurred.
- Send attestation forms with module-level detail attached, and agree a five-day return window with branch managers as their agreed service level.
- Write a dated request to the vendor for bank-format completion files, with a fallback manual check if the date slips past the next quarter.
Results
The date validation went live in the second month and removed most rejected entries before they reached a coordinator. Rework dropped from one record in five to about one in twenty over the following quarter, and the counter showed which steps still caused returns. Branch managers usually returned forms within the five-day window, and the remaining delays traced to two branches with leave coverage gaps. The vendor agreed to a format change but moved its date later than requested, so the manual check ran for one extra cycle. Renata's team now reports rework alongside queue time at every review.
Questions for discussion
- Which of the three options would you have ranked first if rework had not been measured, and what would that ranking have cost the bank?
- When a vendor's timeline sits outside your control, how would you choose between waiting for compliance and building an internal fallback?
- Where in your own handoffs would returned work be invisible in current reporting, and how would you surface it?
Case Study 4
Rolling out a development policy against open resistance
Osprey Mutual Insurance, a regional mutual insurer with 2,700 employees and a seven-person learning team
The situation
Osprey Mutual's leadership decided that every people leader would hold a documented development conversation with each direct report each quarter, recorded in the talent system with a shared competency rating. The policy was announced in a single town hall and followed by a two-page guide. Adoption in the first month was low. Claims supervisors said the quarterly cycle duplicated the existing annual review. Several senior managers argued that the competency framework did not match their work. Mariam, who leads the learning team, was asked to make the rollout succeed within the quarter.
The challenge
Resistance and confusion were tangled, and Mariam needed to separate them. Some objections were about workload, some were about whether the framework fit frontline roles, and some were simply a reaction to being told what to do in a town hall. The policy could not be softened without losing its audit purpose, so the question was sequencing: which groups needed to see the process first, which needed a different version of it, and how uptake would be measured without turning the rollout into a compliance exercise.
Analysis
Claims supervisors were reacting to a real duplication: their annual review already covered most of the competencies, and the quarterly conversation added a second rating for the same work. That is a design fault, and Mariam took it to the policy owner rather than defending the timing. Senior managers raised a different concern about fit, which the team tested by asking eight of them to map their own roles against the framework. Six reported that two or three competencies had no observable behavior in their area. The town hall had sent the message that the policy was settled, which made every objection sound like defiance.
The sequence followed the concerns. The claims supervisors went first, with the annual review folded into the quarterly conversation so that one rating served both purposes. Senior managers received the local examples before any further rollout, and their feedback was written into the guide. Only then did the policy return to the full population, with a target that mattered: ninety percent of conversations documented within the quarter, measured in the talent system and reported by team rather than by individual. Uptake was tracked weekly for the first month, then monthly, because the early numbers mostly reflected how quickly people had read the guide.
Actions taken
- Fold the annual review into the quarterly conversation for claims supervisors, so one rating covers both and the duplicated competencies are removed from the form.
- Ask each senior manager to write two local examples for the competencies they cannot observe, and publish them in the guide before the wider rollout begins.
- Run one thirty-minute briefing for each team, led by a peer manager who has already completed the cycle, rather than another town hall.
- Publish a weekly uptake table by team, showing documented conversations against the ninety percent target, and share it with each senior manager before the wider group sees it.
Results
Documented conversations reached 74 percent in the next quarter, short of the target but up sharply from the first month. Claims supervisors reported that the single rating removed most of their duplication complaints. Senior managers' local examples were adopted into the guide and cited by two other business units when they asked for their own version. Two teams stayed below half, and the uptake table showed that both had a manager on extended leave with no delegate named. Mariam's team assigned delegates within a week, and the figures for those teams moved above the target by the following month.
Questions for discussion
- How would you decide which objections deserve a design change and which call for better explanation?
- What risks come with letting a peer briefing replace the leadership message, and how would you limit them?
- If uptake had stayed flat after the delegate fix, what would you have checked before changing the policy itself?
Case Study 5
Choosing among three funded options with thin evidence
Halden Works, a mid-sized maker of industrial pumps with 1,500 employees, whose learning office reports to operations
The situation
Halden Works' learning office was asked to recommend how to spend its discretionary budget for the coming year. Three proposals were on the table. A hands-on simulator for maintenance technicians would cut the time new hires spent on live equipment, but its vendor had only two reference sites. A second apprenticeship cohort would widen the pipeline of machinists, though the first cohort had not yet finished and its completion rate was unknown. A supervisor development pilot would address rising complaints about shift handovers, but the only measure of its effect would be a survey with a low response rate. Each sponsor believed their option was the priority.
The challenge
Gregor, who directs the learning office, had to recommend one option or a sequence of them to the operations committee, which would not revisit the question this year. The budget covered one of the three proposals in full, or two if the pilot were scaled down. The data was incomplete for every option, and waiting for more of it would mean the budget went unspent. He needed criteria the committee would accept and a way to stage the commitment so that a wrong guess would be cheap to reverse.
Analysis
Gregor began by setting criteria that did not depend on the missing data. He settled on four: effect on the operating constraint each option claimed to address, the cost of being wrong, how quickly the result would be visible, and whether it produced usable evidence. Under those criteria the simulator scored well on speed but poorly on evidence, since the vendor's two sites were not comparable to Halden's plants. The apprenticeship cohort was strongest on cost of being wrong, because its commitments could be paused at the end of each term. The pilot's weakness was measurement, and the workflow data on handovers was not yet captured consistently across shifts.
Risk was uneven across the three options. The simulator's main risk was a sunk commitment with a single vendor and no local proof, so it could only be funded after a short trial on one line. The apprenticeship expansion carried a timing risk, since the next cohort would start before the first reported, and that could be held by freezing intake. The pilot needed a better measure before any scale-up. Gregor recommended holding the apprenticeship intake at its current size, a supervisor pilot on one shift with a handover audit, and a decision point on the simulator once the trial had run for a full quarter.
- 1Define criteriaFour weighted tests agreed before any option is discussed on its merits.
- 2Fund the trialSimulator runs one line for a quarter with agreed time-to-competence measures.
- 3Hold intake steadyApprenticeship cohort stays at current size until first completion data arrives.
- 4Redesign the pilotHandover audit replaces the survey as the main measure of supervisor effect.
- 5Review at quarter endCommittee compares evidence against criteria and releases the next tranche.
Actions taken
- Write the four criteria and their weights into a one-page decision note, and send it to the operations committee before any option is discussed on its merits.
- Fund a six-technician simulator trial on one production line, with a fixed set of time-to-competence measures agreed with the line supervisor in advance.
- Hold the apprenticeship intake at its current size until the first cohort reports completion, and record the reason for the hold in the committee minutes.
- Redesign the supervisor pilot around a handover audit and a survey with a response-rate target, so the next decision rests on observed handovers.
Results
The committee approved the three-part plan at its next meeting. The simulator trial ran for a full quarter on one line, and time-to-competence came in below the old baseline, though the sample was too small to commit on. The apprenticeship intake held flat until the first cohort reported a completion rate high enough to reopen it the following cycle. The supervisor audit found that handovers on one shift were the main gap, which redirected the next budget request. The criteria table is now a standing item in the committee's minutes.
Questions for discussion
- Which criterion would you weight most heavily if the committee's main concern were reputational rather than financial, and why?
- At what point would you stop a staged commitment that is producing mixed signals, and who should make that call?
- How would you explain a decision to hold a promising option to a sponsor who believes the data is already sufficient?
Learning Resources
Professional CASE STUDIES
China, India, United States, Indonesia, Pakistan, Nigeria, Brazil, Bangladesh, Russia, Japan, Mexico, Philippines, Vietnam, Egypt, Turkey, Germany, United Kingdom, France, Italy, Canada, South Korea, Australia, Spain.