From $4.9M in Hidden Losses to a Scalable Training System

Onboarding was tribal. New hires learned by watching others, knowledge lived in people's heads, and there was no scalable system for bringing people up to speed consistently. As the team grew, three failure modes appeared simultaneously.

Exit interview data made the driver clear: this wasn't a hiring problem. It was a retention and experience problem — rooted in how (or whether) people were onboarded.

Dissatisfaction, performance failures, and abandonment — three of the four categories — are directly addressable through better onboarding and training. That insight became the foundation of the business case.

Before presenting any solution, I needed to answer one question for leadership: what is this actually costing us? I built a cost model across two dimensions — the direct cost of onboarding time lost to each separation, and the total cost of replacing employees who left.

That $792,485 was only the onboarding hours cost. The fuller picture included the cost of actually replacing each employee who left — and what the trajectory looked like if the separation rate continued.

The second part of the presentation made the case that eLearning — specifically EdApp's microlearning platform — was the right lever. I grounded this in external research so leadership wasn't being asked to take my word for it.

These weren't presented as guarantees — they were presented as directional benchmarks showing that the investment in a structured LMS system had a strong track record of return across industries and company sizes.

After securing leadership buy-in, I designed and implemented the LMS in EdApp — structured learning paths, role-specific content, completion tracking, and a safety training module built to directly address the TRIR risk.

The design principle was simple: knowledge should live in the system, not in a person. 40 hours of peer training time per new hire — the single largest cost driver in the separation model — needed to become structured, trackable, and reusable.

This project didn't start with a clear brief. It started with something closer to: "onboarding is a mess, can you fix it?" A request like that is a starting point, not a requirement. My first instinct wasn't to open a spec doc — it was to figure out whose pain this actually was and what had finally made it visible.

The data was already there. Exit interviews, separation rates, a rising TRIR — but no one had connected it to onboarding yet. Before writing a single user story, I needed to understand the problem well enough to define it precisely. That meant asking questions most people skip.

I pulled everything that existed: separation data, exit interview themes, the cost model built in sections 01 and 02. Then I talked to the people actually doing the work — trainers, peer mentors, a few recent hires. They surfaced friction that never showed up in a ticket or a leadership meeting. That ground-level input shaped what went into the MVP and what stayed out of it.

Once the problem was defined clearly enough to spec against, I translated the findings into requirements and worked with stakeholders to prioritize what would actually move the needle in the first release versus what belonged in a future iteration. That conversation happened before a single story was written — because scope decisions made late cost far more than scope decisions made early.

For this project, the final specification included:

The LMS implementation addressed the three cost drivers identified in the business case: onboarding time waste, high separation rates, and safety incident exposure.

Beyond the metrics, the less visible outcome was cultural: onboarding stopped being a burden passed from one senior person to the next, and became something the organisation owned, tracked, and could improve over time.