Finding why new-hire training stalled

The problem

New hires were assigned a required custom skill track on a third-party learning platform. Managers believed completion was low, but had no reliable figure to support the concern.

Given an access account, I reviewed the 60 earlier hires who had been assigned the track and checked how many had finished it.

Three of 60 had completed the track. Five percent. The platform's records showed course completions, but not when the others stopped working, whether they had ever started, or what had interfered.

That was enough to establish the scale of the problem and nothing else. There was no usable record of work on unfinished courses, so there was no way to see where in the track people were dropping out, or when.

Tracking the next cohort

Thirteen new hires were about to onboard. I built an Airtable reporting database and a training-log procedure for that cohort, so we could follow progress from the start rather than inspect completion at the end.

I chose Airtable because its low-code tooling let me stand up the database quickly, connect an HTML form to it, and share records directly with the people responsible for onboarding.

I reviewed those records weekly alongside platform completions. Because hires started on different dates, I compared activity by onboarding week rather than by calendar date.

What the data showed, and what it didn't

A few of the 13 logged no training at all at the start. Managers followed up to ask whether they had opened the track, understood it was required, and knew when they were expected to work on it. I asked the onboarding coordinator and the hires' managers what instructions they had given, then compared those accounts against the onboarding message and the automated enrollment email from the platform. Some hires had been told directly that the track was required. Others had only received the automated email.

Most hires did begin. Some logged short sessions on most workdays; others worked in larger blocks with several days in between. Across the cohort, time-log entries fell off around weeks five and six.

A missing log entry did not establish that training had stopped. For hires whose logs went quiet, I checked for new course completions on the platform and asked whether they had trained without submitting the form. Some had missed entries. Most of the gaps we investigated, though, reflected little or no time on the platform. The decline in logs was largely a decline in training, not a decline in logging.

I then asked hires with gaps what had changed in their schedules and what training time they had planned, and asked their managers when project assignments and recurring meetings began and whether those had displaced training. The rise in project work around weeks five and six accounted for much of the timing. Several hires were also holding out for a large block of time to train, and other commitments kept taking those blocks.

The time logs showed when activity declined. The follow-ups established what was happening during the gaps. Neither would have been sufficient alone.

What we changed

1. Blocked training time on calendarsManagers reserved recurring time for the track, so hires had a regular place to work on it as project load increased, rather than waiting for an opening that might never come.
2. A recurring check-in cadenceManagers reviewed progress, asked about unfinished courses and gaps in activity, and agreed the next sessions. The check-in also gave them a place to confirm the hire understood the track was required.
3. A manager-facing dashboard and two alert rulesIf a hire hadn't submitted a first training log by the end of their third workday, Airtable notified the hire and their manager. If a hire who had started went five workdays without another entry, it notified them again. The hire's alert linked to the logging form and asked them either to record training they'd done or to resume the track. The manager's alert showed the last recorded session and the outstanding courses, so the gap could be addressed in a check-in. A new time entry reset the inactivity count, so the same gap didn't fire a fresh alert every day.

Result

By the due date, 12 of the 13 hires had completed the full track. 92 percent.