n o ren
Systems & Organizations

The On‑Call Carousel That Stalls Learning

When a team hands off night‑shift alerts to a fresh pair every week, the speed of incident resolution rises but deep expertise evaporates.

The on‑call carousel swaps a small group of engineers into the emergency queue for a short, fixed stretch, then rotates them out to regular work. At first glance the cadence looks perfect: fresh eyes, reduced burnout, and a clear hand‑off point that keeps the day‑shift focused. What most leaders miss is that the rotation creates a hidden “knowledge silo” where the most recent responders hold the only mental model of the system’s quirks, while the rest of the organization watches from the sidelines. Because the rotation is short, the responders never get the chance to document the why behind their workarounds, and the next group inherits a patchwork of undocumented fixes. Over time the incident backlog becomes a revolving door of temporary hacks, and the team’s collective understanding erodes faster than any single engineer can relearn it.

When the carousel spins, the metric that usually celebrates faster mean‑time‑to‑restore begins to mask a growing “re‑learning tax”: each hand‑off forces the new on‑call pair to spend a chunk of their shift deciphering yesterday’s ad‑hoc decisions instead of building lasting solutions. The result is a paradox where a tool designed for speed quietly fuels long‑term slowness.

The hidden cost surfaces most sharply during a high‑impact outage that lasts beyond a single rotation. The engineers who were on‑call at the start have already moved onto feature work, leaving the current crew to chase clues without the context of earlier patches. The organization spends days untangling a knot that could have been prevented with a single, well‑captured insight.

Short rotations create fresh focus but also fragment system knowledge across many engineers.
Without intentional documentation, each hand‑off adds re‑learning time that outweighs the burnout reduction.

Ignoring the knowledge‑sil​o effect turns every incident into a fresh learning curve, inflating the true cost of reliability.

The carousel also incentivizes surface‑level fixes, because the rotating crew never sees the long‑term consequences of their shortcuts.

1
Open the most recent on‑call post‑mortem document, locate the “unresolved follow‑up” section, and count how many items remain open after the next rotation; a decrease signals that knowledge is staying in the team.
2
In your incident tracker, tag the first three alerts handled by a new on‑call pair and note whether the resolution included a documented “root‑cause” note; an increase indicates better knowledge capture.

The carousel concept originated in fire‑fighter shift design, where rapid crew changes keep responders alert. Translating it to software teams works for fatigue but clashes with the need for deep, cumulative understanding of complex services.

A counter‑measure is to pair each rotation with a mandatory “knowledge‑handoff” sprint, where the outgoing crew records a concise diagram of any interim changes; this adds a small fixed cost but pays back in reduced incident latency.