Most adherence reporting punishes the wrong people
The hidden assumption in every adherence score
Adherence is one of the three or four metrics every TL pack carries. Most operations don’t question it. The score is calculated, reported, used in coaching conversations, factored into performance reviews. The conversation is always about whether the agent’s number is high enough; almost never about whether the metric itself is measuring what we think it is.
What it actually measures is two things stacked on top of each other. The agent’s behaviour — were they where the schedule said they should be? — and the schedule’s quality — was the schedule a reasonable thing to ask of the agent in the first place? When the schedule is right, adherence cleanly measures the agent. When the schedule is wrong, adherence punishes the agent for the planning team’s error. Most operations have a schedule that’s wrong in ways the dashboard can’t see, and so most operations punish agents for things the agents didn’t do.
How this actually plays out
A team leader pulls an agent into a coaching conversation about adherence. The agent has been at 82% against an 88% target. The TL frames it as “you need to be in your seat when the schedule says you should be.” The agent privately knows three things the TL doesn’t. The 9:30 break is when the queue empties for ten minutes every day, so taking it then is operationally fine and the schedule’s 9:45 break window mostly catches the post-break recovery. The 14:00 system-update slot is when the screens freeze on half the agents anyway, so the “aux time” in their record includes those frozen-screen minutes. And the Tuesday morning shift the schedule put them on is one nobody on the team actually works because the volume has been off for three months, so they’ve quietly been swapping to Wednesday morning.
The agent isn’t adhering badly. The schedule is wrong. But the coaching conversation hangs on the agent, the agent’s record gets the “adherence” flag, the TL spends time trying to fix a behaviour that isn’t the problem, and the planning function never finds out that their schedule needed re-designing. The credibility loss compounds in three places at once: with the agent (who feels punished unfairly), the TL (who knows it doesn’t add up but can’t articulate why), and the planning team (whose schedule keeps producing the same complaints).
The four warning signs that adherence is punishing the wrong people
Adherence is rising while service level is falling. Already a diagnostic in the schedule that ages out. If everyone’s adhering more closely to a schedule that’s wrong, SL gets worse not better. The very people who are now being praised for high adherence are arriving at desks the operation doesn’t need filled at the times the schedule says it does. The metric is rewarding compliance with the wrong instructions.
The same shifts produce low adherence across multiple agents. If the 8:00 start on a particular team consistently scores 70–75% across five different agents, the agents aren’t the problem. The shift is. There’s something about that start time — a transport pattern, a system-bring-up timing, a TL’s actual on-floor arrival — that the schedule is ignoring. The fix is to redesign the shift, not to coach five agents on the same thing they can’t individually control.
The high-performers and the low-performers have similar AHT, quality, and CSAT. If the agents at the top and bottom of the adherence league look the same on every output metric, adherence isn’t a behaviour-quality signal. It’s measuring something orthogonal to actual performance — usually willingness to comply with a schedule the agent quietly doesn’t believe in.
Adherence drops on the days the operation runs hottest. The intuitive read is “agents go off-task when it gets stressful.” The honest read is usually “the schedule was built against an average day and the busy days break it.” The agents are adapting in real time to a reality the schedule didn’t plan for. Calling that “poor adherence” punishes them for filling a gap the schedule left.
The redesign that fixes it
Four moves. None is technically difficult; all of them require the planning team to accept that adherence is a joint measure of agent and schedule, not just agent.
1. Measure schedule fit alongside agent adherence. Add a second number to the adherence dashboard: the proportion of intervals where the schedule’s coverage matched the operation’s realised need. When schedule fit is below 90%, agent adherence on those intervals is uninterpretable and shouldn’t drive coaching. When schedule fit is above 95%, agent adherence is the clean signal it’s meant to be. This single change reframes the conversation.
2. Replace rigid timestamps with flex windows where the operation tolerates it. The 10:00 break becomes “between 9:45 and 10:15 depending on the floor.” The 14:00 coaching slot becomes “within the 13:30–14:30 window.” Adherence is measured against the window rather than the timestamp. The agent has the autonomy to fit their personal rhythm to the operation; the schedule is more deliverable; adherence stops being a bureaucratic compliance score.
3. Treat persistent low adherence on a specific shift as a schedule-design signal. When the same time-of-day or day-of-week shows low adherence across multiple agents over multiple weeks, the planning team owns the diagnosis. The default question becomes “what is the schedule asking that the agents can’t deliver?” rather than “why are these agents non-compliant?”
4. Stop reporting adherence below shift level until quality is high. Aggregate the metric. Report adherence at team, day, and skill level. Only drop to per-agent reporting when the team-level number is stable and high — which is the signal that the schedule is good enough for individual variation to be meaningful. Per-agent adherence in a context of poor schedule fit produces conversations that are technically defensible and operationally destructive.
The honest exception
Adherence is the right metric to lean on in two specific situations. The new-starter cohort — where the agent is genuinely learning the structure of the day and compliance with a known-good schedule is part of the development. And the persistent behavioural outlier — the experienced agent whose adherence is structurally low against a schedule that everyone else on the team adheres to comfortably. In both cases the metric is doing what it’s designed to do: measuring agent behaviour against a baseline that’s already been validated.
The argument here isn’t that adherence is a bad metric. It’s that adherence is a useful metric only when the schedule it’s measured against is honest. Without the schedule-fit check, adherence reporting routinely punishes the wrong people, and the planning function unwittingly burns the credibility it most needs.
The conversation worth having with operations leadership
The framing is concrete. “Our current adherence reporting measures the agent against the schedule, but the schedule itself isn’t graded. That means when the schedule is wrong we’re coaching agents for the planning team’s error. The fix is to grade the schedule alongside the agent — one new metric, schedule fit, reported on the same page. When schedule fit is high, agent adherence is the clean signal we’ve always wanted. When schedule fit is low, the action is on the planning team, not the floor.”
This conversation is uncomfortable for the planning function because it accepts shared accountability for a number the planning team has historically used to evaluate the floor. It’s also the conversation that builds the credibility the planning function most lacks — the willingness to grade its own work honestly. Operations that adopt the discipline find the floor relationship lifts within a quarter. Operations that don’t spend years coaching agents on a metric the agents can’t individually control.
Conclusion
Adherence is the most-reported, least-questioned metric in contact-centre operations. The metric isn’t broken — the way it’s reported is. Most adherence conversations assume the schedule is right and the agent is the variable. Most of the time the schedule has some wrongness baked into it that the dashboard doesn’t show, and the agent gets punished for the planner’s error. The fix is structural: grade the schedule alongside the agent, replace rigid timestamps with flex windows, treat persistent low adherence on a specific shift as a design signal, and stop dropping to per-agent reporting until schedule fit is high. None of this is technically difficult. All of it requires the planning function to accept shared accountability for a metric it has historically pointed at the floor. The operations that take the move find their floor relationship transforms; the ones that don’t keep losing the trust they most need.
Pair this with the schedule that ages out, the cost of perfect adherence, and adherence and conformance.