Overload and burnout

Workload data starts with the simplest summary: the team average. Overtime here means hours worked minus the 40-hour contract; nobody worked under contract in any week.

For the six coordinators over the last eight weeks of the peak: Olha 38, Taras 4, Maryna 36, Bohdan 2, Dmytro 2, Sofiia 4 hours of overtime, 86 in total. Six people × 8 weeks = 48 person-weeks. 86 / 48 ≈ 1.79 hours a week per person. On average, harmless.

That's the trap. An average over six people and eight weeks smooths away exactly what a lead needs to see: one person whose weeks keep getting longer, or a trend that started a month ago. Two people can have similar totals and very different stories. So use the average to answer one question only ("is the whole team overloaded?") and then look at each person's weeks in order.

Inside the team average, look at each person, and at the direction of their weeks, not only the total.

Olha has the most overtime over eight weeks, 38 hours. Maryna has a similar total, 36, but a very different shape: her week never gets shorter, from 40 hours to 49. In the first four weeks: 40 + 41 + 42 + 44 = 167 hours, 7 of overtime, 1.75 a week. In the last four: 45 + 47 + 48 + 49 = 189 hours, 29 of overtime, 7.25 a week. Four times as much, in a month, and still rising.

And the loudest voice isn't the evidence. Taras, who posted that everyone is drowning, worked 4 hours of overtime in eight weeks and closed the most tickets, 740. His complaint is still worth hearing (the peak was hard), but it doesn't tell you who is most loaded. Hear the complaint, then check the data.

Long weeks feel productive, but output per hour usually falls as the hours stretch. Tired people work slower, make more mistakes and spend more time recovering from interruptions. So rising hours with flat or falling output is one of the clearest early signs of overload.

Maryna's numbers show it. In the first four weeks of the period she closed 245 tickets in 167 hours: 1.47 an hour. In the last four she closed 226 in 189 hours: 226 / 189 ≈ 1.20 an hour. Twenty-two more hours, nineteen fewer tickets. At the same time her out-of-hours messages went from 11 to 50, 17 of them in the last week alone.

Read these signals together, per person and over time. One long week is noise; hours that never fall, late messages that multiply and output per hour that drops, all for a month, are a pattern. It's also a practical argument: extra hours from an overloaded person buy less and less work.

Burnout builds up from chronic work stress that nobody managed: exhaustion, growing distance from the work, a sense of doing it worse. Only a health professional can assess it, so it's not the lead's word to use about a person.

What a lead can see are early signs of overload: longer hours, late messages, slower output, skipped breaks, unused leave, withdrawal, new irritability. One sign is noise. Several together, for weeks, are a reason to talk.

Maryna's data shows several at once: hours that never fall, out-of-hours messages up from 11 to 50, tickets per hour down from 1.47 to 1.20, for four weeks. The fair conclusion is a private check-in about her workload, which the lead can change, not "she's burned out and should take sick leave". And neither "the average is 1.79 hours, so no problem" nor "Taras is the most overloaded because he says so" fits the data.

A lead can't remove stress from a peak season, but four levers are genuinely in their hands. Workload: move, drop or postpone work, and say no upward. Clarity: who owns what, so nothing falls to "whoever cares most". Control: let people decide how and when they do the work. Fair recognition: notice the work nobody sees.

At Klasna, Maryna answers EU families' evening emails "because otherwise nobody does", and until recently also picked up reopened tickets after her shift. Olha confirms the evening inbox never had an owner. Each lever has a concrete move: make the evening inbox a weekly rotation with one named owner during that person's shift (clarity, spread load); agree with Iryna to postpone the reply-template rewrite planned for July (workload); let Maryna choose which evenings she covers (control); say in the team update who has been carrying the evening emails (recognition).

Two moves look like help and aren't. A resilience webinar leaves the load exactly where it was. Checking someone's chat activity every evening is surveillance: it adds pressure and erodes trust.

Some signs go beyond workload: weeks of poor sleep, feeling hopeless, panic attacks, a health scare. A lead isn't a clinician and shouldn't try to be one. The job has four parts.

Say what you've noticed, without labels: "You mentioned weeks of bad sleep, and you've been writing late most nights", not "you're burning out". Suggest professional help: a doctor or a psychologist, and any support the company offers. Keep it confidential: what someone tells you in a one-to-one stays there, unless someone's safety is at risk. Still use the levers you control: take work off, clarify ownership, offer flexibility. Professional help doesn't replace a lighter week.

Context matters in Ukraine. Night air-raid alerts and power cuts add load that never shows in the hours column: a person may work 42 hours and sleep four a night. Ask, don't assume, in both directions. And avoid the well-meant traps: amateur advice on sleep, meditation apps instead of real help, or telling the team "to support them".

The workload data includes one more person: the lead. It's easy to read everyone else's rows and skip your own. Don't.

At Klasna the lead worked 59 hours of overtime in eight weeks, every week above 45 hours, the marker the team uses for a long week (not a medical threshold). And the out-of-hours messages: the lead sent 12 + 13 + 11 + 14 + 15 + 13 + 16 + 14 = 108. Everyone together: Olha 46, Taras 5, Maryna 61, Bohdan 3, Dmytro 6, Sofiia 5, the lead 108 = 234. So 108 / 234 ≈ 46.15%: almost half of the late-night traffic comes from the lead.

That number changes the reading of Maryna's late messages too. Some of them may be replies to the lead. A team where the lead writes late is a team where answering late feels expected, so any plan to reduce overload has to include the lead's own week.

A team learns what's normal from what the lead does, not from what the lead says. When the lead writes at ten at night, people learn that replying at ten at night is expected, whatever the working agreement says.

At Klasna the lead worked 59 hours of overtime in eight weeks (7.38 a week) and sent 108 of the team's 234 out-of-hours messages: 46.15%. That one person's habits set the norm more than any rule could. Iryna sees it too: she can't have her Support lead running on 47-hour weeks into the summer, and asks what the lead is dropping and what they need from her.

The fix is the same set of levers, applied to yourself. Schedule messages for the next shift. Say no to one thing, with Iryna's agreement. Take your own leave, visibly. A lead who is visibly overloaded can't credibly protect anyone else.

How it fits together: read people, not averages; watch several signs over weeks; use the levers you control; refer beyond workload; and fix your own week first. Next, the module's case, Ticket ownership, Olha and Taras: two rules that both claim one refund ticket.