Overtime is work past the agreed standard week — 40 hours in most of the IT services world. The legal shell varies: employment law regulates and prices it for employees (statutory multipliers, caps), while B2B contractor agreements — the CEE IT norm — leave it to whatever the contract says, which makes an explicit internal policy more important, not less.
Three questions a policy has to answer in writing:
- Is it compensated, and how — extra pay, time off in lieu, or explicitly nothing beyond the agreed fee.
- Is it billable — client-caused overtime on time-and-materials work can be, at rates agreed up front; discovering the answer during an invoice dispute is the expensive path.
- Who approves it, before it happens — overtime nobody asked for is a planning failure being paid for twice.
The management read matters more than the accounting: a deadline week happens; overtime that repeats every sprint means the plan books people past their real capacity, estimates run systematically hot, or scope is creeping. All three are cheaper to fix at the plan level than to pay for at the burnout level — attrition is the deferred invoice.
In Helia, overtime pay isn't modeled — but the plan side is guarded: booking someone past 100% triggers an explicit over-allocation warning at edit time (40 hours = 100%), and the capacity matrix and weekly forecast flag overbooked people and weeks, so planned overload is visible before it turns into logged overtime.