Scope creep is rarely one big unpaid feature. It is twenty small ones: a "tiny" extra report, a support task nobody scoped, a meeting cadence that doubled, an integration with three surprise edge cases. Each is too small to argue about; together they are why a project that looked healthy at signing closes at half its planned margin.
How it lands depends on the billing model. On fixed price it is a direct margin loss — the fee is fixed, the hours are not. On time and materials the hours do get billed, but the client's budget expectation doesn't move with them — the damage arrives as an invoice dispute. On a retainer it shows up as over-servicing: a "40-hour" month quietly consuming 60.
Detection is quantitative, not emotional:
- Plan versus fact. Compare planned allocations against logged hours per project per month; a persistent gap in one direction is creep (or padded plans — also worth knowing).
- Margin by project. Delivery cost climbing against flat revenue is the classic signature.
The defense is the change-request habit plus written scope with explicit exclusions — and a delivery manager empowered to say "gladly, as a change request".
In Helia, the ingredients of that comparison sit in one system: planned allocations in the capacity matrix, locked timesheet hours as the fact, invoice lines labelled "from timesheets" versus estimated, and per-project margins in the profitability view — so the drift shows up while it is still a conversation, not a write-off.