scope creep
Also called: feature creep
The slow growth of a project's scope through small additions that nobody decided on as a change, so the date and the budget stay the same while the work grows.
Scope creep is the slow growth of a project's scope through small additions that nobody decided on as a change. Each addition looks reasonable on its own: one more page, one more report, a free extra for customers. The date and the budget stay the same while the work grows, so the overrun shows up late, usually as a missed date or quietly lowered quality.
Creep comes through side doors: a request in a corridor or a chat, a sponsor's idea taken as an order, a team member polishing beyond what was agreed, or simply a brief without an out list, where anything not explicitly excluded seems allowed. It is not the same as change. Changes are normal and often right; creep is change that nobody priced and nobody decided.
The defense is light and procedural, not a committee. Name what is out in the brief. Route every request through one door, as a short written change request to the project lead, whoever asks, the sponsor included. Write a five-line impact note in days and money, with the estimates coming from the people who would do the work. Let a named person decide: the sponsor for date, budget and scope, the lead below an agreed threshold. Then record the decision, including the option that was dropped.
Saying no is only one outcome. A good impact note often ends with "yes, if we drop X" or "yes, after launch", which keeps the date and the budget honest.
Example
In the first week of January 2027, two requests reach the lead of Klasna's physics launch through side doors.
A free trial lesson for families on the waiting list (from the marketing lead). Each lesson still pays the tutor 430 UAH and costs 31 UAH of video and AI, so 461 UAH that nobody pays for. One hundred trial lessons: 100 × 461 = 46,100 UAH, which is 88.65% of the 52,000 UAH left under the 600,000 ceiling, plus four days of the platform lead's work for a "trial" booking type in the month before opening.
A page where parents see their child's progress (from Leo, the sponsor, in the kitchen). This is the parent progress dashboard that was proposed as a must in October (14 likely days, 85,000 UAH) and cut: together with video solutions it would have put the musts at 159 likely person-days against 150 available, 9 over.
Both go through the same door: a written request, an impact note, a decision by Leo within two working days, and an entry in the decision log. Leo decides: no free trial before opening; look again in March with real booking data. Without the door, both would have entered the work with the same date and the same ceiling.
Common mistakes
- Freezing all change. Creep is unpriced change; a freeze just pushes requests into side doors.
- "It's small, just do it". Small additions add up; size each one in days and money before saying yes.
- Letting the sponsor skip the door. The sponsor's requests need an impact note most, because people say yes to them fastest.
- Pricing only money. In the last weeks before a fixed date, a few days of a key person cost more than the cash.
- No record of what was dropped. An idea rejected without a log entry comes back as new a month later.