Every website project changes as it goes. Some of that is healthy: you learn things. Some of it is a project quietly becoming twice the size for the same fee. Telling them apart is what keeps the relationship intact.
Scope creep is work added after the scope was agreed without adjusting the price or timeline. It is best handled by defining the scope precisely up front, including what is excluded, and agreeing a change process before you need one. Most disputes come from things nobody discussed rather than from bad faith.
What causes it
- A vague brief. If the scope says a modern website, everything is arguably included.
- Undefined page counts. A few pages becomes eleven.
- Just one more thing, repeated across weeks until it is another project.
- New stakeholders arriving mid-project with fresh opinions.
- Discovery revealing genuine complexity, which is legitimate and still needs re-agreeing.
- Unlimited revisions as a selling point, which has no natural end.
Reasonable change or creep?
| Situation | Which |
|---|---|
| Copy longer than planned needs a layout adjustment | Reasonable |
| Adding a booking system not in the brief | New scope |
| Fixing a bug in delivered work | Included, always |
| Fourth round of revisions after three were agreed | New scope |
| A page added because a service was missed | Usually reasonable, small |
| Multi-language after a single-language quote | Substantial new scope |
The test is whether a reasonable person reading the original scope would have expected it. Bug fixes always fall inside. New capability does not.
Preventing it
- Specify the page count exactly, listing the pages.
- State the number of revision rounds included.
- Write down what is explicitly excluded. This is the most useful and most often skipped part of any proposal.
- Name one decision-maker.
- Agree the change process before you need it, including how changes get priced.
- Sign off each stage in writing so it does not reopen.
Handling it once it starts
For the client: ask what this does to the timeline and the price before saying yes to your own idea. Most people would happily drop a request if they knew it added two weeks.
For the agency: flag it immediately and in writing, with the cost and time impact, then let the client decide. Absorbing changes silently builds resentment and trains the client to expect it. Raising it three months later is worse.
When absorbing it is right
Small changes, genuine oversights on your side, and anything that materially improves the outcome for a trivial cost are worth absorbing without a conversation. Nickel-and-diming every adjustment damages a relationship worth more than the invoice.
The distinction is size and pattern. One small favour is goodwill. Ten is a project you underpriced and are now paying for.
Frequently asked questions
What is scope creep?
Work added to a project after the scope was agreed, without a corresponding change to the price or timeline. It is usually gradual, and it usually comes from things nobody discussed at the start rather than from anyone acting in bad faith.
How do I avoid scope creep on a website project?
Define the scope precisely, list the exact pages, state how many revision rounds are included, write down what is excluded, name one decision-maker, and agree a change process before you need it.
Should agencies charge for every change?
No. Bug fixes and small adjustments should be absorbed as a matter of course. Charging for every minor tweak damages the relationship. What should be charged for is new capability and additional rounds beyond what was agreed.
What are unlimited revisions?
A selling point with no natural end, which usually means either the agency has priced in a lot of margin or the project stalls indefinitely. A defined number of rounds with a clear price for extras serves both sides better.
Who is responsible for scope creep?
Usually both. The agency for a vague proposal that did not state exclusions, and the client for requests that were never checked against the scope. Preventing it is mostly about writing things down before anyone is invested.