Use case
Catch scope creep while you can still do something about it
Scope creep is almost never one decision. It is eleven small yeses, each one reasonable on the day, none of them written down. Then there is a conversation at the end of the project where the client is certain it was included and you are certain it was not. Spectr flags this work as it comes up and keeps a record of what you decided about each one.
You are both being honest. That is what makes it so bad
The client genuinely remembers the reporting export being discussed as part of this phase. You genuinely remember ruling it out. It happened out loud, in a call eight weeks ago, and the only record is a transcript nobody has opened since. So a scope question turns into a trust question, and that costs far more than the feature would have.
What it actually looks like
A £60k fixed-fee build. Twelve weeks. You are in week five.
Week 5 call
Near the end, the client asks whether the dashboard could also export to Excel. You say you will look at it. Everyone moves on.
Same day
You import the call. Spectr flags the Excel export as outside agreed scope, with a confidence score, and links to the exact part of the transcript.
Same day
You mark it confirmed out of scope and defer it to phase two. That takes about fifteen seconds.
Week 6
Three more small asks come up across two calls. Two are genuinely in scope, so you dismiss those flags. One is not, and you approve it back in with a change order.
Week 11
The client asks where the Excel export is. You open the record: raised week five, agreed as phase two, here is the transcript.
Week 11
It is a two-minute conversation. Phase two gets quoted instead of absorbed.
Four scope asks that would have gone unrecorded have turned into four decisions logged in under a minute each - which is the difference between quoting phase two and doing it for free.
Flagged when it comes up, not at the end
Spectr reads each meeting against what the project actually agreed and flags work that falls outside it. Each flag carries a confidence score, so a borderline call looks borderline instead of being presented as a finding. Timing is the whole point. An ask raised the week it happened is a conversation. The same ask raised at handover is a dispute.
Every flag gets a decision
Confirm it as out of scope, push it to a later phase, dismiss it as a false positive, or approve it back in. Nothing sits in a maybe state, and the record of who decided what is there when someone asks in week eleven.
The evidence is the transcript, not your memory
A flag points back to the part of the meeting it came from. That is what turns "we agreed this was out" into something you can show someone. It also keeps the conversation about a record rather than about who remembers better, which is the difference between a change order and a fight.
On every plan, including free
Scope tracking is not behind a tier. It is available on every plan including Free, because the projects most exposed to scope creep tend to be the smallest ones with the least room to absorb it.
Works with
Frequently asked questions
How does Spectr know what is in scope?
From the project's requirement documents and the spec sheets you already agreed. A tagged requirement version is the strongest signal, which is a good reason to tag one at sign-off.
What if the flag is wrong?
Dismiss it. The confidence score exists so a marginal flag looks marginal - you should be overruling it sometimes.
Can I show a client the scope record?
The record lives in Spectr with its history intact, and requirement documents export to PDF. There is no client-facing view yet.
Does this work if we do not do fixed-fee work?
Yes, though the value changes. On time-and-materials the record is less about protecting margin and more about explaining where the hours went.
Also worth reading
See it on your own project
Free plan available - five spec sheets a month, two seats, no card required. 30-day free trial on Starter.
Free plan available. 30-day free trial on Starter.