Scope drift: why the agreed scope and the invoiced scope diverge
Scope drift is usually a records failure before it is a people failure: the agreed scope lives in a chat or a PDF while the work and the invoice live somewhere else. What a structured quote has to contain, why the fix is a new quote rather than an edit, and where quoted against billed still lies to you.
Most agencies do not lose money on the work they agreed to do. They lose it on the work nobody wrote down. A line added in a call, a revision round that ran to four, a format nobody mentioned — each reasonable on its own, and collectively the reason a project sold at one number was invoiced at another. Calling that carelessness or dishonesty is usually wrong, because it treats a records failure as a character failure.
Scope drift is structural before it is moral. The agreed scope lives in one kind of place — a chat thread, a PDF, a paragraph in an email — while the delivered work lives in a task list and the invoice lives in the billing system. Those artefacts share no structure, so there is nothing to compare. This is an argument about where the agreement lives and what it must contain.
Why a scope agreed in a chat or a PDF cannot be compared against anything
A PDF is a rendering of a decision, not the decision: prose with a number at the bottom, carrying no field anything else can read. A comparison needs two things that line up — a set of items and a set of amounts — and prose gives you neither. “A landing page with a CMS, responsive, three rounds of revisions” holds a scope and a price, but no quantity, no rate, no tax treatment, and nothing a task or invoice line can attach to. So when the fourth revision arrives, the honest question — inside the agreement or outside it? — has no answer, only an opinion. Hold the same agreement as a structured quote with items, tax treatment, terms, and a send state, and the quote says three, the delivery record says what was done, the invoice says what was billed. The gap becomes a fact rather than a memory.
Where the agreed scope lives, and what each location can be compared against (structural comparison — no measured outcomes)
| Where the agreement lives | Comparable against delivered work and billing? | What drift looks like when it cannot |
|---|---|---|
| A message thread or email | No — no items or amounts to line up | The addition is remembered by whoever agreed it and forgotten by whoever prices it |
| A PDF proposal | No — no quantity, rate, or tax field to match | A delivered item cannot be shown to sit inside the quote, so it is absorbed silently |
| A structured quote record | Yes — each quoted line reads against work and billing | The gap is a number on a project, visible while it can still be priced |
What a structured quote has to contain before the comparison is possible
Four things, each load-bearing. Line items with a quantity and a rate, because a scope you cannot count cannot be compared with work you also cannot count. Tax treatment, because two totals that match can cover different taxable values, and comparing them without the breakup invites an argument you cannot win. Terms, because “overdue” is meaningless without the date the money was promised. And a send state, because a quote never sent is an internal draft, not an agreement. The quote, the invoice, and the payment are one commercial history seen at three moments, and the moment that goes missing is the first one.
How a scope becomes a task list, and where the divergence enters
Delivery is where the translation happens. A won opportunity becomes a project carrying the client, the scope, and the commercial history; the scope breaks into task lists with priorities and dates; each task has exactly one owner, because a task assigned to a team is a task nobody is behind. Status is read back from the tasks actually done, and finished work becomes an invoice without a second person re-deriving what was delivered. A good chain, and also where the loss happens. A task is not a line item. The list gets re-cut by the person doing the work, who is not the person who set the rate — for good reasons, since the rate-setter is rarely in the room when the client changes their mind. When a request arrives as “one more round, this time a different format”, something has to become a task or the work has no home. A task appears tracing to no quoted line, done well by the same people who would have done the quoted work. The record is not lying: it is showing a task no quote described, and the invoice inherits the ambiguity.
Four ways the delivered list diverges from the quoted list (structural, not measured behaviour)
| Divergence | What the records show afterwards |
|---|---|
| An addition — a request arrives that nobody quoted | A completed task with no quoted line behind it, and a billed total above the quoted total |
| A re-cut — the delivery owner regroups the list and adds implied acceptance steps | Individually reasonable tasks, collectively larger than the quote: drift with nobody overstepping |
| A reduction — the client drops a piece | A smaller completed list and a lower invoice, which reads as a loss until the agreed drop is noticed |
| A never-invoiced item — delivered, no document raised | A fully delivered project with billed below quoted, and no task-list gap to explain it |
Quoted against billed turns an anecdote into something you can price
One project cannot tell you anything. Everyone involved remembers it as a success or a disaster, and memory is not evidence. Read a quarter of them, each carrying what was quoted against what was billed, and the argument changes character: you are looking at the same type of work priced the same way each time. That is where drift stops being a complaint about a person and becomes a pricing decision — add an acceptance step to the standard quote, quote revisions as a line, or move the rate. The comparison is only as good as the quoted state behind it, though. If the quote was raised as a proposal document there is no quoted value in the system, and the honest conclusion is that this profitability question cannot be answered from here — a process fix to make first, not a reporting problem to solve later.
Why the honest fix is a new quote, not an edit to the old one
There is a version of this story where the original quote is quietly updated to include what got added. We do not recommend it. The original is what the client accepted, and often the copy they still hold; edit it and your two versions differ, which is a harder conversation than the one you were avoiding. It also destroys the only record of what you under-quoted, and a revised total tells the delivery team nothing about which part of the work is now being paid for. The pattern we can stand behind is narrower: raise a new quote for the additional scope, get it accepted, attach it to the same project so the economics show both pieces. A reduced scope works the same way in the other direction. Be clear about what this is not. There is no change-request record, no milestone record, and no deliverable record in NoxOrigin — no contract document editor, no redlining, no e-signature. If your buyer wants a signed change order, that document belongs in a contract-lifecycle tool built for it. What the operating record can do is make the commercial half of the change priced, accepted, and attached to the project that absorbed the work.
What a scope change looks like when it is a record
- The additional work is quoted as its own quote, with its own items, tax treatment, and terms
- The original quote is left exactly as it was sent and accepted
- The new quote attaches to the same project, so quoted against billed shows both pieces
- The invoice for the extra work comes from the new quote, so the client sees which agreement it belongs to
- No quote is back-dated so the addition can look like part of the original agreement
The words this article keeps separate
- Agreed scope
- What the two sides committed to, in a form that can still be read later.
- Quoted value
- The accepted quote total, as a structured record with items, tax treatment, terms, and a send state.
- Drift
- The gap between the accepted quote and what was delivered and billed — a measurement of comparability, not of honesty.
What we would need to count before saying anything about drift
Without these, any statement about scope drift is opinion
- How many projects carry more than one accepted quote, and how many of those are additions
- The distribution of billed against quoted per project, grouped by work type
- Projects where the delivered list grew after work started, and how often that became a second quote
- Projects that closed with completed tasks and no billing event against them — agreed work never invoiced
What we have not measured
To be explicit: we have no measured drift rate, no measured distribution of billed against quoted, no measured cost base, and no customer data. Nothing here reports what agencies actually do. It argues about the structure that makes the comparison possible, and lists what would have to be counted before a claim about behaviour was more than an assertion. Quoted value exists only where quotes were raised as records. If a number about scope creep is offered to you as a benchmark rather than as a reading of your own records, ask which records produced it, over what period, and how work type was handled.
Frequently asked questions
Is scope drift just a discipline problem?
Mostly not. It is a comparability problem. When the agreed scope is a chat thread or a PDF, there is no object to put the invoice next to, so nobody can say whether a delivered item was inside the agreement. The argument then becomes about character instead of about records, and it gets settled by whoever holds the conversation. Raising the agreement as a structured quote is what makes the question answerable at all.
What should a quote contain so it can be compared with what was billed?
Line items with a quantity and a rate, an explicit tax treatment, payment terms, and a send state. The first lets a quoted line be matched to delivered work; the second stops a like-for-like total comparison hiding a different taxable value; the third gives the word overdue a date to hang on; the fourth separates an agreement the client accepted from a draft you never sent.
Should I edit the original quote when the client adds work?
No. The original is the document the client accepted and often the copy they still hold, and editing it destroys the record of what you under-quoted. Raise a new quote for the additional scope, get it accepted, and attach it to the same project. Both quotes then sit on the project and the project economics show the whole engagement.
Does NoxOrigin have a change-order or change-request record?
No, and we would rather say so than imply otherwise. There is no change-request record, no milestone record, no deliverable record, no contract document editor, no redlining, and no e-signature. The pattern we support is a new quote attached to the same project, which covers the commercial side of a change. If you need a signed change order for your buyer, that belongs in a contract-lifecycle tool.
What does quoted against billed actually tell me?
That the work which left the door stayed, or did not stay, inside what was sold — per project. It is a delivery observation, not a financial statement: it says nothing about whether the work was worth doing, whether the rate covered the cost, or whether the relationship was profitable. We do not capture delivery time or a cost base, so it cannot be read as margin.
Can the system tell me whether the extra work was worth doing?
No. There is no delivery time capture and no cost recorded against a task, so worth is a judgement rather than a reading. Quoted against billed is useful for the part you can answer — should the next quote for this kind of work carry an acceptance step, a revisions line, or a different rate — and you supply the value judgement.