Scope, Change Orders and Reporting Process for Agencies

Agency margin leaks in two places, and delivery is neither of them: the scope written before anyone had properly read the brief, and the second round of revisions nobody logged as a change. I begin by reading your last ten signed statements of work against the hours actually booked to them. On the three most recent engagements of this kind, the gap between hours sold and hours worked ran between nineteen and forty-one percent, and almost none of it was ever raised with the client as a change. What the read finds: Which client, which service line, which account lead. Usually one of the three explains most of the loss. The cause is rarely incompetence. It is one scope sentence phrased so that a reasonable client and a reasonable account manager read it differently, then repeated across every engagement because the template carried it forward untouched. The five documents: An intake brief listing the questions that must be answered before any number is quoted, and the rule that no number leaves the building until they are. A statement of work with an exclusions section that is actually filled in, the section most agencies leave as a heading and the one clients quote back at you when a dispute starts. A change-order trigger written as a threshold rather than a judgement call, so a mid-level person can raise it without asking permission first. A weekly client report showing work against scope rather than activity. An escalation path with names in it and a stated response time. Installed where you already work: Whatever you run now, whether Harvest, Float, ClickUp, Notion, HubSpot or a shared drive. I do not recommend software and I take no referral income from tool vendors. A process that requires a migration will not survive the migration. Two weeks of shadowing while the first live briefs go through the new intake, because the first three uses decide whether a process sticks or gets quietly abandoned. Refusals: No pitch decks, no new-business collateral, no positioning work for the agency itself. No pricing recommendation before I have seen at least two quarters of utilisation data; anyone giving you one without it is guessing with your revenue. I will not write process for an agency under roughly eight people. Below that, partners hold it in their heads faster than a document can hold it, and formalising it makes them slower. Not for: Agencies where a founder personally approves every statement of work and intends to keep doing so. The documents will be correct and unused, and you will have paid to have a bottleneck described rather than removed. Also not for in-house teams. The client and supplier boundary is the thing being repaired here, and an in-house team does not have one; that work is a different engagement. What the partners have to do: Sign the exclusions list personally. It is the only document that costs you revenue in the short term, because it makes you say no to work you have been absorbing, and it is the one that stops the bleeding. Every engagement of this kind reaches a moment where a partner wants an exception for a favourite client. Granted twice, the process is dead and the fee was wasted, and I will say so at the sixty-day review rather than at the end. Reviews: At sixty days and again at a hundred and twenty, I read every change order raised since installation. A period with no change orders raised at all is not success. It means the trigger sits at the wrong number, or nobody trusts it enough to pull.