Agency Advisory

Whether the agency relationship is structured to work — and whether you could leave it if it did not.

Overview

How I approach this work

Agency disputes are rarely arguments about effort. They are arguments about structure: the agency defines the metrics, reports on its own performance, and often holds the ad accounts, the pixel, and the historical data, so the client has no independent way to judge the work and no clean way to walk away from it. That is not a character problem, it is a contract-and-measurement problem, and it is fixable. I have spent years on both sides of this relationship — client-side running retainer work across paid search, paid social, email, and CRO, and as an operator selecting and managing external partners. Engagements are project-scoped rather than ongoing retainers, because the point is to leave you able to run the relationship yourself.

The diagnostic comes before any verdict. What was the agency actually contracted to do, what is being delivered against that, what is being reported, and does any of it connect to revenue. That means reading the master agreement and the statements of work properly — deliverable definitions, notice periods, exclusivity, rate cards, and who owns the accounts, the creative files, the tracking, and the data when the relationship ends. Ownership is usually the most consequential clause in the file and the least discussed, because it decides whether leaving is a decision or an ordeal. Then the commercial model, which shapes behaviour more than any brief does: a percentage-of-spend fee rewards a larger budget rather than a better one, a deliverable-based retainer rewards volume of output, and a performance fee is only as honest as the conversion definition underneath it. Then measurement, read independently rather than through the agency reporting layer, because a scorecard the graded party built is not evidence. Then the people — who was in the pitch, who is actually on the account, at what seniority, and how much of the contracted time is genuinely landing. From there the work goes wherever the finding points: restructuring scope and governance so the existing agency can do better, splitting capabilities between in-house and specialist by cost and institutional memory, or running a proper search — brief, longlist, scorecard, references, and a paid trial before a full commitment. I work the other side of this too, delivering technical work behind agencies for their own clients, which is part of why the read stays neutral. The two failure modes are sacking the agency on day one and defending indefensible spend, and both come from deciding before looking.

Deliverables

What a typical engagement produces

Concrete artifacts from this kind of work.

Contract and commercial review: Master agreement and statements of work read against what is actually being delivered — deliverable definitions, notice and termination, exclusivity, rate cards, and any markup on media, tooling, or production that is not visible on an invoice.
Account and data ownership audit: Who holds the ad accounts, the pixels and tags, the creative files, the analytics property, and the historical data — established in writing, with a plan to move anything held in the wrong name before it becomes leverage.
Independent performance diagnostic: Results measured outside the agency reporting layer, against the conversion definitions the business would choose rather than the ones that flatter the account, with explicit callouts for what is working, what is not, and what is currently unmeasurable.
Fee model and cost benchmarking: What the relationship costs against what the equivalent in-house function would cost, and what the fee structure quietly incentivises — because a percentage-of-spend model and a deliverable retainer pull in different directions before anyone writes a brief.
Scope and delivery reconciliation: Contracted deliverables lined up against delivered work over a real period, including who was staffed on the account at what seniority versus who appeared in the pitch.
In-house and outsource split: A written recommendation on which capabilities belong inside the business and which are better held by a specialist, based on cost, volume, speed, and where institutional memory needs to live.
Agency search and selection: A brief worth responding to, a longlist and shortlist, a scoring framework applied consistently, reference calls that ask the useful questions, and a paid trial project before a full commitment.
Transition and onboarding plan: The sequence that makes a handover survivable — access and data migration, knowledge transfer, overlap period, and the first ninety days written down so a new partner is productive rather than merely appointed.
Governance and reporting cadence: The operating structure around the relationship: a single accountable owner on your side, a KPI framework tied to outcomes the business cares about, review cadence, and an escalation path agreed before it is needed.

First thing I check: Who owns the accounts and the data
Performance read: Independently, not from agency reporting
Engagement model: Project-scoped, not a retainer

Related areas

Other parts of my practice that overlap with this one.

Performance Consulting

Paid acquisition measured on what it returns, not on the numbers the platforms grade themselves with.

Brand Consulting

Positioning first, identity second, and a system that survives contact with production.