Agency management: foundations & core concepts

The invoice says "agency fees, Q3". Ask four people around the table what that money bought and you get four answers: ads, thinking, extra hands, speed. Agency spend is one of the largest lines a CMO controls, and it is the line most often defended with the least precision. Before anyone can manage an agency well, someone has to be able to say what is being bought, in what shape, on what commercial terms, and for how long. That is the ground this lesson covers.

What agency management actually means

Agency management is the discipline of designing, operating and adjusting the relationship between a brand and its external creative, media, strategy or production partners. It covers four things: the shape of the roster, what sits inside each scope, how the partner gets paid, and how the work is judged. None of it is administrative. It is organizational design applied to people who do not report to you, and it sets the quality of most customer-facing work your brand puts out.

The word "agency" covers a wide spectrum. A full-service agency handles strategy, creative and sometimes media under one roof. A creative agency sells ideas and the craft to make them. A media agency plans and buys inventory, and earns a good part of its money on the volume it places. A production company executes what someone else designed. Specialists own single terrain: CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition →, influencer, retail media, packaging design. The large holding companies (WPP, Omnicom, Publicis, Dentsu, IPG) own agencies in every one of those categories, which is why the same parent can show up three times on a single roster with three different logos and three different rate cards. Treating all of those types as if they staff, price and think the same way is the most common and most expensive error in this discipline.

The four foundational concepts every CMO must own

*1. Roster shapes*

Rosters come in a handful of recognizable shapes. The agency of record (AOR) model gives one agency the brand across most or all output on a continuing contract, and expects it to know the business almost as well as the client does. Toyota has run its US advertising through Saatchi & Saatchi since the mid-1970s, one of the longest client-agency relationships in the industry. The dedicated unit takes that further: the agency builds a team that works on one client only, often with its own name and address, as TBWA did when it created Media Arts Lab for Apple. A project roster keeps no continuing appointment and briefs work out in sprints to whoever fits the task. A hybrid keeps an internal studio for volume and iteration while holding outside agencies for the bigger platform ideas.

Each shape buys something different. AOR buys accumulated context and one voice. The project roster buys optionality and speed, at the cost of re-explaining your business every time. The dedicated unit buys context plus confidentiality, at a price only a very large and very consistent spender can carry. Which shape suits your situation is a decision with its own selection criteria, handled elsewhere in this module.

*2. What you are buying, and the brief that names it*

Strip the relationship back and an agency supplies four things: senior talent whose cost is spread across several clients, an outside read on a business you are too close to, production capacity that flexes with your calendar, and in media, buying scale you do not have alone. Everything else on the invoice is a variation of those four.

The document that converts them into work is the brief. Not the contract, not the scope of work: the brief. Mark Ritson has argued for years that most client briefs are not briefs at all but wish lists, and the diagnosis holds. A real brief names one audience with specificity, one thing that audience currently believes or does, the shift you want, the constraints that are genuinely fixed (budget, channels, legal, timing), and how you will both know it worked. Nike's "Just Do It", written at Wieden+Kennedy in 1988, came out of a conversation about widening Nike past competitive runners to anyone who exercises at all. A business problem went in and a line came out. That order is set by the brief, and it is the client's job, not the agency's.

*3. Scope of work and fee models*

The scope of work turns the brief into a commercial object: deliverables, volumes, timelines, revision rounds, the named team and the fee. Four fee models cover almost everything you will sign:

• Commission on media billings, historically 15 percent, now rare in creative and heavily discounted in media.
• Retainer, usually built bottom-up from full-time equivalents: this many strategists at this share of their time, priced on agreed salary bands with an overhead and profit multiple on top.
• Project fee, a fixed price for a defined piece of work with a defined end.
• Performance-linked money, a bonus pool sitting on top of a reduced base and released against agreed metrics.

Most large advertisers now run a blend, and the blend is where the argument lives. The failure mode is scope creep: the slow expansion of what the agency does with no matching adjustment to the fee. That erodes agency margin, and margin decides who is assigned to your account. Senior creatives and strategists get moved off unprofitable business quietly, and you find junior teams working on your hardest brand problem. Write the revision rounds, the attendance list for reviews, and the line between an amend and a new project into the SOW itself, because those three ambiguities cause most of the drift.

*4. Performance frameworks and accountability*

Agencies behave according to what you measure. Count outputs (ads produced, launches per quarter) and you buy volume. Measure outcomes (consideration lift, cost per qualified lead, incremental sales) and you buy accountability, but only if you have built the measurement infrastructure first. Industry compensation surveys have shown performance-linked fees in a large share of major advertiser contracts for years now, with the design varying wildly: some tie to business results, some to a subjective quality score, some to both. A usable scorecard mixes business results, quality of thinking, and delivery discipline. The better ones run in both directions, with the agency scoring the client on brief quality, decision speed and how often approvals get reopened. If your side of that scorecard is bad, the fee model will not save you.

Real-world cases with actual results

Apple and TBWA are the reference case for what an AOR relationship becomes at its best. Chiat/Day made "1984"; TBWA\Chiat\Day made "Think Different" when Steve Jobs returned in 1997; Media Arts Lab was then built as a unit working on Apple alone. It works because Apple supplies unusual strategic clarity (one product story at a time, no discount positioningpositioningThe mental space you want your brand to occupy in your target customer's mind relative to alternatives.View full definition →, ever) and the agency keeps a small senior team carrying years of accumulated brand context. "Get a Mac", from 2006 to 2009, ran through a period when Mac sales grew considerably faster than the wider PC market. That output came from a stable, deeply briefed relationship rather than from a competitive pitch.

Toyota and Saatchi & Saatchi show the other value of longevity: the agency understands a structure most outsiders never see. US automotive advertising runs on three tiers (national brand, regional dealer associations, individual dealers), each with different money, approvals and timing. An agency five decades into that account does not need the tier explained, and it knows which model cycle a launch is landing against. Context of that kind is expensive to rebuild and is the main thing you write off when you re-pitch.

The counter-case is the roster nobody has counted. Several global advertisers, on first audit, discovered four-figure numbers of marketing suppliers across their markets, most engaged locally with no shared brief, terms or performance data. The cost is not only duplicated fees; it is that no single partner holds enough of the picture to be useful strategically.

CMO action items

• List every agency and production supplier you pay, with its fee model, annual spend and the one business outcome it is accountable for. If any line has no outcome next to it, you have found your first cut.
• Rewrite the standard brief template so it forces one audience, one behavioral shift, the real constraints and a success measure onto a single page. Run it live on the next campaign and compare first-round creative against your last three.
• Reopen your largest SOW and check three things: revision rounds, the definition of a new project, and whether the named senior team in the contract is the team actually in the room.
• Set a twice-yearly review that scores partners on results, thinking and delivery, and that lets them score you back on brief quality and decision speed.

Common mistakes that kill results

• Treating the agency as a production resource. Bring them in only at execution and you waste the thing you are paying most for: an outside read on the business problem. Agencies briefed on the commercial challenge, rather than on a creative task, produce work of a different order.
• Managing the relationship through procurement alone. Cost optimization and quality optimization are separate conversations that need separate owners. Agency talent follows margin. Compress it too far and your senior creatives are on another account within two quarters, with nothing in the contract broken.
• Changing the brief after creative development has started. A pivot two weeks into concepting is not a course correction, it is a write-off, and it is the largest single source of wasted agency money. Fight the alignment battle internally before the brief goes out, not after the work comes back.