The agency creative workflow at 15 brands: where the week actually goes

Fifteen brands, one creative team. On paper it's a portfolio; in practice it's a queue. Every brand on the roster expects the same thing this month as last month — fresh concepts, backed by something real, delivered in a deck that looks like you thought hard about their business specifically. Your team knows how to do that for one brand. The problem is arithmetic: the loop that produces good creative for one brand doesn't compress just because you're running it fifteen times. This guide traces where the week actually goes, and what to do about it.

The loop, one brand at a time

Strip away the meetings and the Slack threads, and the work that produces a round of concepts for a single brand looks nearly identical at every agency:

• Pull performance. Export from Meta, Google, wherever the brand runs. Reconcile naming conventions, dedupe the variants, get it into something a human can read.
• Read it. Which creatives won, which died, which are quietly fading. Not the campaign totals — the creative-level story: hooks, formats, tones.
• Check rivals. Skim the ad libraries for the brand's competitors. What are they running more of? What did they kill?
• Assemble the brief. Objective, audience, what worked last round, what to avoid, mandatories. Half of this step is archaeology in old decks and email threads.
• Write the concepts. The actual creative thinking — the part everyone got into this business for.
• Format for the client. The deck, the rationale slides, the brand's template, the send.

None of these steps is hard. All of them take real hours, and only one — writing the concepts — is the job your clients think they're paying for. The rest is reading and assembly: necessary, repetitive, and invisible in the final deliverable.

Now multiply by fifteen

Karl Sakas, who advises agencies on operations, benchmarks a dedicated strategist at 8–12 clients — and as few as 2–3 when accounts are high-touch. Most teams running 15 brands aren't staffed the way that math suggests. Which means the loop above doesn't actually run fifteen times a month at full quality. It can't.

So a quiet triage happens, and nobody ever announces it. The three retainers who email the most, spend the most, or scare you the most get the full loop: fresh performance reads, real competitive checks, concepts written from zero. The other twelve get the compressed version — last month's deck with new dates, a brief assembled from memory, concepts that rhyme with the previous round because nobody had time to find out what changed. The quiet brands aren't getting worse work because anyone decided they matter less. They're getting it because attention is the scarce input, and squeaky wheels set the allocation.

At fifteen brands, quality doesn't follow the retainer. It follows the noise.

That's the churn risk hiding inside a healthy-looking book. The quiet brand that leaves rarely complains first — the deck just stopped being about them, and eventually they noticed.

The workflow patterns that actually help

Before any software enters the picture, three patterns move the needle. Agencies that hold quality at scale tend to run all three.

Batching first. Running the same step across many brands beats running the whole loop brand by brand — pull all fifteen performance reads on Monday, do the competitive skims Tuesday morning, write concepts in protected blocks. Gloria Mark's research at UC Irvine found it takes people about 23 minutes to fully refocus after an interruption, and the loop as most teams run it is one long chain of interruptions: platform tab, spreadsheet, ad library, deck, Slack, repeat. Same-step batching keeps your head in one mode long enough to be good at it.

Templates second. One brief format across the whole book, one deck skeleton per client type. The point isn't uniformity for its own sake — it's that a strategist opening brand number eleven at 4pm shouldn't spend the first twenty minutes deciding what a brief looks like.

Decision logs third, and this one's the sleeper. Write down what you tried, what won, and why you believed it — per brand, in one place. Without it, the learning from every round evaporates into whoever happened to be in the room. Six months later someone re-pitches an angle that already failed, or kills a direction that was quietly working, because the reasoning lived in a strategist's head and that strategist is on another account now.

Where AI genuinely fits — and where it can't

Look back at the six steps. Three of them are reading and assembly: pull performance, read it, check rivals. One is formatting. Those four are automatable in the honest sense of the word — a machine can group ads by creative, grade each one against the brand's own KPI, sweep the rival ad libraries overnight, and assemble the brief with what changed since last round already filled in. That's the layer AgentMark runs: it reads your live accounts read-only, grades every creative against each brand's bar, watches competitor libraries and buyer personas, and drafts the concept round in your brief format — with each concept carrying a receipt for which winning pattern grounded it.

The other two steps don't automate, and you should distrust anyone who claims they do. Taste — knowing which of eight decent concepts is the one — is your team's job. Client trust gets built in rooms and on calls, not in software. And the final cut is a human call every time, because a machine can tell you what wins; it can't tell you what will. Nobody can. The honest claim is always "this direction wins right now," never "this will work." The value of automating the reading isn't that it replaces judgment — it's that judgment finally gets applied to all fifteen brands instead of three.

Ship the rationale with the work

There's a client-facing layer here that most agencies underinvest in. A concept delivered without its reasoning invites a taste debate: the client's CMO doesn't love the hook, and now you're defending creative on vibes. The same concept delivered with its grounding attached — here's the pattern in your own account this is built on, here's what it beat — changes the conversation from "do we like it?" to "do we believe the read?" You'll still lose some of those conversations. You'll lose fewer.

It also pays off at renewal. An agency that ships every round with its rationale, into one place the client can revisit, walks into the renewal conversation with a year of cited creative decisions on file. AgentMark's version of this is a white-labeled client portal where approved work lands with its grounding attached — but the principle holds whatever tooling you use: the rationale is part of the deliverable, not an internal artifact.

The compounding effect of a system that remembers

Here's what changes when the reading, the assembly, and the memory are systematic instead of heroic. Round one for a brand starts from a real read instead of a recollection. Round three starts from what rounds one and two proved. The decision log writes itself, because every graded creative and every shipped concept is already on the record. And the quiet twelve stop getting last month's deck, because a fresh read for brand number fourteen no longer costs a strategist's Thursday.

The loop doesn't disappear — your team still briefs, still writes, still makes the final cut, still sits in the rooms where trust gets built. What disappears is the version of the week where most of the creative hours go to finding out what happened instead of deciding what to do about it. At fifteen brands, that's the whole game.