What You Are Actually Buying From Your Paid Media Agency
Marketing leaders evaluate agencies on case studies. The thing that determines the work is the operating model. Who is on your account and how they connect.
How a paid media agency is structured determines what it can produce, and most marketing leaders are evaluating the wrong things. Case studies. Pitch decks. The credentials of whoever is in the room for the new-business meeting. None of those tell you what the actual work product will be three months in, when the senior team has rotated off the account and the work is being done by people the brand has not met yet.
The thing that determines the work is the operating model. Who is on the account every week. How they are structured. What they can see across channels. Whether anybody on the team is responsible for how the channels connect. Most agencies sell one of two operating models, and the difference between them is the difference between an agency that can deliver full-funnel paid media and one that cannot.
Most brands are buying without realizing they are choosing. The pitch is interchangeable across the category. The operating model is not. Once you know what to look for, the model the agency is selling is visible in the first thirty minutes of an evaluation meeting.
Most brands are buying an operating model without knowing it
The default agency operating model in 2026 is the specialist-per-channel model. A senior strategist sells the engagement. After contract signature, the account is staffed with channel specialists. The paid social specialist runs Meta and TikTok. The paid search specialist runs Google and Performance Max. A media planner sits at the top with light coordination duties. Each channel specialist is rotated across multiple accounts, anywhere from four to ten brands per specialist.
The agency does not call this an operating model. The brand does not see it as one. But everything that follows is downstream of this structure. How budgets are reallocated week to week. How reporting gets built. Who owns total revenue. What gets prioritized when channels conflict.
The other model, the one fewer agencies operate, is the pod model. A dedicated team is assigned to a single brand. The pod includes channel specialists, but it also includes a senior orchestration role whose entire job is the connections between channels. The specialists do not sit across multiple accounts. They sit on the brand. Their incentives align to brand-level outcomes, not channel-level KPIs. When a CMO evaluates an agency without understanding the operating model the agency runs, they are choosing between two structurally different products that look identical in the pitch.
The specialist-per-channel agency optimizes the wrong thing by design
The specialist-per-channel model has a structural problem the agencies operating it cannot easily solve. Every specialist is incentivized to optimize what they can see. The Meta specialist optimizes for Meta-reported ROAS. The Google specialist optimizes for Google-reported CPA. The retail media specialist optimizes for retail-attributed sales. None of these specialists see the brand-level revenue picture. None of them are accountable for how their channel interacts with the others.
The result is predictable. Each channel reports clean dashboards while blended ROAS deteriorates. Brand search gets cannibalized by Performance Max because nobody owns the conflict. Audience overlap drives up CPMs because two specialists are buying the same audience without coordinating. Upper-funnel investment gets cut because no one has the data to defend it across the channels it actually feeds. We see this pattern in nearly every audit we run for a brand considering switching agencies.
This is not an execution problem. The specialists are usually good at what they do. The problem is that the structure rewards channel KPIs and the brand needs a system. Replacing the social agency with a different social agency does not fix the org chart. The org chart is the issue, and the predictable mistakes that follow from a siloed structure repeat from one specialist agency to the next.
The pod model puts cross-channel accountability on one team
A pod is built around a single brand. The team includes channel specialists for paid social, paid search, retail media, and CTV / programmatic, but the dedicated structure is what enables the work. The specialists are not rotated across multiple brands. They sit on one brand long enough to build deep account knowledge. They have time to coordinate. They share a single P&L view of the work.
The role that makes a pod different from a dedicated team is the orchestration role. We call this role the Associate Director of Media. Their job is the connections between channels. Where Meta and Google are competing for the same audiences and what to do about it. When Performance Max is cannibalizing brand search and how to wall it off. How upper-funnel CTV spend translates into branded search lift two months later. Whether retail media is incremental to the broader paid program or pulling sales that would have happened on DTC.
The orchestration role does not just coordinate. It is accountable for the system. When blended ROAS deteriorates, the orchestration role is the one whose performance is tied to it. When upper-funnel needs to be defended to a CFO, the orchestration role builds the case. When channels disagree on budget, the orchestration role decides.
This accountability cannot be added on top of a specialist's existing channel responsibilities. A Meta lead who is also "in charge of cross-channel" is not in charge of cross-channel. The orchestration job is full-time or it is not real.
Cross-channel orchestration cannot be a part-time job
The most common pattern in the specialist-per-channel model is to add a media planner or growth lead who carries cross-channel responsibility on top of their other duties. In practice, that role gets pulled into the urgent channel issues every week and never reaches the cross-channel work. The cross-channel work requires sustained attention to data that does not show up on any single platform's dashboard. It requires running the triangulation between platform reports, blended ROAS, MMM outputs, and incrementality reads. It requires modeling scenarios for next quarter's budget. None of that happens in the gaps between channel meetings.
When agencies tell brands their senior strategist will stay involved or that there is a growth lead coordinating across channels, the brand should ask: how many other accounts is that person on? What percentage of their week is dedicated to this one brand? Who owns the cross-channel reporting and is it built fresh every week, or pulled from a template that the same person uses for five other clients? The answers separate agencies that have a real orchestration function from agencies that have a job title.
| Question | Specialist-per-channel agency | Pod-model agency |
|---|---|---|
| Who owns full-funnel revenue? | No one explicitly | Dedicated orchestration role |
| How many accounts is each specialist on? | 4-10 | 1 |
| Where is budget reallocated cross-channel? | When the brand asks | Weekly, by orchestration role |
| How is reporting structured? | By channel | By business question |
| Who diagnoses portfolio-level underperformance? | Whoever has time | Orchestration role accountable |
The questions that reveal an agency's operating model
Five questions surface the operating model in the first thirty minutes of an evaluation meeting. None of them are the questions agencies prepare for, which is part of why they work.
First, who is on this account every week, and how many other accounts are they on? The answer reveals whether the team is dedicated or rotated. Second, who owns blended ROAS and full-funnel revenue, and what is in their compensation tied to that ownership? If the answer is "everyone owns it" or "the head of media coordinates," there is no real ownership.
Third, walk me through how a budget reallocation between paid social and paid search happens. Who makes the call? With what data? On what cadence? The answer reveals whether the agency has a real cross-channel decision process or whether budget moves only when the brand asks. Fourth, show me the cross-channel reporting that goes to a current client. The structure of the report tells you what the agency is actually orchestrating against. A report organized by channel is built by a channel-specialist agency. A report organized by business question is built by a pod.
Fifth, when a channel underperforms, what is the agency's process for diagnosing whether it is a channel problem or a portfolio problem? The agencies that can answer this with specificity have the orchestration function. The agencies that cannot, do not. We have published the longer evaluation playbook for marketing leaders running a formal RFP.
The pitch is interchangeable across agencies. The operating model is not. Once you know what to look for, the model is visible in the first thirty minutes.
In-house, agency, hybrid: which model fits which brand
The operating model question does not resolve in favor of one structure for every brand. The right model depends on spend level, growth stage, internal capability, and how much of the strategic work the brand wants to retain.
In-house works when the brand has the spend to justify dedicated headcount, the recruiting capacity to attract senior talent in a competitive market, and the strategic clarity to direct the work without an external partner. Most brands underestimate the cost of in-housing because they model only the agency fees they will save and not the 12 to 18 months of ramp time, the tool stack, the measurement architecture, or the pattern recognition that agencies bring from working across many brands. In-house at scale is excellent. In-house at $5M to $30M of paid spend is usually expensive and slow.
Agency works when the brand needs strategic depth, cross-channel orchestration, and access to patterns from other accounts. The right agency for this is a pod-model agency, not a specialist-per-channel one. The reason brands hire agencies and then complain about agency work is usually that they hired the wrong operating model.
Hybrid is the default for most brands at $10M to $100M of paid spend. The brand keeps strategic functions in-house and outsources channel execution and measurement infrastructure to a pod. Hybrid works only if the in-house team treats the agency as part of the operating model, not as a vendor. When the agency is treated as a vendor, the in-house team takes back the strategic decisions but does not take back the cross-channel orchestration that the agency was supposed to provide, and the work plateaus. QRY's Impact Methodology is built around the pod plus orchestration structure for exactly this reason.
Frequently asked questions
- How do I tell if an agency is running a pod model versus a specialist model in a sales meeting?
Ask how many accounts each person on the proposed team is currently on. A pod-model agency staffs people one-to-one with brands. A specialist agency rotates each specialist across four to ten accounts. The answer is concrete and disqualifying. Agencies that hesitate or generalize, with answers like "it varies by role," are usually running specialist coverage with pod-style language.
- Is the pod model just more expensive than specialist-per-channel?
Per-hour, often yes. Per-outcome, usually no. The pod model produces fewer hours of total work but those hours are concentrated on one brand and include orchestration that the specialist model does not provide. Brands that compare on fee structure alone miss that they are comparing two different products.
- What about in-house teams that work with channel specialists at multiple agencies?
That is a common structure for brands at $30M to $100M in paid spend. It works only if the in-house team has its own orchestration function. If nobody in-house owns blended revenue, the structure has the same problem as a specialist agency: each external partner optimizes its own channel and nobody connects them.
- Does the operating model matter as much for smaller spend levels?
It matters at every spend level above the point where the brand is running multiple paid channels. Below $1M to $2M in paid spend, a single channel specialist running the account often works fine. Above that, the connections between channels start to matter, and the specialist-per-channel structure starts to underperform. The threshold is when the brand has more than one channel that genuinely contributes to revenue.
- What contract terms protect against the operating model going backwards?
Three. First, named individuals on the account by role, not titles alone. Second, the percentage of each named individual's week that is dedicated to the brand. Third, a quarterly review process that audits whether the staffing matches the contract. Without these, every agency reverts to its default model after the first quarter, regardless of what was sold.
The agency a brand chooses is the operating model that brand is choosing. The pitch deck rarely explains this, and the case studies almost never do. The way to evaluate is by asking what is on the account every week and who owns the connections between the channels.
If your team is reviewing agency proposals on Friday, the questions to bring to those meetings are not about the work the agency has done before. They are about how the team will be staffed and who is accountable for the parts of the program that no individual specialist can see.
Get smarter about paid media
Strategy and data for senior marketers. No spam.
Founder & CEO
Samir Balwani is the founder and CEO of QRY, a full-funnel paid media agency he started in 2017. He has 15+ years of advertising experience and previously led brand strategy and digital innovation at American Express. He writes on paid media strategy, measurement, and how agencies should operate.
