Marketing Procurement's Role in Agency Selection: Who Owns Which Decisions?

Agency selection rarely collapses because one function seized control. It stalls because nobody wrote down who decides what at each stage. AAR Group, which advises brands on agency reviews, notes that difficulty tends to appear when roles are left undefined at the outset and creative judgement and commercial judgement are argued as two separate debates instead of parts of a single decision (AAR Group). This article maps marketing procurement agency selection stage by stage: which decisions marketing should own, which procurement should own, who is consulted, and which deliverable closes each one. It then works through the two questions that produce the most friction in practice, how to sequence capability assessment ahead of pricing and how to score value rather than cost, before covering the handoff after signature. If you are building a longlist while sorting out internal ownership, the Growth Partner Index directory shows how documented agency criteria can anchor that first conversation.

TL;DR: Who Owns What in Marketing Procurement Agency Selection

• Marketing owns the brief, the strategic and creative fit judgement and the final recommendation. Procurement owns commercial structure, benchmarking, contract terms and supplier risk. Both co-own the scorecard weights and shortlist criteria, agreed before any agency is contacted (AAR Group).
• The most common failure is undefined roles rather than a power struggle: creative and commercial judgement debated as separate arguments instead of parts of one decision.
• Assess team and strategic capability first, then open financials. Running both in parallel lets the cheapest proposal colour the capability read (TrinityP3).
• Score value against benchmarks and the cost of failure. Lowest-bid selection tends to raise long-term operating cost through quality and stability problems (Art of Procurement, RightSpend).
• The contract, rather than the pitch, carries promised quality into delivered quality, so procurement's contract work is a marketing outcome lever (Veza Network).
• Weighting shifts by agency type: creative selection leans on marketing judgement, while media and production selection carry heavier procurement-weighted commercial criteria.

GPI's view is that agency selection goes wrong when a function wins the argument rather than when the decision is made on evidence. Every claim an agency makes, whether about creative capability, media efficiency or data practice, should be judged against what it can document, how the result was produced and what it cannot prove. That standard is neutral between marketing and procurement, which is why it works as shared ground. It also means results-based terms and performance KPIs should be chosen to answer a specific business decision, and buyers should resist reading attribution as proof that the agency caused the outcome. Decision rights should follow the evidence each function is best placed to evaluate.

Why 'Procurement vs Marketing' Is the Wrong Frame

Two kinds of expertise, one decision

The two functions contribute different expertise, and neither substitutes for the other. AAR Group describes procurement's contribution as value, risk, governance, efficiency and supplier accountability, and the marketer's as judgement about the brand, the customer and the creative work (AAR Group). A selection that leans entirely on one set of skills will be either commercially exposed or strategically hollow. The useful question is which of those judgements each specific decision depends on most, because that determines who should hold the pen.

One caveat belongs here. Some marketing teams carry real commercial expertise of their own, particularly where a marketing operations or agency management role exists. The map in this article should be adapted to that reality rather than copied, and the adaptation itself should be written down.

Where the friction actually comes from

Friction between the functions is usually structural rather than personal. AAR's observation is that trouble arrives when roles are left undefined at the start, and the symptom is that creative merit and commercial merit get debated as two independent arguments, with each side defending its own (AAR Group). Once that split exists, every later stage inherits it. Marketing reads a low fee as a threat to quality, procurement reads a strong creative preference as a threat to leverage, and the final meeting becomes a negotiation between colleagues rather than a decision about an agency.

The cost of leaving roles undefined

Undefined roles are expensive in their own right. WBR's Procurecon commentary points out that selecting and managing agencies is already complex, and that without clear guidelines and policies the process becomes slow and inefficient (Procurecon). Weeks lost to internal disagreement carry a business cost that no fee saving offsets, and the source offers no quantification, so treat this as a directional warning rather than a measured figure.

The practical reframe is to ask who owns each decision and what closes it, rather than who owns the process. Before the next selection, hold a 30-minute session in which marketing and procurement each write down the decisions they believe they own. Compare the lists and log every disagreement. Those disagreements are the raw material for the RACI exercise in the next section.

The Decision Rights Map: A RACI for Each Selection Stage

The table below is a GPI working framework built from the cited advisory guidance, and it is presented as a starting point to adapt rather than an industry standard. It uses the RACI convention: Responsible does the work, Accountable makes and owns the call, Consulted gives input before the decision, Informed hears the outcome. Every stage ends with one named deliverable, because a decision without an artifact tends to get reopened.

The division of expertise in this table follows AAR Group's description of procurement's commercial and governance role alongside marketing's brand and creative judgement (AAR Group). The separation of stages 4 and 5 follows TrinityP3's rule that financial assessment should wait until confidence in the team and strategic process is established (TrinityP3). The emphasis on early joint definition of goals and timelines draws on 2023 guidance from Procurement Tactics, which is qualitative best practice rather than a measured outcome (Procurement Tactics).

Stage 1: Defining the need, the brief and success measures

Marketing is accountable for the brief because only marketing can say what the business needs the agency to change. Procurement is consulted at this stage for a specific reason: a brief that cannot be contracted or measured will cause trouble at stage 5. Procurement Tactics' 2023 guidance holds that the two functions get better results when they cooperate from the very beginning, defining goals, expectations, timelines and priorities together (Procurement Tactics). In practice that means procurement reads the draft brief and flags any success measure that an agency could not reasonably be held to, and marketing decides whether to sharpen it or accept that it will be a judgement criterion rather than a contractual one. The stage closes when the brief is signed by both leads and the finance partner has confirmed the budget envelope.

Stage 2: Market scan and longlist

Procurement is responsible for the mechanics: identifying candidates, running conflict and financial-stability checks, and collecting comparable documentation from each. Marketing is accountable for the outcome, which is whether the longlist actually reflects the brief. A longlist assembled on procurement criteria alone tends to favour agencies that are easy to contract, and a longlist assembled on marketing enthusiasm alone tends to favour agencies someone already knows. The closing deliverable is an approved longlist with a one-line rationale per agency, so that anyone later can see why each name is present.

Stage 3: Shortlist criteria and scorecard weights

This is the stage that makes shared ownership durable. Both functions set the weights for their own domain, both sign the total, and the scorecard is locked and dated before any agency sees the RFP. The point of locking is to remove the possibility that AAR describes, where commercial and creative arguments run separately and each side re-weights after seeing what it does not like (AAR Group). Once prices or creative responses are visible, any change to the weights should require the executive sponsor's approval and a written reason. The contents of the scorecard are covered later in this article; here the decision is simply that the weights exist and cannot move.

Stage 4: Capability and chemistry evaluation

Marketing is accountable and responsible, with brand and channel leads acting as evaluators. Procurement is consulted, and its brief is narrow: observe whether the proposed team is realistically resourced, whether the working method could actually be delivered at the scope described, and whether any subcontracting or dependency creates delivery risk. Procurement does not introduce price at this stage. The deliverable is a set of signed capability scores per agency, completed before any commercial document is opened by the marketing evaluators.

Stage 5: Commercial evaluation and negotiation

Procurement is accountable and responsible, legal is responsible for terms, and marketing is consulted on one question: is the scope the agency has priced the scope the brief actually asked for? Marketing's role here is to defend scope realism, since a fee that looks low because the agency quietly narrowed the deliverables is a false comparison. Marketing does not reopen creative preferences at this stage. TrinityP3's sequencing rule places financials after strategic confidence for exactly this reason (TrinityP3). The deliverable is a commercial memo showing each finalist's fee against benchmark, scope-adjusted, with any terms that remain open.

Stage 6: Final recommendation and approval

Marketing owns the recommendation. Procurement owns sign-off on commercial and risk terms. When the two disagree, the escalation path is a named executive sponsor identified in the RACI before the process began, and the sponsor decides on the record. Escalation should never default to whoever argues longest or whoever has the more senior title in the room. The approval record names both signatories and the sponsor.

How to run the RACI workshop

1. Circulate the six-stage table one week before the workshop and ask each lead to fill in named individuals for every R and A cell.
2. In the workshop, compare the two versions and mark every cell where the leads disagree or where two people are marked Accountable.
3. Resolve each conflict by asking which function's evidence the decision depends on most; where genuinely equal, name a co-accountability and a tie-break owner.
4. Name the executive sponsor for stage 6 escalations and confirm their availability across the selection calendar.
5. Agree the deliverable that closes each stage and who files it.
6. Attach the completed RACI to the signed brief and circulate before any agency is contacted.

Treat any cell that still holds two Accountable parties after the workshop as an unresolved conflict; it will resurface during the live selection.

Sequence Capability Before Commercials

Write the evaluation calendar so that no commercial document is opened by marketing evaluators until capability scores are signed. The reasoning follows from the RACI: procurement holds price, marketing holds capability judgement, and the order in which those two readings happen changes both of them.

Why parallel evaluation distorts both judgements

TrinityP3 argues that separating the assessment of commercial viability from the assessment of strategic capability is essential, and that financials should only be opened once the buyer is confident in the agency's team and process (TrinityP3). When both are visible at once, the cheapest proposal gains an unearned glow and the most expensive one is read more harshly than its capability deserves. The reverse also happens: a team the evaluators liked in the room gets its fee waved through without proper benchmark scrutiny. Sequencing protects both functions from contaminating their own judgement.

The recommended order is:

1. Define the problem and the success measures, and confirm they are the ones in the signed brief.
2. Test core competencies and working method through practical exercises rather than presentations.
3. Build confidence in the specific team that would work on the account, including senior time commitments.
4. Only then open financials, benchmark them and negotiate.

What to test instead of speculative creative

TrinityP3's critique of the traditional pitch is that it rewards a single high-stakes creative moment rather than a proven, sustainable way of working (TrinityP3). The alternative is to test how the agency solves problems: a working session on a real brief with the real team, a walkthrough of how a past campaign moved from insight to execution, and references contacted about how the agency behaved when something went wrong. During these rounds procurement's job is to watch for delivery risk and resourcing realism, such as a senior team presenting while a junior team would deliver. Price stays sealed.

This argument applies most strongly to creative and advertising selection. For media buying, where much of the value sits in contractable and auditable terms, speculative work is less of a distortion and commercial evaluation can carry more of the load, which the agency-type section below addresses.

When the timeline forces compression

Sequential evaluation assumes a calendar that allows it. When the timeline is compressed, the answer is a seal rather than a merge. Procurement receives and holds pricing while capability rounds run, and marketing evaluators sign their scores before pricing is released to them. Procurement can begin benchmark preparation in parallel because that work does not require knowing which agency marketing prefers. The rounds happen in the same fortnight, but the information flow still runs in one direction. If the gate cannot be held, record the loss of decision quality as a risk in the approval memo.

Resolving Cost Against Creative Quality: Build a Value Scorecard, Not a Price Ranking

Agree the capability threshold and the criterion weights in writing before the RFP issues, then run a mock score on last year's incumbent to see whether the scorecard would have chosen them. If the answer surprises both leads, the scorecard needs work before it meets a live proposal.

Why lowest bid is a false economy in marketing services

General procurement guidance from Art of Procurement warns that a low-cost supplier can look like a win on paper while poor quality, instability and reputational exposure raise operational costs over time (Art of Procurement). This is a heuristic rather than an empirical law, but in agency terms the mechanisms are familiar: rework when the first round of work misses, churn of the people on the account when the agency cannot afford to retain them at the agreed fee, and lost time while the marketing team compensates for both. None of those costs appear in the fee comparison; they show up in the following year's results.

Benchmark to find the right price, not the lowest one

RightSpend, which sells spend-management tooling and therefore has an interest in benchmarking, frames the discipline as paying the right price for the right work rather than driving prices down (RightSpend). The vested interest is worth noting, but the principle stands on its own and it splits ownership cleanly. Procurement owns the benchmark comparison: rate cards, blended rates, scope-adjusted fees against comparable engagements. Marketing owns the definition of the right work, which means confirming that what is being benchmarked is the scope the brief asked for and the seniority the capability rounds established. A benchmark applied to the wrong scope produces a precise answer to the wrong question.

Weighting the scorecard: who sets what

The table below is a diagnostic structure. Weights are deliberately absent because they belong to your organisation, and any number presented here would be invented.

The threshold logic follows TrinityP3's sequencing rule that capability confidence precedes financials (TrinityP3), and the value framing follows RightSpend's right-price principle (RightSpend).

The rule that resolves cost against quality is simple to state: a proposal must clear the capability threshold before its commercial score counts at all. Below the threshold, price is irrelevant, however attractive. Above it, the commercial score does its work among agencies that could all deliver.

A hypothetical illustration, with invented numbers for clarity only: suppose three finalists price the same scope at 100, 85 and 70 index points. If the agency at 70 fails the capability threshold, it exits before the commercial score is calculated, and the decision is made between 100 and 85 on their combined fit and value scores. Reading this as a 30 percent saving foregone is the mistake the threshold exists to prevent, because the saving was never available at the required quality.

Results-based terms: where they help and where attribution breaks

Procurecon's commentary suggests marketing procurement can negotiate contracts around results rather than the completion of listed tasks, giving the company more control over its marketing budget (Procurecon). The limitation deserves equal weight. In multi-channel marketing, attributing an outcome to one agency's work is difficult, and a fee tied to a metric the agency cannot control turns into a lottery for both sides. Use results-based components for outcomes the agency demonstrably controls and that can be measured without contested attribution, and keep the rest as scope-based fees with clear acceptance criteria. Procurement drafts the mechanism; marketing decides which outcomes qualify. Neither should treat a measured lift as proof that the agency caused it.

How Ownership Weighting Shifts by Agency Type

Before adopting the RACI for a new category, annotate each stage with the lead function for that category and confirm with both leads that the change is deliberate. The stages and the RACI structure stay constant. What shifts is the relative weight of commercial versus judgement-based criteria, and therefore which function carries more of the scoring at stages 4 and 5. The categories on the GPI directory, which span paid media, creator, Amazon and creative agencies, illustrate why a single fixed weighting cannot serve every selection. The grid below is a conceptual application of the cited principles, not a set of measured findings.

The creative row reflects TrinityP3's point that the anti-pitch critique applies most strongly to creative and advertising selection and less to media buying (TrinityP3). The production row reflects RightSpend's benchmarking principle, which is most legitimate where units of work are comparable (RightSpend). The PR and influencer row reflects the 2023 Procurement Tactics guidance on early joint definition of expectations (Procurement Tactics).

Creative and brand agencies

Marketing judgement dominates capability scoring, and procurement's largest contribution comes later, in contract structure and intellectual property terms. Usage rights, ownership of work product and what happens to unfinished concepts at exit matter more in this category than in any other, and they are procurement's to draft from marketing's stated needs.

Media planning and buying agencies

Transparency of fees, rebates and inventory arrangements, along with access to data and reporting, are contractable and auditable. Procurement's weight rises accordingly, and commercial evaluation can legitimately start earlier. Marketing still owns the judgement about planning quality, which no audit clause replaces.

Production and specialist services

This category sits closest to conventional procurement. Unit-cost benchmarking is legitimate because units of work are comparable, and procurement can lead more stages, including parts of the longlist and shortlist. Marketing remains accountable for whether the output fits the creative direction it serves.

PR, influencer and content partners

These are judgement-heavy categories with measurable delivery components attached, such as publication volumes, disclosure compliance and reporting cadence. The right approach is hybrid weighting with a KPI schedule that names an owner for every metric before the market is engaged, so that neither function later disputes what was promised.

Contract, Handoff and Who Owns the Relationship After Signature

Draft a one-page relationship charter during negotiation, naming an owner for each recurring decision, and attach it to the contract as a schedule. Selection is complete only when the operating relationship has clear owners, which comes after the approval record.

The contract converts pitch quality into delivery quality

Veza Network makes the point directly: the contract is what turns the quality shown in a pitch into the quality delivered on the account, and without it the conversion depends on the agency's goodwill, which is a procurement risk rather than a procurement instrument (Veza Network). This is an advisory position rather than a measured outcome, but it reframes procurement's contract role as a marketing outcome lever. Procurement owns contract structure, rate cards, change control and audit rights. Marketing owns scope, deliverable definitions and acceptance criteria. The KPI schedule is co-owned, with results-based components limited to agency-controlled, measurable outcomes as described in the scorecard section.

Durability enters before signature. Team stability commitments, key-person clauses that name the individuals marketing evaluated in stage 4, notice periods for staffing changes and exit terms that protect data and work in progress are all procurement-drafted from marketing-defined needs. A team that impressed in the room and left within a quarter is a selection failure that the contract could have priced.

Handoff from sourcing to operating

After signature, day-to-day relationship management moves to marketing. Procurement retains ownership of periodic commercial review, benchmark refresh and contract compliance. The relationship charter is the handoff artifact. It lists the recurring decisions that will arise, scope changes, fee disputes, performance reviews, staffing changes and renewal, and names an owner and a consulted party for each. Without it, the first fee dispute lands on whichever person the agency emails, and the ownership clarity built during selection dissolves.

Governance cadence and long-term value

Procurement Tactics' 2023 guidance holds that marketing procurement should not be treated as a single transaction, and that longer supplier relationships tend to improve communication, consistency, trust and performance (Procurement Tactics). The source offers no quantified performance data, so this stands as an enduring principle rather than a measured effect. In practice it argues for a governance cadence agreed at signature: a quarterly performance review led by marketing with procurement present, an annual commercial review led by procurement with marketing present, and a renewal decision with a named owner and a date. Scheduled reviews, rather than reviews triggered by a crisis, are what keep the partner viable.

Data Transparency and AI Readiness: Whose Question Is It?

Add a due-diligence schedule to the RFP requiring agencies to list data ownership terms, reporting access, subcontracting arrangements and the AI tools in use across production and media. Procurement verifies the answers; marketing scores their strategic relevance. No captured evidence quantifies how buyers currently assess AI readiness in agency selection, so what follows is a framework rather than a description of industry practice.

What procurement can verify

Questions about who owns the ad accounts and data, who holds reporting access, what is subcontracted and which AI tools touch the work are contractable. Because they are contractable, they are verifiable, and procurement should verify them against documentation rather than accept them from a slide. Veza Network's point about the contract carrying pitch promises into delivery applies here as much as to creative quality (Veza Network).

What marketing must judge

Whether an agency's use of AI and data actually improves strategic and creative output is a judgement, and it belongs in the capability rounds. An agency can pass every disclosure check and still use its tools only to produce ordinary work faster. Marketing should ask to see how a tool changed a decision on a past piece of work, and what it got wrong.

Using documented agency profiles as evidence

The joint rule is the one that runs through this article: every claim an agency makes about data or AI capability should be assessed against its evidence, its methodology and its limitations. Documented agency profiles give both functions a starting evidence set that does not depend on pitch assertions. For a sense of what documented agency evidence looks like in practice, see how a profile such as AB Marketing Group on GPI records criteria buyers can compare against pitch claims. The profile is an illustration of documented fields a buyer could request from any candidate; it makes no claim about that agency's AI capability.

How GPI's Evidence Standard Supports Shared Decision Ownership

The argument reduces to four practices. Write decision rights down per stage, with one deliverable closing each. Assess capability before commercials, and seal pricing when the calendar will not allow a clean sequence. Score value against a capability threshold and a benchmark rather than ranking on price. Use the contract, and a relationship charter attached to it, to carry pitch promises into delivery and to name owners after signature.

GPI applies the same discipline to agency evaluation that this article asks marketing and procurement to apply jointly: claims are assessed against documented evidence, stated methodology and acknowledged limitations. That standard gives the two functions a shared evidence base instead of competing instincts. To define what counts as evidence in your own scorecard, the GPI methodology explains how the Growth Partner Confidence Score weighs documentation, method and limitations.

The RACI and scorecard here are frameworks drawn from expert guidance, not measured outcomes. Adapt them to your organisation's structure, and record the adaptation.

Frequently Asked Questions

Who should have the final say if marketing and procurement disagree on the shortlist?

A named executive sponsor identified in the RACI before the process began, deciding on the record. Marketing is accountable for whether the shortlist reflects the brief and procurement for whether each candidate passes due diligence, so a disagreement usually means one of those two tests is being applied to the other's domain. The sponsor's job is to name which test the disputed agency actually failed.

Should procurement attend chemistry and capability sessions?

Yes, as a consulted observer with a narrow brief: delivery risk and resourcing realism. Procurement watches whether the team in the room is the team that would deliver and whether the working method could survive the scope. It does not raise price, consistent with TrinityP3's rule that financials wait for confidence in team and process (TrinityP3).

How do we stop agencies from lowballing fees to win and then renegotiating?

Threshold-first scoring removes most of the incentive, because a low fee cannot rescue a proposal that fails capability, and benchmark comparison flags a fee that sits implausibly below comparable engagements (RightSpend). Contract change-control and fixed rate-card terms handle the rest, since renegotiation then requires a documented scope change rather than a conversation. As Veza Network puts it, the contract is what converts pitch quality into delivery quality (Veza Network).

Can procurement lead selection for a specialist category marketing does not understand well?

Procurement can lead more stages for production-like categories where units of work are comparable and benchmarking is legitimate. Marketing remains accountable for fit, because the commercial and governance expertise procurement brings does not substitute for brand and customer judgement (AAR Group). Where marketing lacks category knowledge, bring in an internal channel lead or an independent adviser as a consulted evaluator rather than transferring accountability.

How often should decision rights be reviewed after the RACI is agreed?

At two fixed points: once at the end of each selection, as a short retrospective on which cells caused friction, and once whenever a lead changes in either function. The RACI should not be reopened mid-selection except through the sponsor, because re-weighting after prices or creative are visible is the failure pattern the locked scorecard exists to prevent.

Is a results-based contract realistic for a creative agency?

Only for components the agency controls and that can be measured without contested attribution, such as delivery against agreed milestones or defined quality gates. Procurecon's commentary supports results-based negotiation as a way to gain budget control (Procurecon), but tying creative fees to business outcomes shaped by media, pricing and distribution asks the agency to bet on factors it does not influence. Keep those outcomes in the review conversation rather than the payment mechanism.