Vetting multi-location marketing agencies: an unranked, unpaid pilot approach

Set the test boundary before issuing the brief

The profiles help identify questions, but they do not show how an agency performs for your locations. Keep the comparison unranked: ask every candidate to run the same limited routine campaign under the same terms, then assess observed work against criteria set before the test begins.

Define unpaid as no agency fee for pilot work, and state separately how any campaign spend or other client cost will be handled. Put the participating locations, access permissions, campaign window, required outputs, and pass conditions in the invitation. A location checklist can help make the information request consistent.

Make the campaign a real delivery test

Give each agency the same brief and the same available record of twelve months of spend and results from three to five locations that differ in market size.

Ask the team to execute a routine campaign across those locations, rather than merely describe its preferred approach. Include a local exception and a bad data input for the team to handle during delivery.

Set the pilot start and end dates in advance. The work begins once the agreed inputs and access are ready and ends when the campaign results and location reports are returned. Require the same outputs from each candidate: campaign activity, comparable location reporting, an account of staffing and fees, and a usable export and handoff record.

Before launch, write down what counts as success. Criteria can include whether the campaign follows the brief in each location, whether the local exception is handled as agreed, whether reporting definitions permit comparison, and whether account access and exports work. Set any outcome target from your own records rather than inventing a benchmark. Related analytics reporting can inform the measures.

DDB Worldwide: expose the local operating model

A large network and major multi-market clients are profile signals, not proof that local work will run smoothly. During the test, track who makes local decisions, how fees are presented, and whether your team can access and export the location data needed to review results.

Moroch Partners: identify the people doing the work

The profile points to local connections for national and multi-location brands. Require the proposed account team to carry out the agreed campaign, and record who handles each location, what comparable outcomes the team can support, and whether the work matches the staffing plan presented before the pilot.

R&R Partners: watch how regional capability is applied

The profile refers to multi-unit or franchise experience and regional capability. Use the local exception in the pilot to see how the team responds within your operating model. Record any relevant client examples, possible conflicts, and whether assigned staff understand the franchise structure and markets involved.

RPA: examine measurement while work is underway

RPA describes national and local media work alongside performance services. Ask the assigned team to show how results are measured and reported across the participating locations as the campaign proceeds. Compare the delivered reporting with the agreed definitions, rather than assuming the profile establishes an account measurement plan.

Team One: follow each service through its handoff

Team One lists strategy, media, data, social, and production. During the campaign, note which proposed people deliver each relevant service and who takes responsibility at every handoff. A broad service list alone does not show how work is divided among locations or whether the proposed team has capacity.

Walz Tetrick: check the proposed resources against delivery

The profile describes customized work spanning markets and locations. Compare the people and technology proposed for the pilot with what the campaign actually requires. Record whether the team can provide relevant program examples and whether its demonstrated capacity fits the location count and reporting needs.

Keep the comparison tied to observable work

Use the same six operating areas for each candidate: measurement, local execution, data access, account staffing, cost structure, and exit arrangements.

Record what the team did, what evidence supports its claims, and what remains unresolved. A franchise strategy process can help shape requirements, while a location registry question may inform reporting context without proving campaign performance.

Separate profile claims from pilot findings

The six candidates appear in 4As agency profiles. Treat those descriptions as prompts, not independently verified evidence. A pilot gives you direct observations of the proposed team at work, but does not establish that results will transfer to other locations, markets, or periods.

Use the searchable 4As Agency Search database to identify candidates, then invite incumbents and an agency research list to respond to the same terms. Keep agency claims, customer statements, measured records, and independent findings distinct. For each performance claim, record its criteria, commercial relationship, product version, study population, and method.

If a case study reports a change in cost per booked job, compare the stated method with your own twelve months of spend and leads, split by location. A different baseline market or period does not establish the result for your business. Keep unresolved claims separate from pilot observations so another reviewer can reconstruct the decision.

Agree on access and exit before sharing records

Set terms for account access, data exports, confidentiality, ownership, and deletion before sharing location-level information. Apply the same terms to every candidate, and make clear which permissions allow review and which could change accounts or campaign activity.

Use CISA's cybersecurity performance goals and incident response plan as discussion prompts, not proof of agency capability. Further review may include vendor risk, identity management, log management, control assessment, risk assessment, risk management, contingency planning, and cybersecurity incident handling. Treat unresolved exit conditions as open decision points.

Make selection depend on the same evidence

No agency paid to appear, and table order signals nothing about quality, fit, size, price, or recommendation. Base any appointment on current, documented multi-location evidence and the pilot criteria shared with every candidate. Record the reasons for the decision and retain the pilot outputs for review.

Before selecting a team, confirm named account staff, reproducible evidence, clear exit terms, a full export, account transfer, and deletion duties. Review those handoffs alongside delivery and measurement. For lead-related assessment, lead outcomes provide related context, while the pilot shows how each candidate handled your own brief.