$ man how-to/agency-evaluation-checklist

The agency engagement starts strong. New team, fresh campaign ideas, promises of pipeline growth. Three months in, the campaigns are running but the results plateau. Six months in, you are paying the same retainer for the same playbook. The agency is not doing anything wrong exactly - they are just optimizing for campaigns shipped, not pipeline generated. This is the structural misalignment. Agencies bill for activity. You pay for outcomes. Their incentive is to keep campaigns running. Your incentive is to generate revenue. These two things are correlated but not identical. A go-to-market engineer spots this gap during the first audit call.

PATTERN

1. Who owns the tool logins? If the agency controls your Clay, Instantly, or CRM credentials, you are locked in. You should own every login. 2. Can you see credit consumption? Ask for a monthly breakdown of enrichment credits, email sends, and LinkedIn connection requests. If they cannot produce it, they are not tracking it. 3. What happens if you leave? Can you export all workflows, templates, and data? Or does it stay in their accounts? The answer reveals how portable your investment is. 4. How many other clients share your GTM engineer? If the answer is 8-10, your campaigns are getting maintenance, not engineering. An engineer managing 10 workspaces cannot optimize any of them. 5. Are they building infrastructure or running campaigns? Campaigns are one-time. Infrastructure compounds. Ask what assets you will own in 6 months that you do not own today. 6. Can they explain why each tool was chosen? If the answer is "we use Clay for all clients," that is a default, not a recommendation. Each tool choice should be justified by your specific needs. 7. What is their testing methodology? A/B testing subject lines, send times, audience segments? Or running the same playbook every month? 8. Are they measuring leading or lagging indicators? Reply rates are leading. Revenue is lagging. If they only report on vanity metrics (emails sent, open rates), they are not connected to your outcomes.

ANTI-PATTERN

The agency will not share tool logins. This is the biggest red flag. If you are paying for tools through the agency and cannot access them directly, your data and workflows are hostage. Reporting only on activity metrics. "We sent 5000 emails this month" is not a result. "We generated 12 qualified meetings from 5000 emails" is a result. If the agency reports on volume without conversion context, they are billing for busywork. Identical playbook across clients. Your ICP is different from their other clients. Your messaging should be different. Your qualification criteria should be different. If every client gets the same 3-email sequence template, the agency is running a factory, not a service. Resistance to audits. A good agency welcomes an independent evaluation because it validates their work. An agency that pushes back on audits has something to hide. That resistance alone tells you what the audit would find.

PRO TIP

The best agency engagements look like partnerships, not vendor relationships. You own all the tools and logins. The agency builds in your accounts. They produce monthly reports with credit consumption, conversion rates, and ROI analysis. They test hypotheses, not just run campaigns. A go-to-market engineer who audits agencies is not trying to replace them. Sometimes the audit conclusion is "your agency is doing excellent work - here are three things to optimize." Other times the conclusion is "you need to restructure this relationship." Either way, you have an independent assessment from someone with no stake in the outcome. The checklist above is not about finding fault. It is about aligning incentives. When both sides are measured on the same outcomes, the relationship works. When incentives diverge, results plateau.