You're three months in. The kickoff was strong. The team seemed capable. But something has shifted.
Deliverables are coming in late or just slightly off. The strategic thinking you saw in the pitch isn't showing up in the work. Your point of contact has changed twice without explanation. You're spending more time correcting and redirecting than you expected. Results are flat and the explanations are getting longer.
You haven't said anything yet because you're not sure if this is a real problem or a rough patch. You don't want to damage the relationship over something that might resolve itself. And honestly, the idea of starting the selection process over is exhausting.
So you wait.
This is the moment most agency relationships are won or lost. Not in the pitch. Not at signing. Here, at the first sign of friction, when the path of least resistance is to say nothing and hope it improves.
It almost never improves on its own.
The difference between a rough patch and a real problem
Not every early struggle signals the wrong agency. Engagements have ramp-up periods. New teams take time to learn your business, your voice, your internal dynamics. Some friction in the first sixty to ninety days is normal and doesn't mean the relationship is broken.
The question isn't whether there are problems. The question is what kind of problems they are.
Execution problems are fixable. Late deliverables, miscommunication on scope, work that needs more revision than expected — these are process issues. They have process solutions. A direct conversation about expectations, a tighter feedback loop, a clearer briefing format. If the agency responds to direct feedback with genuine adjustment, the relationship is probably worth continuing.
Structural problems are not fixable through better communication. If the team working your account isn't the team that pitched, if the strategic capability you were sold isn't actually present in day-to-day execution, if the agency's business model requires them to understaff your account to maintain their margins — those aren't communication problems. They're structural ones. More conversations won't change them.
The hardest part of this diagnosis is that agencies are often very good at making structural problems feel like execution problems. Another kickoff. A new account lead. A revised project plan. These gestures buy time without changing the underlying dynamic, and they can keep a broken engagement alive for months longer than it should be.
How to run the diagnosis
Before you decide anything, get specific about what's actually happening. Vague dissatisfaction is hard to act on. A clear pattern is not.
Go back to the original proposal and the criteria you used to select this agency. Score the actual engagement against those same criteria today. Not against your feelings about the relationship, not against what you hoped would happen — against the standard you applied when you chose them.
If you ran a structured evaluation, this is straightforward. Your scoring criteria are documented. Apply them to what's happened in the last ninety days and see what the numbers show. If you didn't run a structured evaluation, you're doing this from memory, which is harder but still worth doing.
Ask yourself three questions. Are they delivering what they proposed, at the quality level they demonstrated in the pitch? Is the team working your account the team you evaluated? And when something goes wrong, does the agency surface it proactively or do you find out when it's already a problem?
Honest answers to those three questions will usually tell you whether you have an execution problem or a structural one.
The conversation you need to have before you make any decision
If the diagnosis points to execution problems, have a direct conversation before you do anything else. Not a gentle check-in. A specific, documented conversation about what isn't working, what needs to change, and what the timeline is for seeing that change.
Most clients avoid this conversation because it feels confrontational. It isn't. It's the most professional thing you can do, for the agency and for yourself. Agencies can't fix problems they don't know you're tracking. A clear conversation gives them the opportunity to respond. It also creates a documented record of what was discussed, which matters if the relationship doesn't improve and you need to make a harder decision later.
Come to that conversation with specifics. Not "the work hasn't felt strategic enough." Instead name the deliverable, the date, the gap between what was proposed and what was delivered. Agencies respond to specifics. They can negotiate around generalities indefinitely.
Set a defined review point. Not an open-ended commitment to monitor the situation. A specific date, four to six weeks out, at which point you'll assess whether the issues have been addressed. This signals that you're serious, gives the agency a fair window to respond, and gives you a clean decision point rather than an indefinite waiting period.
When to stop trying to fix it
There are three situations where continuing the engagement is the wrong call regardless of what the agency promises.
The first is when the structural problem is confirmed. If the senior team that pitched you is not involved in your account and there is no credible path to changing that, the engagement is unlikely to improve. The agency's economics don't support a different model for your account size, and no amount of escalation will change that math.
The second is when trust is gone. Trust in an agency relationship is built through consistency — doing what they said they would do, when they said they would do it, at the quality they represented. When that pattern breaks repeatedly and the explanations keep changing, the relationship has lost the foundation it needs to function. You can re-engage with a new agency faster than you can rebuild trust with one that's lost it.
The third is when the cost of staying exceeds the cost of leaving. This calculation is more than financial. Factor in the internal time you're spending managing the relationship, the opportunity cost of results that aren't materializing, and the organizational credibility you're spending defending a choice that isn't working. Sometimes the most expensive decision is the one that keeps an underperforming engagement alive for two more quarters because the switch feels hard.
How to make the exit clean
If you decide to end the engagement, do it clearly and professionally. Review your contract for notice periods, deliverable ownership, and data access provisions before you say anything. Know what you're entitled to — creative files, analytics access, platform credentials — and make sure those handover terms are part of the exit conversation.
Document the transition. What's in progress, what's been completed, what's been paid for and not yet delivered. This protects you in the handover and gives your next agency a clear picture of where things stand.
And document why the engagement ended. Not for the agency. For yourself and your organization. The same structured evaluation you should have run at the start is worth running at the end — what did this engagement teach you about what to weight differently next time? That information is more valuable than it might feel in the moment when you're relieved it's over.
The agencies that feel like failures often contain the clearest instructions for what to do better on the next decision.
Where to start
If you're currently evaluating whether an agency relationship is worth continuing, the free Agency Comparison Scorecard can help you reapply your original criteria to the current engagement — the same framework works for evaluation and for diagnosis.
If you're heading into a new search after a difficult engagement and want a process that protects you from the same outcome, the full Evaluation Pack gives you weighted scoring, documented trade-offs, and a stakeholder-ready recommendation your leadership can approve with confidence.
— The Clarity Brief