After reviewing dozens of agency relationships, I see the same two patterns.
First, most switches happen 8 to 14 months too late. Not because leaders miss the problems, but because they don’t have objective criteria to separate growing pains from structural failure.
Second, most advice is either generic (“poor communication!”) or a sales pitch (“fire them and hire us!”).
This article gives you a practical diagnostic you can run on any agency, including ours.
The short answer
Fire your agency when structural issues persist after direct attempts to fix them. Use the 8-signal diagnostic to tell growing pains from failure.
How to use this diagnostic
Score each signal as 0 (not present) or 1 (present).
0-2 signals: Optimize the relationship
3-5 signals: Have the conversation; give 60-90 days
6-8 signals: Start your search
Structural issues means missing research, no decision framework, no outcomes mapping, or no cross-channel integration. One critical structural failure can outweigh several minor issues.
The 8-signal agency performance diagnostic
Most agencies deserve 6 to 9 months to show results. But certain patterns indicate structural problems that won’t improve with time.
Keep in mind that timelines vary by channel: SEO and brand work trend longer, while paid media and CRO can show directional lift sooner. Judge pace against channel realities and your scope, after access, tracking, and baseline are in place.
Here are the 8 signals to look for, with specific methods to verify each one.
Signal 1: They can’t explain the “why” behind recommendations
What to look for: You get tactics without strategy. They recommend activities but can’t articulate the decision framework or alternatives they considered.
Why it matters: Tactics without strategy is expensive activity theater. You’re paying for expertise in making strategic decisions, not just executing tasks someone else prioritized.
How to verify: Next time they recommend something, ask: “Walk me through why you’re recommending this approach versus the alternatives you considered.”
The answer tells you everything. Strong partners explain the decision framework. They walk you through what they evaluated, why they rejected certain options, and what success looks like for your specific situation.
Red flags: defensiveness, vague answers, or falling back on “this is best practice” without explaining why it’s best for you.
What good looks like: “We’re recommending content clusters over traditional blog posts because your competitive landscape shows weak topical authority across your core topics. Based on your domain strength and resource capacity, we can build defensible authority in 6 to 9 months. Here’s what we considered and why we rejected paid-first strategy.”
Reality check: Early in a relationship, they may not have enough context yet to provide sophisticated strategic rationale. This signal matters most after months 3 to 6, once discovery should be informing strategy.
Signal 2: No research artifacts, only templated deliverables
What to look for: Every client gets the same deliverables with logos swapped in. Recommendations feel generic rather than specific to your business, customers, or competitive landscape.
Why it matters: Template solutions can’t create differentiation. You’re paying for insights about YOUR business, not recycled frameworks that work for everyone.
How to verify: Ask to see the discovery research that informed your strategy. Specifically: stakeholder interview insights, competitive analysis findings, customer research results, or market assessment.
If they can produce artifacts, you’ll see things like stakeholder interview summaries revealing positioning gaps, competitive research identifying opportunities, or customer language patterns that shaped messaging.
If they can’t produce research artifacts, or everything looks suspiciously similar across clients, that’s your answer.
What good looks like: “Here’s the competitive white space analysis showing where your three main competitors are weak. Here are the verbatim customer quotes about what matters in their buying decision. This positioning emerged from patterns across 12 stakeholder interviews.”
Reality check: Some deliverable formats ARE standardized, and that’s fine. Brand guidelines follow consistent structures. Audit reports use similar frameworks. What matters is whether the insights inside those deliverables are genuinely yours or template fill-in-the-blank.
Signal 3: They report activity, not outcomes
What to look for: Dashboard of metrics that don’t connect to business results. They’re proud of rankings, traffic, or impressions but can’t explain how those lead to revenue.
Why it matters: Marketing exists to drive business results, not generate impressive-sounding numbers for quarterly reviews.
How to verify: Ask them to connect their metrics to your business goals. “Show me how this ranking improvement translates to customer acquisition” or “Walk me through the path from this traffic increase to revenue.”
Strong partners map every metric to business outcomes. They distinguish between leading indicators (things predicting future results) and lagging indicators (results themselves). They show you the logic connecting marketing activity to revenue.
Warning signs: changing the subject, defending vanity metrics, or showing you charts that go up and to the right without business context.
What good looks like: “This keyword ranking improvement puts us in position 3 for terms your customers actually search. Based on search volume and your historical conversion rates, we’re projecting 15 qualified leads per month from this cluster. At your typical close rate, that’s roughly $180K in annual recurring revenue.”
Reality check: Leading indicators matter early in relationships. You can’t show revenue impact in month two. But by month six, they should be mapping metrics to business outcomes and adjusting strategy based on what drives actual results.
Signal 4: You’re managing your agency instead of them advising you
What to look for: You’re the one pushing for strategy discussions, asking for accountability, and ensuring deadlines are met. They’re reactive rather than proactive.
Why it matters: You’re paying for expertise and advisory, not just execution hands. If you’re managing them like employees, you’re not getting the strategic partnership you’re paying for.
How to verify: Track who initiates strategic conversations over 90 days. Who brings up performance concerns first? Who identifies opportunities? Who pushes for uncomfortable but necessary discussions?
That pattern is the story. Strong partners own the relationship. They surface opportunities and concerns before you see them. They push for conversations about what’s working and what isn’t.
Problem pattern: waiting for you to ask questions, defensiveness when you raise concerns, or needing constant direction on next steps.
What good looks like: They schedule quarterly strategy reviews without you asking. They identify conversion bottlenecks before you see them in reports. They recommend course corrections proactively. You feel like they’re protecting your interests, not just completing assigned tasks.
Reality check: Some clients genuinely prefer hands-on involvement in everything. If you want to be deeply involved in execution details, make that clear. But strategic advisory should still be proactive, not reactive.
Signal 5: No integration across channels (coordination theater)
What to look for: PPC team doesn’t talk to SEO team. Brand positioning doesn’t inform website architecture. Customer insights from one channel never make it to other channels.
Why it matters: Real integration creates compound value. When brand informs web, web informs marketing, and marketing feeds back into brand, you get exponential value. Coordination is just separate departments sharing a client.
How to verify: Ask how brand strategy informed your website architecture. Ask how paid campaign insights shaped organic strategy. Ask how customer research from discovery flows into all subsequent work.
You should see clear connections. Brand positioning directly shaped page hierarchy. Customer language from research appears in ads, web copy, and content. PPC performance data informs which organic opportunities to pursue. One discovery feeding all subsequent decisions.
If each service works independently, that’s coordination, not integration. The logo designer never spoke to the web team. The content writer never saw the customer research. Each team starts fresh rather than building on shared knowledge.
What good looks like: “All specialists on your project consumed the same discovery research, not filtered briefs. Your brand positioning directly informed information architecture decisions. Customer pain points identified in stakeholder interviews appear across all channels because every team member understood them.”
Reality check: Some specialized contractors shouldn’t coordinate. If you hire a freelance video editor for a specific project, they don’t need integration with your SEO strategy. But strategic partners should demonstrate genuine integration, not just coordination.
Signal 6: They overpromise timelines then miss them repeatedly
What to look for: “We can have this done in 6 weeks” becomes 6 months. Deadlines consistently slip. Either they don’t understand the actual complexity, or they’re optimizing for winning business over accurate delivery estimates.
Why it matters: Either they don’t understand their own work well enough to estimate accurately, or you’re getting sales-led estimations that ignore delivery realities. Neither is acceptable at the price point you’re paying.
How to verify: Compare their initial timeline estimates to actual delivery across multiple projects. One delay happens. A pattern of 2 to 3x timeline misses indicates something deeper.
Strong partners provide honest timelines accounting for real complexity. They under-promise and over-deliver. They explain what affects timeline and communicate early when delays are likely.
Persistent pattern of missing deadlines by significant margins, blaming external factors, or making new promises they also can’t keep? That’s either poor estimation capability or sales-led promises that delivery can’t honor.
What good looks like: “This website will take 4 to 6 months from kickoff to launch. Here’s what affects timeline: discovery depth, revision rounds, your internal review process, content readiness, and technical complexity. We’ll update you every two weeks on progress against milestones.”
Reality check: Occasional delays happen. Clients cause them too through slow feedback or changing requirements. What matters is the pattern. Do they consistently miss promises by large margins? That’s the signal.
Signal 7: No strategic pushback (they say yes to everything)
What to look for: They never challenge your assumptions or bad ideas. Every request gets an agreeable “sure, we can do that.” They’re compliant rather than advisory.
Why it matters: You’re paying for expertise. Part of expertise is knowing when client requests won’t work and having the confidence to say so constructively.
How to verify: Think about the last time they disagreed with you or challenged an assumption. If you can’t remember, that’s your answer.
Partners worth the investment push back when it matters. They explain why certain approaches might not work. They offer alternatives based on experience. They protect you from expensive mistakes, even at risk of tension.
Never disagreeing? You might like the compliance, but you’re not getting the expertise you’re paying for.
What good looks like: “I understand why you’re thinking about a complete rebrand, but based on what you’ve shared about business priorities, I’d recommend addressing conversion optimization first. Your brand isn’t the bottleneck. Traffic quality is. Let’s solve that before investing in brand work.”
Reality check: Some agencies are yes-men by design. If you explicitly want order-takers, not advisors, that’s your choice. But don’t pay advisory-level prices for execution-level service.
Signal 8: Retention requires negotiation every renewal
What to look for: They never suggest expanding scope based on opportunities discovered. You have to push for additional services. Renewals feel like renegotiations rather than natural continuations.
Why it matters: Good partnerships grow naturally as agencies understand your business better. They see opportunities to help beyond initial scope. If they’re not suggesting ways to expand value, they’re not thinking strategically about your success.
How to verify: Have they proactively suggested additional ways to help based on what they’ve learned about your business? Or do all scope discussions originate from you?
Partnerships grow naturally as partners understand your business better. They see opportunities to help beyond initial scope. They suggest reasonable expansions based on discovered opportunities, not aggressive upselling.
Only doing exactly what’s contracted, never suggesting additional ways to help, renewals feeling like starting over rather than building on shared knowledge? You’ve got a vendor, not a partner.
What good looks like: “Over the past six months, we’ve noticed your website is converting at 1.2% when industry benchmark is 2.5%. We think conversion rate optimization would give you faster ROI than expanding paid spend. Here’s what that would look like.”
Reality check: In mid-market and enterprise organizations, annual renewals and legal refresh are normal. The red flag isn’t the paperwork – it’s when value feels negotiable every time because progress isn’t compounding. Scope creep is also real. Structured expansion based on discovered opportunities is different from boundary violations. Partners respect scope while identifying legitimate opportunities to add value.
What to do with your diagnostic results
You’ve worked through the 8 signals. Maybe you identified 2, maybe 6, maybe somewhere in between.
Now the question: what does that actually mean for your decision?
0 to 2 signals identified: You probably have a good partner
Focus on optimizing the relationship rather than searching for problems. Most successful agency relationships have minor areas for improvement.
Clear communication about expectations, regular strategy discussions, and mutual investment in the partnership will strengthen what’s already working.
3 to 5 signals identified: Time for a serious conversation
This is the zone where you need to address concerns directly before deciding to switch. Use the conversation framework in the next section to discuss patterns you’re seeing.
Give them 60 to 90 days to demonstrate meaningful improvement. That’s long enough to be fair, short enough that you’re not wasting months if nothing changes.
6 to 8 signals identified: Start your search
These patterns indicate structural problems that rarely improve. You’ve likely already attempted to address concerns without seeing change.
The gap between your expectations and their delivery is too wide to bridge with better communication. Time to find a partner whose approach aligns with what you need.
The conversation framework (if you’re addressing issues rather than switching)
Most people either stay silent about concerns until they explode, or they avoid the conversation entirely by switching agencies. Here’s the middle path.
How to frame the conversation
Approach this as problem-solving, not accusation.
“I’ve been evaluating our partnership objectively and want to discuss some patterns I’m seeing. My goal is to understand if we can address these together, because I value what’s working.”
This frames the conversation as partnership evaluation rather than performance review. You’re assessing fit, not just their performance.
Questions to ask for each signal you’ve identified
For Signal 1 (no strategic “why”): “Can you walk me through the decision framework behind your recent recommendations? I want to understand how you’re prioritizing among options and what alternatives you considered.”
For Signal 2 (no research artifacts): “Can I see the discovery research that informed our strategy? Specifically the stakeholder insights, competitive analysis, or customer research you gathered?”
For Signal 3 (activity, not outcomes): “Help me understand how the metrics you’re tracking connect to our business goals. Can you map the path from these numbers to revenue impact?”
For Signal 4 (reactive, not proactive): “I’ve noticed I’m usually the one initiating strategy discussions. I’d value your proactive input on what’s working, what’s not, and what opportunities you see.”
For Signal 5 (no integration): “How are different parts of our work informing each other? Can you show me where brand informed web, or where one channel’s insights shaped another’s strategy?”
For Signal 6 (missing timelines): “We’ve seen consistent timeline misses. Help me understand what’s causing this pattern and how we can improve estimation accuracy going forward.”
For Signal 7 (no pushback): “I’d value your expert opinion more, even when it challenges my assumptions. If you see risks in approaches I’m suggesting, I need you to tell me directly.”
For Signal 8 (no natural expansion): “Based on what you’ve learned about our business, where do you see opportunities to add value beyond current scope? What gaps are you noticing?”
What to listen for in their response
Ownership versus defensiveness. Do they acknowledge patterns and take responsibility? Or do they blame external factors, make excuses, or deflect?
Concrete action plans versus vague promises. “We’ll do better” isn’t an action plan. “We’ll implement weekly strategy check-ins starting next Tuesday” is.
Timeline commitments versus vague improvements. “Give us some time” isn’t a commitment. “You should see these changes within 30 days” shows accountability.
Understanding versus confusion. Do they grasp what you’re saying? Or do they seem genuinely confused about concerns you’re raising?
Timeline for improvement
Give them 60 to 90 days to demonstrate meaningful change.
That’s long enough to be fair. Most structural changes take 30 days to implement and another 30 to show results. It’s also short enough that you’re not wasting half a year hoping things improve.
Document the conversation. Note what was agreed. Set a specific check-in date to assess whether you’re seeing the changes discussed.
When to walk away
Walk away if you see these responses:
They’re defensive and won’t acknowledge patterns you’re raising. They make vague promises without concrete action plans or timelines. They blame you or external circumstances without taking ownership. After 60 to 90 days, you see no meaningful change from baseline patterns.
These responses indicate the relationship won’t improve. It’s not about whether they’re a good agency in general. It’s about whether they’re the right agency for you.
Common questions about firing your agency
How long should I give an agency before evaluating?
Six months minimum. Most agencies need 90 to 120 days just for onboarding and discovery. Real results typically show around month 6 to 9.
Evaluate too early and you’ll mistake normal onboarding challenges for structural problems. Wait too long and you’ve wasted a year on the wrong partner.
The 8-signal diagnostic is most accurate after month 6. That’s when patterns become clear.
What if they’re good at execution but weak on strategy?
This depends on what you hired them for.
If you hired them as an execution partner and knew you’d provide strategy, this might be fine. Not every engagement needs strategic advisory.
But if you’re paying for strategic partnership and receiving only execution, that’s a mismatch. You’re paying advisory-level rates for execution-level service.
Be honest about what you need. If you want strategy included, it needs to be part of the relationship. If you just need execution hands, different partners (and different pricing) might serve you better.
Should I tell them I’m evaluating other agencies?
Honesty about the relationship is fair. Announcing you’re interviewing competitors is usually counterproductive.
“I’m evaluating whether our partnership is meeting needs on both sides” is honest and appropriate. It signals that you’re assessing fit seriously.
“I’m talking to three other agencies” usually damages the relationship without adding value. It feels like leverage-seeking rather than honest evaluation.
Make the decision based on your assessment, then communicate it clearly.
How do I transition to a new agency without losing momentum?
Most agency transitions take 60 to 90 days of overlap for knowledge transfer.
Smart transitions involve: documenting what’s working (so new agency can maintain it), organizing all assets and credentials, extracting insights from current work, and overlapping relationships briefly if possible.
The cost of transition is real. Factor it into your decision:
Contract obligations: Review your agreement for notice periods, minimum terms, and early termination clauses. Most agency contracts have 30 to 90 day outs.
Budget reality: Plan for 60 to 90 days of overlap (paying both agencies during knowledge transfer), plus new onboarding investment. Sometimes optimizing a current relationship costs less than switching.
Internal politics: If you championed the current agency, address that directly. Frame it as “we gave this a fair shot, here’s what we learned, here’s the data-driven path forward.” Your leadership will respect systematic evaluation over either defensiveness or impulsive switches.
Don’t let any of these trap you in a fundamentally misaligned relationship. That’s sunk cost thinking.
Can these patterns improve, or are they structural?
Signals 1, 2, 3, and 5 are usually structural. Agencies that don’t do research, can’t explain strategy, or don’t integrate services typically won’t develop these capabilities mid-engagement.
Signals 4, 6, 7, and 8 can sometimes improve with direct conversation. Communication patterns and relationship dynamics are more fixable than fundamental methodology.
The conversation framework helps you test whether issues are fixable or structural. Their response tells you what you need to know.
What if I identify 3 to 4 signals but everything else is working well?
This is the toughest zone. You’re not definitely wrong. But you’re not definitely right either.
Use the conversation framework. Address concerns directly. Set clear expectations and timeline for improvement. Evaluate their response and actual changes over 60 to 90 days.
If they respond with ownership and concrete change, the relationship can work. If you see defensiveness or no real change, even 3 to 4 signals indicate enough misalignment to justify switching.
The bottom line
Most agency relationships that end should have ended months earlier.
Use these 8 signals to evaluate any agency. Including us. Including whoever you work with next.
The goal isn’t finding perfection. It’s finding alignment. An agency scoring well on these signals might still not be right for your specific situation. An agency failing 6 to 8 tests is definitely wrong.
Great marketing partnerships compound over time. Poor ones just compound costs.
These signals help you know which one you have.
