When to Fire Your Marketing Agency: A Practical Checklist (and What to Do Before You Do)

If I had a dollar for every bad agency I've dealt with, well, I'm sure you know the end of that sentence. There's a moment in a lot of founder conversations where someone sighs and says, "I think we need to fire our agency." Sometimes that's true. Sometimes the agency really is phoning it in. But often, what looks like an agency problem is really a data problem, a strategy problem, or a budget and expectations problem, and firing the agency without checking those first just means you'll be in the same spot six months from now with a different logo on the invoice.

Here is the short answer. Fire an agency right away if you don't own your ad accounts, if they can't explain what's happening and what they'll do next, or if you catch them dressing up junk conversions as wins. Replace them if they can't speak to your unit economics, never bring new tests, and can't tell you what's incremental. Before you do either, check your tracking, your brief, your budget, and your website, because a new agency won't fix those on its own.

1. Is the agency failing, or is the system around it failing?

Before you fire anyone, separate two problems. The first is an execution problem. The agency may be capable, but the goals were never clear, tracking is broken, the offer and landing pages are weak, sales follow-up is inconsistent, the product or service has quality issues, or the budget is too low to produce a signal. The second is an ownership problem. You don't have senior marketing leadership, so the agency is making strategy decisions by default. Most agency failures I see are really ownership failures. If nobody on your side is accountable for the strategy, the agency will fill the void, whether they should or not.

2. When should you fire a marketing agency immediately?

The first non-negotiable is account ownership. If you can't log into Google Ads, Meta, GA4, Tag Manager, or your CRM, stop. That is not normal. If they run any of your platforms on their own accounts, fire them immediately, because your data and your history leave with them.

The second is trust. If you discover they are counting button clicks or page views as conversions, hiding brand spend inside "non-brand," taking credit for demand you already had, or refusing to show you search terms and placements, end it. Trust is the only thing in the relationship you can't renegotiate.

The third is a pattern of calls with no real answers. If every meeting is a list of tasks, a few charts, and a vague promise to "keep optimizing," you're getting activity, not leadership. The same goes for blaming "the algorithm." Automation is real, which is exactly why management matters more than ever. A good agency can tell you what signals they are feeding the system, what they changed, what they learned, and what they're testing next. If they can't, you're paying someone to babysit a dashboard.

3. What are the warning signs that an agency should be replaced?

Some agencies are pleasant to work with and still need to go. Watch for reporting built around impressions, clicks, CTR, and "engagement," or around leads with no definition of lead quality. That means you're being managed to the wrong outcomes.

A growth partner should also know your margin, your CAC targets, your payback window, your capacity limits, and what good looks like for your business. If they never ask those questions, they can't optimize toward the business. They should be bringing you new angles, offers, and landing page tests, and if the landing page is the bottleneck, they should say so even if it's "not their job." If every conversation ends with "we need more budget" and no plan, that isn't strategy. A credible budget ask sounds like what's working, what's capped, the test plan, and the expected return range.

Incrementality is the big one now. If you're running Performance Max, retargeting, or branded search, your agency must be able to explain what is truly incremental and what is simply claiming credit for demand that already existed. If they can't, you are likely paying for attribution games. Finally, it's normal to onboard an agency, and for very complex or very large organizations that can take months. It's not normal to still feel, after that, that they don't understand your best customers, what makes a lead qualified, your sales process, or what makes you different.

4. What should you check before you fire your agency?

I walk clients through five checks first. They keep you fair about what's actually broken, and they make the transition much easier if you do switch.

Start with tracking. If GA4, your ecommerce or billing system, and the ad platforms all disagree and no one has reconciled them, you're flying a plane with three broken gauges. Confirm that purchases, leads, and booked calls fire correctly, that UTMs and naming are consistent, and that someone has written down how you measure CAC, MER, and ROAS. If tracking is a mess, a new agency won't fix it by magic.

Next, look at the brief. A lot of agencies are working from "we want growth" and "keep ROAS good." Do they know your revenue and margin goals, your acceptable CAC and payback period, and your priority products, audiences, and markets? If not, tighten the brief before you judge the results.

Then check whether your expectations and budget are in the same zip code. Spending $3,000 a month in a crowded DTC category with a cold audience and expecting 3 to 5x ROAS isn't an agency issue. It's asking a test budget for enterprise results. Make sure the budget gives them room to learn and that you're judging performance over a reasonable window.

Fourth, look at the offer and the website. Even the best media buying can't save a confusing offer, a slow site, a weak mobile experience, or a page with no proof. Check landing page conversion rates and where people drop off between add to cart and purchase, or between form start and form submit.

Finally, have one blunt, data-first conversation. Tie your concerns to numbers, such as "CAC has risen from X to Y and we don't see a plan to address it." Ask them what they believe is happening, what they've tried, and what they would do with more freedom, budget, or creative, and ask what they need from you. Some agencies step up and reset. Others show you, in real time, why it's time to move on.

5. How do you score your agency?

If you want a simple, non-emotional read, rate the agency 0 to 2 on each of these eight areas, where 0 means no, 1 means somewhat, and 2 means yes:

• Clear goals tied to revenue and profit

• Accurate tracking and meaningful conversions

• Full, transparent access to accounts and data

• Reporting that drives decisions

• A strong testing roadmap for ads, offers, and landing pages

• Proactive communication and clear next steps

• The ability to explain incrementality, especially for Performance Max, brand, and retargeting

• Lead quality that matches what sales or intake actually accepts

A score of 13 to 16 means keep them, tighten expectations, and scale what works. A score of 9 to 12 means the relationship is fixable with stronger oversight or a reset scope. A score of 0 to 8 means fire or replace them, because you're paying for noise.

6. What should you ask for before you make the final call?

If you're on the fence and want to be fair, run a 14-day audit sprint. Ask in writing for full admin access to every platform, the current conversion definitions and which ones count as primary, the last 60 days of changes to bidding, targeting, creative, and landing pages, the search terms and placement reports, budget split by brand, non-brand, and retargeting, and their next 30-day test plan with what success will look like. If they refuse, delay, or can't produce it clearly, you have your answer. If you'd like to see what a healthy account looks like first, my Google Ads audit checklist walks through it step by step.

7. If you do part ways, what should you have ready?

Don't just pull the plug and hope. Before the contract ends, make sure you have admin access to every ad account, analytics property, and tool, and export at least 6 to 12 months of campaign and ad performance, any tests they ran and what they learned, and their historical reports. That becomes the foundation for whoever comes next.

Then write down, in plain language, your revenue and margin targets, your CAC ceilings and MER floors, your priority products, markets, and segments, and any compliance or brand rules, which matter especially in healthcare and telehealth. That document becomes the brief for the next chapter.

8. Should you hire another agency, a fractional CMO, or someone in-house?

Instead of assuming you just need a better agency, decide what you actually need. Another agency makes sense when your strategy is solid, your data is clean, and you only need better execution in a channel. A fractional CMO makes sense when you need someone to own the strategy, channel mix, budget, and agency oversight and connect all of it to your P&L. An in-house hire makes sense when you have the volume and complexity to justify a full-time senior marketer and you're ready to build a team around them. I compare all three, with costs, in my post on fractional CMO vs. agency vs. in-house.

Many companies between $5M and $500M in revenue end up in the middle: fractional CMO leadership plus a few specialist agencies for execution. Sometimes the best fix isn't replacing the agency at all. When someone senior finally sets the strategy, defines conversions and reporting, and holds the system accountable, many agencies magically get better, because someone is finally steering the ship.

If you're staring at rising CAC, flat revenue, and thin reports and want a neutral read on whether you have an agency problem or a strategy and data problem, that's exactly what I do. You can book a free 30-minute introductory call here, and I'll give you an honest view of what's the agency, what's upstream, and what should change first.

Dr. Linda Orr, PhD, is the founder of Orr Consulting, a fractional CMO and marketing analytics practice serving brands from $5M to $500M in revenue. She built Upwork's first enterprise marketing mix model, covering an $88M annual media portfolio, and spent about 20 years as a marketing professor before moving fully into consulting.