How Do I Know If My Marketing Agency Is Doing a Good Job?

1. Results: Is the agency measuring what actually matters?

Start with the outcome the marketing is supposed to produce.

For an ecommerce company, that may include:

• Revenue
• New-customer revenue
• Customer acquisition cost
• Conversion rate
• Profitability
• Repeat purchase behavior

For a lead-generation company, it may include:

• Qualified leads
• Booked meetings
• Opportunities
• Cost per qualified lead
• Pipeline
• Customers acquired

Clicks, cost per click, impressions, rankings, traffic and engagement still matter. They can help explain why performance changed. But they usually are not the final outcome.

Imagine this monthly update:

Traffic increased 21%. Impressions increased 34%. CPC decreased 12%.

Those numbers could represent progress. They could also represent a month in which sales went down. Without the connection to the business outcome, you cannot tell.

The report test

Open your latest agency report and identify the number it wants you to care about first. Then ask:

Is this the result we hired the agency to create, or is it a metric that helps explain that result?

That distinction matters. A strong report should help you move from marketing activity to business impact.

Not every campaign can be tied perfectly to revenue. Not every business has a short sales cycle. But your agency should understand the outcome the company is ultimately trying to influence.

2. Measurement: Can you trust the numbers?

Before debating whether performance is good, make sure the performance is being measured correctly.

Suppose your agency reports:

5.2x ROAS

That sounds excellent. But the number immediately creates other questions.

• Which sales are included?
• Are existing customers included?
• Is revenue being counted in more than one platform?
• Does reported revenue broadly reconcile with Shopify, your CRM or another source of truth?
• What attribution settings are being used?

Marketing platforms do not simply observe a single, universal version of reality. Attribution assigns credit to different marketing interactions along the path to a conversion. That means two platforms can look at the same customer journey and assign credit differently.

That does not make platform data useless. It means the data needs context.

What good measurement should allow you to answer

The larger principle is simple:

Good measurement follows the business outcome, not whatever metric is easiest to count.

If the numbers look great but nobody can clearly explain what is being counted, start there before drawing conclusions about performance.

3. Optimization: What changed because of what the agency learned?

This is one of the most useful questions you can ask:

What did you change because of what you learned?

Marketing should create decisions. That does not mean an agency needs to touch every campaign every day. Constant activity can be just as unhelpful as neglect.

You are looking for intentional optimization. Depending on the channel, that could include:

• Reallocating budget
• Testing new creative
• Changing offers or messaging
• Refining audiences
• Reviewing search queries
• Excluding irrelevant traffic
• Improving landing pages
• Adjusting campaign structure
• Testing bidding strategies
• Changing email flows
• Revising SEO priorities

A useful conversation might sound like:

This audience became significantly more expensive without improving customer quality, so we reduced the budget and shifted part of it into the segment producing better acquisition economics.

That shows a chain:

Observation→Interpretation→Decision→Action

Compare that with:

We’ve been optimizing the campaigns.

That tells you almost nothing.

Google Ads gives you evidence

If Google Ads is part of the relationship, you do not have to rely entirely on somebody’s memory. Google Ads Change History records account and campaign changes over time. It can help you understand what changed, when it changed and who made the change.

But there is an important caveat:

More changes do not automatically mean better management.

Twenty unnecessary changes are not better than two smart ones. The useful question is:

Do the changes make sense given what was happening in the account?

Search-term data provides another example. Instead of asking only:

Did you check the search terms?

Ask:

What did we learn from the searches people made, and did that change anything?

The point is not the task. The point is the decision.

4. Strategy: Is your agency thinking or just executing?

Execution and strategy are not the same thing.

A company can launch campaigns, produce reports, publish content and hold meetings without making particularly good marketing decisions.

A strong agency should sometimes tell you something you were not expecting to hear. It might say:

We don’t recommend increasing the budget yet.

Google Ads isn’t the biggest problem. The landing page is.

Lead volume is healthy. Lead follow-up is where we’re losing the opportunity.

The agency isn’t the main problem here.

That last answer matters. If leads are coming in but your sales team waits three days to contact them, replacing the advertising agency may not solve the problem. If qualified traffic is reaching your website but customers abandon checkout, buying more traffic may simply buy more abandonment. If your most popular product is out of stock, Meta targeting cannot fix inventory.

A strategic partner separates marketing execution problems from business problems.

One question reveals a lot

Ask:

If this were your own money, where would you spend more, where would you spend less, and what would you fix before spending another dollar?

A strong agency should have an opinion. Not a reckless opinion. Not certainty where certainty does not exist. But a reasoned point of view.

5. Transparency: Can they explain performance clearly?

Marketing can be technically complicated. Communication should not be.

After a useful agency meeting, you should be able to answer four questions:

You do not need to understand every campaign setting or attribution model. That is one reason specialists exist. But technical complexity should never become a shield against accountability.

Good agencies report bad news

Marketing produces failures. Tests lose. Creative stops working. Keywords disappoint. Landing pages underperform. Assumptions turn out to be wrong.

That is normal.

What should concern you is a report where everything is always framed as a win, particularly when the business itself does not feel healthier.

A good agency does not need to make every decision correctly. It needs to recognize what happened, learn from it and respond.

Look for reasoning, not just reporting.

6. Accountability: Does anything happen after the meeting?

This is where many agency relationships gradually lose effectiveness.

A problem is discussed. Everyone agrees it matters. Someone says:

We’ll look into that.

Thirty days pass. The next call arrives. The same issue appears again.

Good accountability does not require a complicated system. It requires a visible chain:

Problem→Decision→Owner→Action→Follow-up

Consider the difference between these two updates.

Weak

We’re looking into the conversion-rate issue.

Stronger

Mobile conversion fell after the landing-page change. We’re comparing performance before and after the release, and we’ll return Friday with the three changes we recommend testing.

The second statement can later be evaluated. That is accountability.

Five questions to ask your marketing agency on your next call

If you are unsure how good the relationship actually is, ask these five questions together:

A strong agency may not have perfect answers. Marketing rarely offers perfect certainty. But you should hear reasoning.

The agency should be able to move logically from:

Data→Interpretation→Decision→Action

Signs your marketing agency may actually be doing a good job

Articles about agencies often jump straight to warning signs. That can lead businesses to look for reasons to fire an agency rather than objectively evaluate one.

There are also positive signals worth recognizing.

They bring you problems before you discover them

They do not wait for the monthly call to mention that an important campaign has deteriorated.

They tell you when a test failed

They are comfortable saying: “We tried it. It didn’t work.” That is useful information.

They disagree with you when there is a good reason

You hired expertise, not automatic agreement.

They recommend reducing or reallocating spend when appropriate

Every solution should not require a larger budget. Sometimes the right recommendation is to spend less.

They understand your business economics

They know the difference between a cheap lead and a valuable customer. They understand that ROAS, CAC, margin, lead quality and lifetime value tell different parts of the story.

They explain why they believe results changed

Not every explanation will be certain. But it should be evidence-based.

Their recommendations evolve

What the agency recommends six months into the relationship should reflect what it learned during those six months.

They separate what they control from what they do not

A good agency takes responsibility for its work without pretending it controls your entire business.

That is a much fairer standard than: Numbers down = bad agency.

Warning signs worth investigating

One awkward meeting does not prove you have a bad agency. One poor campaign does not prove incompetence.

Patterns matter more than isolated mistakes.

But repeated patterns such as these deserve a closer look:

• The explanation for poor performance is always the same.
• “Seasonality” gets blamed without supporting evidence.
• Reports contain lots of metrics but few decisions.
• Nobody can clearly explain what changed in your accounts.
• Problems persist across multiple meetings without an owner or action.
• Every recommendation involves spending more.
• Business performance is declining while every reported metric somehow looks positive.
• Tracking problems remain unresolved.
• You repeatedly discover problems before the agency does.
• Testing has no clear hypothesis or conclusion.
• Reporting focuses heavily on activity rather than outcomes.
• Direct questions consistently receive complicated or indirect answers.

None of these automatically means: Fire your agency.

They mean: Investigate.

How long should you give a marketing agency before judging performance?

You should not evaluate month one exactly the same way as month twelve. The expectations should evolve as the agency learns more about your business.

This is a decision framework, not a promise that every marketing channel should produce a particular result by a particular date. Timelines vary significantly. Paid search, SEO, creative testing and a long B2B sales cycle should not all be judged on the same clock.

The important question is whether the quality of knowledge and decision-making is improving as the relationship matures.

If you have worked with an agency for a year and every conversation still sounds exactly like month two, that is useful information.

What is actually your agency’s responsibility?

One of the easiest ways to evaluate an agency badly is to hold it responsible for everything.

That is not accountability. It is inaccurate attribution of responsibility.

Things an agency may meaningfully control or influence

Depending on the scope of work:

• Campaign structure
• Targeting
• Keyword strategy
• Media allocation
• Bidding
• Creative strategy
• Testing
• Marketing analysis
• Tracking implementation
• Reporting
• Landing-page recommendations
• SEO recommendations
• Email strategy
• Budget recommendations

Things the agency may not fully control

• Product quality
• Pricing
• Inventory
• Fulfillment
• Sales-team performance
• How quickly leads are contacted
• Customer service
• Internal approval delays
• Economic conditions
• Competitive changes
• Changes you choose not to implement
• Broader market demand

This distinction does not excuse poor agency work. It makes the evaluation more accurate.

The question is not merely: Did revenue increase?

The better question is: What was the agency responsible for, and how well did it perform that responsibility?

Should you fire your marketing agency if results are down?

Not automatically.

First determine what changed. A decline could come from:

• Poor execution
• Bad measurement
• Increased competition
• Changes in demand
• Seasonality
• Creative fatigue
• Website problems
• Offer fatigue
• Pricing
• Inventory
• Sales issues
• A strategy that simply stopped working

Then evaluate the agency’s response.

• Did they recognize the problem?
• Did they tell you?
• Can they explain what they believe happened?
• Are they investigating that explanation?
• Did they make a reasonable change?
• Do they have a next step?

A bad month with a strong response can be healthier than a good month nobody understands.

The more concerning situation is: Persistent underperformance combined with weak diagnosis, weak action and weak accountability.

The 10-minute agency check

Before your next agency call, see whether you can answer these six questions.

If the answers are clear, that is encouraging.

If several are unclear, it does not automatically mean your agency is failing. It means you have important questions that have not been answered yet.

And that is worth fixing.

When does an independent marketing review make sense?

An independent review becomes especially useful when the agency’s report and the business results seem to tell different stories.

It can also make sense when:

• You cannot tell whether agency reporting reflects actual business performance.
• Results and platform metrics seem to contradict each other.
• The same problems keep appearing without resolution.
• You are considering changing agencies.
• Marketing spend has become significant enough that guessing is expensive.
• You want validation before increasing your budget.
• Tracking or attribution is questionable.
• Your agency appears to be doing well, but you want to independently confirm it.

That last situation is important. An independent review should not begin with the assumption: “Your agency is bad.” That would not be independent.

A legitimate conclusion can be: “Your agency is doing good work. Keep them.”

Another might be: “The agency is mostly doing the right things, but these three areas need attention.”

Or: “The marketing execution is not the primary constraint. Fix this part of the business first.”

The value of an independent review is that the conclusion should depend on the evidence. Not on whether the reviewer gets to replace your agency afterward.