How to Tell If Your Marketing Agency Is Actually Working

The monthly report lands on the fifteenth, and it is a beautiful document. Twelve slides. Impressions up nineteen percent. Engagement rate climbing. Forty pieces of content shipped, a fresh posting cadence, a content calendar mapped out for next quarter. Every chart points up and to the right. Your account manager walks you through it with real warmth, you nod, and you approve the next invoice, because arguing with a green dashboard feels petty.

Then you open your bank account, and it looks exactly like it did in March.

That gap, between a report that is thriving and a business that is not, is the most common thing we get hired to explain. The founder on the call is rarely angry. They are confused, and a little embarrassed, because they cannot point to the problem. The agency is busy. The agency is responsive. The deck is genuinely impressive. And still, nothing that shows up on the report ever shows up in the pipeline. The suspicion creeps in slowly: I am being managed, not grown. This post is the test that tells you whether that suspicion is correct, before you do anything about it.

How do you know if your marketing agency is working?

You know your agency is working when you can trace real pipeline, leads, booked calls, and closed revenue, inside accounts that would still be yours if you fired them tomorrow. Activity reports do not answer that question; they are built not to. Posts shipped, impressions, and engagement rate measure whether the agency was busy, not whether your business grew. The only honest test runs through four questions, and you can answer all four this week without their help.

Run them in order. Each one is a yes or a no, and a no is not automatically a firing offense. It is a finding.

One: can you log into every account yourself, right now, without asking them? Open a new tab and try. Your ad accounts, your analytics, your domain registrar, your CRM, your email platform. If you can reach all of them with your own credentials and see the agency working inside as a user you added, that is the healthy shape. If getting in requires a Slack message and a wait, or if the login belongs to the agency and you have never seen it, you do not have an agency problem yet. You have an ownership problem, and it is the one that turns expensive the day you leave. We wrote the full version of this in the retainer trap: the account holder, not the contract, holds the leverage.

Two: can you trace a single lead from the first click to a closed dollar? Pick one customer who bought something last month. Ask the agency to show you where that person came from: which channel, which campaign, which ad, and what it cost to acquire them against what they paid you. A working setup answers in minutes, because the plumbing was built to answer it. A broken one produces a platform-reported conversion number that does not reconcile with anything in your bank. Real attribution is unglamorous and specific, and it is the difference between knowing your cost per acquisition and guessing at it. We built exactly this for a med spa client, server-side conversion tracking that reconciled to actual bookings instead of inflated platform numbers, and it produced $1.3M in attributed revenue at 6.7x ROAS precisely because every dollar was traceable. The mechanics live in tracking every dollar from click to close and the Skin & Self case study.

Three: is the work an owned asset, or rented activity? Look at what the last ninety days produced and ask whether any of it keeps working after the retainer stops. A landing page on your domain, a tracking layer inside your accounts, an email list that grew, per-location pages that rank, a review engine that runs on its own: those are assets. They compound. Forty social posts that expired the week they went up, a boosted-post budget with no pixel history you control, a monthly deck: those are activity. Activity ends when the invoices end. The Magna Pest engagement is the clean version of an owned asset doing its job; the acquisition engine kept producing while the business grew from four to eleven locations, because the machine belonged to the client, not to us.

Four: would any of it survive you firing them tomorrow? This is the summary question, and the first three feed into it. Imagine you send the termination email tonight. In the morning, what still works? If the honest answer is a website you own, tracking you can read, a list you can mail, and campaigns running in accounts with your name on them, the agency has been building for your independence, and that is an agency worth keeping. If the honest answer is that your marketing goes dark the moment they revoke access, you have not been growing. You have been renting, and the rent was the whole product.

A working agency builds a machine that would keep running if you fired them. A failing one builds a report that stops the day the invoices do.

Three of those four questions are about your own site and stack, and you do not need the agency in the room to start answering them. Run the free Pre-Flight Check on your site and it reads back your SEO, your local presence, your tracking setup, and your conversion readiness in a few minutes. It will not tell you whether your agency is honest, but it will tell you whether the infrastructure they are billing you to manage is actually in place, which turns out to be a surprising number of the answers already.

Which marketing numbers are theater, and which are real?

Theater metrics measure motion; real metrics measure money. Impressions, reach, follower count, engagement rate, and posts shipped all climb when an agency is busy, and none of them prove a single dollar moved. The real numbers are cost per acquisition, booked revenue traced to source, owned-list growth, and the ratio of what you spent to what closed. If your monthly report leads with the first set and buries or omits the second, that ordering is the tell.

The direction of causation is what separates the two, the same tell we flagged in the retainer piece. Honest reporting starts from your revenue and works backward to the spend that produced it. Theater starts from the spend and works forward to whatever number looked best that month. A report that opens with impressions and mentions revenue on slide fourteen was assembled by someone who already knows which number is weaker.

None of this makes the soft metrics worthless. A growing owned audience is real leverage, which is the entire argument for building an acquisition engine you own instead of renting reach you have to keep repurchasing. The distinction is ownership and traceability. Follower count on a platform you do not control is theater. Subscriber count on a list you can export and mail tomorrow is an asset. Same shape of number, opposite value.

Is a slow month proof your agency is failing?

No. A single slow month is noise, and treating it as a verdict is how good operators get fired for seasonality. This test is a diagnosis, not the exit. What you are looking for is not one bad number but a structural answer to the four questions: whether the machine being built is yours and traceable, or theirs and opaque. A slow month inside a fully owned, well-instrumented system is a data point you can investigate. A great month inside a rented one is a countdown.

The honest failure modes are worth naming, because sometimes the report is not the problem. Marketing that is genuinely working still has flat quarters, especially in seasonal businesses and long sales cycles where the pipeline built in spring closes in fall. If your four answers are healthy and one month came in soft, the correct move is to read the funnel data you can already see, not to fire anyone. Diagnosis first.

Sometimes the diagnosis points back at you. An agency cannot fix a product nobody wants, a sales team that does not call leads back, or an offer priced past the market. We wrote the uncomfortable version of that in when to fire your agency, including the cases where the agency is doing its job and the bottleneck sits downstream in your own operation. The four-question audit earns its keep precisely because it separates the two. If you own the accounts, the tracking answers cleanly, and the assets are compounding, and the number is still flat, the problem is probably not the agency. If you cannot get into your own accounts and nobody can trace a lead to a dollar, the problem is not seasonality.

Run the four questions this week. Most founders already suspect the answer; the audit just turns the suspicion into something you can act on with a clear head instead of a grievance. If the answers come back clean, keep your agency and stop second-guessing the slow months. If they come back opaque, you now know exactly what to fix, and it is not a firing decision yet. It is an ownership one. If you want a second set of eyes on what your current setup actually produces, or a system built so these four answers are yes by default, book a call.