When to fire a vendor: the signals that matter.

You fire a vendor when the relationship stops producing results you can tie to revenue, and no honest conversation fixes it. That is the whole test. Not whether you like them. Not whether they are busy. Whether the money going out is coming back, and whether they can explain what they are doing when you ask.

Most owners wait too long. They keep paying a monthly retainer for another quarter, then another, because switching feels like more work than staying. I have sat in those review meetings. The vendor talks for forty minutes, shows a dashboard full of numbers that go up, and the owner leaves the call unsure whether anything actually happened. That feeling is data. This post is about reading it early enough to act.

What are the signals that a vendor should go?

The signals that matter are the ones tied to money, clarity, and honesty — not to how the relationship feels. A vendor can be pleasant, responsive, and still wrong for you. A vendor can be blunt and slow to reply and still be the best money you spend. Judge the work, not the vibe.

Here are the signals worth acting on:

• You cannot connect their work to revenue. Six months in, you still can't say what the spend produced in leads, calls, or booked clients.
• The reporting is all vanity. Impressions, reach, and "engagement" fill the deck. New clients do not.
• They dodge direct questions. You ask what changed last month and get process language instead of a straight answer.
• The work stalls but the invoice doesn't. Deliverables slip. Billing never does.
• They deliver and disappear. A report lands in your inbox. Nobody walks you through it. Nobody flags a problem before you find it.
• They resist measurement. When you ask to track conversions properly, they call it unnecessary or too complex.

One signal is a conversation. Two or three at once is a decision. If you see all of them, you already know.

How do I tell a slow month from a failing vendor?

You tell them apart by asking whether the vendor can explain the slow month before you raise it. A good vendor sees the dip, names the cause, and tells you the plan. A failing vendor waits for you to notice, then produces a reason on the spot.

Marketing has real variance. Search demand moves with seasons. A San Diego tax firm goes quiet in June and buries itself in April. A roofer's phone rings after the first big storm. A slow month against that backdrop is not failure — it is weather. The question is whether your vendor understands your weather.

Run this test. At your next review, do not lead with the numbers. Ask two questions. First: "What went well last month, and what didn't?" Second: "What are you changing this month because of it?" A vendor doing the work answers both in plain language and specific detail. They name the page that ranked, the keyword that moved, the form that broke. A vendor coasting gives you adjectives — "strong momentum," "great progress" — and no nouns.

The difference between a slow month and a failing vendor is not the result. It is whether someone is steering.

What signals look scary but usually don't matter?

Some signals feel like reasons to fire and usually aren't — chief among them a single bad month, a redesign that ranks lower before it recovers, and a vendor who tells you no. These read as problems. Often they are the vendor doing the job right.

SEO has a lag. When you move a site to a new structure or clean up a mess of old pages, rankings can dip for weeks before they climb past where they started. A vendor who warns you about the dip in advance is not failing. They are being honest about how search works. The failing vendor is the one who promised you would go up and to the right every single month, because nobody can promise that.

A vendor who pushes back also looks like friction and usually isn't. You want to run a discount promotion. They tell you it will train your clients to wait for discounts and cheapen your brand. That is not insubordination. That is the advice you are paying for. The vendor who agrees with everything you say is not managing your marketing. They are managing your feelings.

And one slow month, on its own, is noise. Fire on the trend, not the point. Look at the last two quarters, not the last two weeks. If the line is flat or falling across six months and nobody can explain it, that is the signal. One dip is not.

How do I run the conversation before I fire?

You run the conversation by putting the standard in writing, giving a fixed window to meet it, and defining what success looks like in numbers before the clock starts. Most vendor relationships fail because nobody ever wrote down what "working" meant. You cannot fairly fire someone against a standard they never agreed to.

Here is the sequence I use.

Write the standard down. One page. What are we trying to produce — leads, calls, booked consultations — and how many by when. Attach a dollar figure. "We spend $4,000 a month with you. We need that to produce at least eight qualified leads a month within ninety days." Specific. Dated. Measurable.

Set the window. Give a real timeframe that respects how the channel works. SEO gets ninety days minimum, often more. Paid search can show signal in three weeks. Do not set a window shorter than the channel's honest lag.

Fix the measurement first. If you can't see conversions, you can't judge anything. Before the window starts, make sure calls, forms, and bookings are tracked and attributed. A vendor who resists this step has told you something already.

Have the direct conversation. Sit down. Read them the standard. Ask if it is fair and achievable. If they say yes, the clock starts. If they say the target is wrong, listen — they may be right, and you adjust. If they say measurement is impossible, that is close to an answer.

Decide on the date, not the feeling. When the window closes, look at the number against the standard. Met it, keep going. Missed it with a real explanation and a credible plan, extend once. Missed it with adjectives and excuses, end it.

This is the work a Fractional CMO does for owners who don't have time to run it themselves. Someone who sits on your side of the table, writes the standard, holds the review, and makes the call without a stake in keeping the vendor. That last part matters. An agency will not fire itself.

What does firing well actually look like?

Firing well means you leave with your assets, your data, and your dignity — and the vendor does too. This is not a fight. It is an ending. Handled cleanly, it protects the work you already paid for and keeps a door open you may want later.

Three things to secure before you say goodbye. First, your accounts. Make sure your business — not the vendor — owns your Google Business Profile, your ad accounts, your analytics, and your domain. Owners who skip this find themselves locked out of their own website. Get admin access in writing before you give notice. Second, your data. Export the historical reporting. You want the baseline for whoever comes next. Third, the handover. Ask for a plain summary of what is live, what is scheduled, and what breaks if nobody touches it.

Then give notice against the contract terms, pay what you owe, and thank them. The professional-services world is small. A San Diego firm that burns a vendor on the way out gets a reputation faster than it thinks.

When McShanes Solicitors came to us, part of the early work was untangling what a previous arrangement had left behind — accounts nobody could access, reporting nobody could read, and no clear line from spend to signed clients. Cleaning that up came before any new work. The ending you inherit shapes the beginning you get.

If you are unsure whether the problem is the vendor or the model itself, it helps to understand when you actually need a Fractional CMO and when you don't before you replace one agency with another. And if you are choosing what comes next, the split between a strategist and a delivery shop matters — we cover it in Fractional CMO vs agency: the difference that matters.

Where this breaks down

This framework assumes you have measurement in place. If you cannot see whether leads are coming in, you cannot fairly judge any vendor — and the first fix is your own tracking, not the firing. It also assumes the vendor was ever set up to succeed. If you handed them a broken website, no clear target, and a budget too small for the channel, the failure may be yours. Fix the setup before you blame the operator. And firing does not solve a strategy problem. If you replace a vendor without knowing what you actually need, you will hire the same problem with a new logo.

Fire on the signals that matter — money, clarity, honesty. Ignore the ones that don't — a slow month, a warned-about dip, a vendor who tells you no. Write the standard down, run the conversation, decide on the date. That is oversight. That is the job.