How to Evaluate Your Marketing Agency’s Performance

Published On: November 4, 2026Categories: Blogs

To evaluate your marketing agency’s performance, judge it on numbers that end in revenue: cost per real lead, lead-to-close rate, and cost per new customer. Pull those numbers from your own CRM or call log, not from the agency’s platform report. Then fill in the same scorecard every month and watch the trend.

Most agency reports are full of numbers that look like results and aren’t. A “conversion” might be someone who stayed on a page for a minute. If you can’t read those reports, you end up judging the agency on gut feel, or on how much you like the account manager. Below are the marketing agency KPIs worth tracking, how to read conversion numbers, realistic funnel rates, and a scorecard you can fill in yourself.

Key takeaways

  • Three numbers do most of the work: cost per real lead, lead-to-close rate and cost per new customer.
  • Before you trust any conversion number, find out what it counts. A real lead is always a whole number.
  • Judge extra spend by what the last dollars bought (incremental cost), not by the monthly average.
  • Use realistic funnel rates to set targets, then score the agency against them every month.

The marketing agency KPIs that matter

Picture yourself signing the agency’s check at the end of the month. What number does your mind land on? If it’s something you care about, like new customers or booked jobs, the agency is measured on the right thing. If it lands on nothing, or on “80,000 impressions,” you’ll resent that check no matter how hard they work. I call it the check-writing test, and it’s the fastest way to sort useful KPIs from noise.

The numbers that pass it:

Cost per real lead

Total agency spend (fees plus ad spend) divided by the number of leads a human being on your team would call a lead. A phone call longer than a minute from a new number, a form with a real name and a real need, a booked appointment. Not newsletter signups, not chat widgets opened and closed.

Lead-to-close rate

Of the real leads, how many became paying customers? This one is shared: the agency controls lead quality, your sales team controls follow-up. If it drops, find out which side moved before you blame anyone.

Cost per new customer

Total spend divided by new customers from marketing. Compare it with what a customer is worth to you over time, and you know whether the math works.

Revenue from marketing

If your CRM tags the source of each sale, this is the cleanest number of all. If it can’t yet, put that on the list.

Click-through rate, cost per click, bounce rate and impressions help the agency diagnose problems. They don’t belong at the top of your report.

Metric Type What it tells you Where it belongs
Cost per real lead Revenue-tied What you pay for a conversation with a possible buyer Top of the report
Lead-to-close rate Revenue-tied Whether the leads are the right people, and whether sales follows up Top of the report
Cost per new customer Revenue-tied Whether the marketing pays for itself Top of the report
Click-through rate Activity Whether the ad message gets attention Agency working notes
Impressions Activity How many times the ad was shown, not who cared Agency working notes
Followers and likes Activity Very little about sales Footnote, if anywhere

One prerequisite: the ad accounts, analytics and call tracking should sit under a login your business owns, so you can check the numbers yourself.

How to read the conversion numbers

Most agency reports lead with “conversions.” That number fails in three common ways.

Find out what is being counted

A conversion is whatever someone set up as a conversion in the ad platform or analytics. That can be a form fill. It can also be a page view, a scroll, or someone spending two minutes on the site. Those are engagement goals, not leads. Ask the agency for a list of every conversion action in each account and what triggers it. This is what I call clarity of conversions, and in my experience it’s the first thing that falls apart when someone checks.

A decimal means it isn’t a person

If the report shows 37.62 conversions, something other than people is being counted. You can’t have 0.62 of a customer. Decimals usually come from attribution models splitting credit, or from engagement events mixed in with leads. When you see one, find out what’s behind it and have the agency move that action to secondary so it stops inflating the total. The rule is simple enough to remember: don’t trust the fraction.

A zero doesn’t mean nothing happened

The opposite mistake is just as expensive. A channel can show zero attributed conversions while still sending buyers who come back later through other routes, so pause it for a set period and watch your total leads before you cut it. That’s the idea behind don’t trust the zero.

Real leads, counted as whole numbers, also show you when a campaign is broken. In one agency case study, the ads kept running and drawing clicks while the landing page talked past what searchers wanted. A count of real leads exposes that kind of problem much faster than a click report does.

Blended cost vs. incremental cost

When an agency asks for a bigger budget, it will usually show you the average cost per lead. That average is called blended cost, and it hides the thing you most need to know: what the extra money bought.

Each additional dollar in a channel tends to produce less than the one before it, because the most ready buyers get reached first. That’s the law of diminishing returns, and here’s how it shows up in a report.

Example only: in March you spend $10,000 and get 80 real leads, so your cost per lead is $125. In April the agency raises spend to $14,000 and you get 92 leads. The blended cost per lead is about $152. Up some, nothing alarming. But look at the increase alone. The extra $4,000 bought 12 extra leads, which is about $333 per lead. If the most you can afford to pay for a lead is $200, the last $4,000 lost money, and the blended number made it look fine. A lot of wasted marketing budget hides exactly like this, inside an average nobody questions.

So ask for incremental cost any time spend goes up: extra dollars divided by extra leads. Warning signs include cost per lead rising faster than lead volume, the same people seeing your ads over and over, and revenue flat while spend climbs. Scale until the incremental cost per lead reaches what you can afford, then stop.

Realistic funnel benchmarks

You can’t score an agency without targets, and the targets have to be realistic. In my experience, each step of the funnel converts at a fairly predictable rate. From lead to qualified to closed, roughly a third makes it through each stage, which is why I call it the law of thirds. These are the working ranges I use:

  • Impression to click: 0.5% to 3%
  • Click to lead: 10% to 30%
  • Lead to close: 20% to 35%

These are planning ranges, not promises, and your industry and price point will move them. They still tell you which stage is broken. If clicks are healthy and almost nobody becomes a lead, the problem is usually the landing page or the offer, not the ads. If leads are healthy and almost nobody closes, look at lead quality and at your own follow-up.

Use them to work backward from the customers you need. Example only: you want 10 new customers a month. At a 25% lead-to-close rate, you need 40 leads. At 20% click-to-lead, that’s 200 clicks. At a 2% click-through rate, that’s 10,000 impressions. Now you have a target for every stage, and a way to see where the agency is falling short. Raising lead-to-close from 25% to 33% would cut the leads you need from 40 to about 30, without spending a dollar more on ads. Fixes like that are the cheapest way to improve your marketing ROI, because they make the spend you already have work harder.

A monthly scorecard you can fill in

This is the marketing agency scorecard I’d want on my desk every month. Most of the numbers come from your own records, which is the point.

Line Where it comes from Target This month Last month Notes
Main KPI (one number, in writing) Agreed with agency Has it changed?
Total spend (fees + ads) Invoices
Real leads CRM or call log Whole numbers only
Cost per real lead Spend / real leads
Incremental cost per lead Extra spend / extra leads Only when spend changed
Leads closed CRM or sales team
Lead-to-close rate Closed / real leads 20% to 35% Agency or sales issue?
Cost per new customer Spend / leads closed
Revenue from marketing CRM source field
Platform conversions Agency report Any decimals? Any zeros?

Fill it in yourself, or have someone on your team do it, before the agency meeting. If platform conversions and real leads drift apart, ask why. A single month tells you little. Three months of the same line moving the same direction tells you a lot.

How often to evaluate

Monthly for the scorecard. Quarterly for decisions.

A monthly look catches broken tracking or a spike in cost per lead before it eats a quarter of budget. But one month is noisy, especially if you close only a handful of deals. Use the quarterly review to decide what to cut, test and scale.

A brand new campaign needs time to collect real leads before you can judge it on closed sales, so agree up front on what an early read looks like (lead volume and lead quality) and when you’ll judge it on customers. And keep the strategy moving. If nothing changes for six months, you get no new data to make decisions from.

Contracts, account ownership, meeting format and what happens when the numbers miss are governance, and I cover them in how to hold your marketing agency accountable. The short version: open every meeting with the main KPI and close by confirming it. If you’d like to see how a fractional CMO handles that relationship, here’s how I work alongside an existing ad agency.

When the numbers will mislead you

Judging an agency purely on these KPIs is the wrong move in a few situations.

Your tracking is broken

If calls aren’t tracked or forms don’t reach the CRM, your scorecard will punish the agency for your own data gaps. Fix the plumbing first. A money leak audit is one way to find where leads and spend are disappearing.

Your sales cycle is long

Deals that take six months to close make monthly cost per new customer swing wildly. Lean on cost per real lead in the short term and judge closed revenue over a longer window.

The problem is on your side

When lead-to-close drops while lead volume and quality hold steady, look at how fast and how well your team follows up. An agency can’t close a lead nobody called back.

The agency’s job isn’t leads

An agency hired for brand work, PR or a website build shouldn’t be held to cost per lead. Agree on a KPI that fits the work, tied to revenue as closely as you honestly can.

The downside of all this is time. Somebody has to pull the numbers every month, and the first few scorecards will raise more questions than they answer. That’s still cheaper than paying an agency for years without knowing whether it worked. Some owners hand the job to a part-time marketing executive instead. If you do, score that person too: here’s how to judge a CMO’s results by the same standard.

Check your own blind spots first

Before your next agency review, it helps to know what to look for. The Blind Spot Brain loads my marketing laws into ChatGPT, Claude or Gemini, so you can paste in your agency report, your ads or your numbers and ask what doesn’t add up. The laws behind it are explained on the marketing laws hub.

Frequently asked questions
A KPI is a performance indicator, a number you agree to judge marketing by. Good marketing KPIs connect to revenue: cost per real lead, lead-to-close rate, cost per new customer and revenue from marketing. Clicks, impressions and followers are activity measures, not KPIs.
Start with three: cost per real lead, lead-to-close rate and cost per new customer. Add revenue from marketing if your CRM can track it. Pick one of them as the main KPI, put it in writing, and have every report start with it.
Compare what you spent with the agency against real leads and closed customers from your own CRM or call log, not the agency’s platform report. Check what each conversion counts, judge new spend by its incremental cost, and fill in the same scorecard every month.
It depends on what a customer is worth to you. Divide the most you can afford to pay for a new customer by the number of leads it takes to close one. That is your allowable cost per lead. Industry averages tell you far less than that number does.
Fill in a scorecard every month and look for the trend. Do a deeper review each quarter, where you decide what to cut, test or scale. Give a brand new campaign enough time to collect real data before you judge it on closed sales.
Spend, real leads, cost per real lead, incremental cost per lead, leads closed, lead-to-close rate, cost per new customer and revenue from marketing, plus the main KPI and whether it changed. Each line should show this month, last month and the target.
Not on that number alone. Zero attributed conversions can mean broken tracking, or ads that educate buyers who come back through another channel. Test it first: pause the channel for a set period and watch what happens to your total leads.
Get a second read on your agency’s numbers

If you’d like someone who doesn’t sell ads to go through your agency reports and your real lead data with you, book a free 2-hour strategy session. You’ll leave knowing which of your numbers you can trust and where the money is leaking. Rather talk now? Call 214.466.8332.

Phoenix Build Expo 2026
Upcoming Build Expo

Phoenix Build Expo 2026

📅 December 9–10, 2026

📍 Phoenix Convention Center - F & G · Phoenix, AZ

Reserve My Seat View Conference

Free · In-person · Limited seats