Agency Fees Scale With Spend. Do the Hours? The Math From $5k to $200k a Month

At $200,000 a month in Google Ads spend, a 15% management fee is $30,000 a month. That is 10 times the fee on $20,000 in spend, even though the search-term reviews, budget checks and reporting do not take 10 times as long. The 10–20% range commonly quoted for agency management makes 15% a useful midpoint, not an unusually expensive quote. It is also a fee paid on top of what you pay Google.

I used to defend this model. I told clients that more spend meant more risk and more work. It sounded reasonable until I tracked my hours. So I lined up WebFX’s published spend-based tiers, a reported $5,000 minimum for Directive and a reported $3,000 starting price for KlientBoost against an illustrative task-by-task workload at $5k, $20k, $50k, $100k and $200k a month. Those are comparison points, not five observed client accounts. The question is whether the fee follows the work, or merely follows the budget.

Three ways a management fee follows your budget

Most quotes I have seen reduce to three structures:

• Percent of spend: The agency charges a share of the ad budget, often within the 10–20% range. You pay that fee in addition to media spend.
• Spend-tiered retainer: The quoted monthly fee changes when your budget crosses a band, sometimes switching to a percentage above a threshold.
• Flat retainer: Management costs a set amount. You still need to check which work, especially landing pages, sits outside it.

WebFX gives us a published example of the second model. Its Lite tier costs $750 a month for $100 to $5k in spend; Pro costs $975 or 15% for $5k to $30k; and Enterprise costs $4,500 or 12% above $30k, whichever amount is greater in each case. It also lists setup fees of $1,200, $2,250 and $5,800 by tier. Landing-page performance testing is a separate $1,800-a-month add-on.

That last line changes what the management price means. If page testing sits outside the retainer, the base fee does not buy the whole job of improving conversion performance. A comparison of agency offers also describes Directive’s landing-page testing as separately priced and says it publishes no prices or contract length. Read the scope beside the price. A lower-looking management line can leave the work that affects CPA on another invoice.

What the other published prices do—and do not—tell us

Directive and KlientBoost are useful reference points, but neither published starting figure is an hourly rate. They describe different offers for different buyers. Treating them as interchangeable with a Google Ads management quote would make a tidy table and a bad decision.

Independent reporting puts Directive at a reported $5,000 minimum retainer and a $5,000–$6,500 monthly startup package. The same review describes typical disclosed projects at $10k–$49k and its fit as funded B2B SaaS buying multi-channel demand generation, rather than small or single-channel budgets. I read that as a filter, not a flaw. On a hypothetical $8k monthly Search budget, a $5k retainer would equal 62.5% of ad spend. If all you need is Search management, the minimum tells you to look elsewhere.

A SaaS agency roundup lists KlientBoost as starting at $3,000 a month with PPC, landing-page A/B testing and CRO bundled; it also reports a $4,000–$6,000 base and $5,000 typical monthly price. The bundle matters more than the starting-number contest. WebFX lists page testing as an add-on; KlientBoost’s reported offer includes it. Those are different scopes.

I used to tell clients that creative and landing pages were someone else’s job. I was wrong. For lead gen and SaaS, a page can explain more of a CPA move than a bid adjustment. Before comparing retainers, ask who owns the page test. Otherwise you are comparing two prices for two different jobs.

The work curve: my estimated hours at five spend levels

Here is the floor that rarely makes it into a proposal. An agency operations guide puts even a small account at 8–12 hours a month and onboarding at another 15–25 hours of senior work. It recommends percentage pricing above roughly $15k a month in spend and a $1,500–$5,000 flat fee below that.

My logs have generally followed the same shape: onboarding is heavy; steady-state checks are lighter; more spend can mean more volume without creating a new kind of task. The table is my illustrative operator estimate, anchored to that published small-account floor. It is not a study of agency timesheets. I use four weeks per month to turn the weekly tasks into the monthly totals, and I average lumpy structure, tracking and page work across steady state. Month-one builds and separate landing-page projects are not in these totals.

An account with more countries, product lines or feed complexity can take substantially longer. These columns hold that complexity roughly steady to isolate the pricing question. The useful distinction is not that larger budgets create no work. They do. It is that the work need not rise in proportion to the money passing through the account. Ask a prospective agency which tasks grow when your budget grows, and by how much.

The effective hourly rate rises faster than the hours

For the next table, I multiply each illustrative spend level by 15%, then divide the fee by the estimated steady-state hours above. I also apply WebFX’s published tier rules to the same hours. Every effective rate is arithmetic on those two inputs, not a rate an agency has quoted or evidence of what any particular team earns. Setup fees, onboarding and separately billed page work are excluded.

From $5k to $200k, spend and the illustrative 15% fee rise 40-fold. Estimated hours rise from 11 to 31, less than threefold. The calculated effective rate rises from about $68 to $968 an hour. WebFX’s lower percentage above its threshold softens the high-end figure, but it does not make the hours track the fee.

Minimums can reverse the pressure at the low end. One operator’s comparison describes agency fees of 15–20% with $3k–$10k minimums. A minimum can make the effective percentage higher than the headline rate on a smaller account. Check the greater of clause before treating a quoted percentage as the price.

There is another limit to this calculation. It divides the fee by all estimated management hours, including routine checks and reporting. It does not measure what you pay per hour of real decisions. If only some of those hours involve deciding what to change, the rate for decision-making time would be higher. I cannot put a defensible number on that without an account’s actual work log. The table gives you the question to ask for it.

When percent of spend still earns its keep

I will defend percent-of-spend pricing in two circumstances. First, a small account can be expensive to serve relative to its budget. At $5k a month, 15% produces a $750 fee; against my estimated 11 hours, that is about $68 an hour before the agency’s overhead. The small-account workload does not vanish because the ad budget is small. That is one reason minimum fees exist.

Second, the scope may genuinely grow with spend: more countries, product lines, margin tiers, feed work or landing-page tests tied to traffic volume. If a proposal names the added work at each band, the percentage can be shorthand for a larger job. The broader structures are covered in this pricing-models guide. A percentage is easier to defend when the proposal says what grows with it.

The problem is a rising fee on an unchanged scope. An operator serving smaller stores warns that the person who sells the work may not be the person doing the weekly account work. That does not prove your account gets a junior handoff. It tells you to ask who will read the search terms.

The incentive problem is plainer. The same operator describes cutting a garden store’s spend from $20,000 to $4,000 a month while revenue rose 42% after traffic went to a quiz rather than category pages, then notes that a percentage-of-spend agency would earn less for recommending that cut. One account is not a universal result. The fee arithmetic, though, is unavoidable: spend falls, so the percentage fee falls. If your agency has never recommended spending less, ask what rule would make it do so.

Ask for the work behind the quote

Ask for a price tied to scope, not just dollars spent. A useful quote tells you:

• who touches the account and who makes the decisions;
• how often someone reviews search terms, budgets and anomalies;
• who builds and tests ads, feeds and landing pages;
• what changes when spend crosses the next pricing band; and
• which work triggers a separate fee.

If landing-page testing is excluded, price it explicitly, as WebFX does with its add-on. If it is included, say what testing the fee covers. For the B2B retainer-plus-media version, this Directive Consulting review examines what its minimum filters for.

For a different pricing benchmark, groas lists a flat $1,499 a month per service regardless of ad spend. That does not make every scope equivalent. It gives you a concrete comparison: when a percentage quote costs more, ask what additional work and decisions the difference buys.

Then demand an action log, not just a monthly deck. CTR screenshots and a paragraph about “optimizations” do not tell you much. I want dated entries: a keyword added, a bid target changed and why, a budget moved, a test started or stopped. The log lets you compare the promised work with the work done. If the quote cannot show what decisions the fee covers, the percentage alone is not enough information to sign.

The number to take into the decision

As a quick quote check, divide the proposed monthly fee by 20 hours. That gives you a common yardstick, not the account-specific rates in my table: my estimates range from 11 hours at $5k in spend to 31 at $200k. For the modeled rates, use the hours your account actually needs. Ask the agency to show its own estimate, including strategy, page work and the first month’s build.

My hours are estimates anchored to a published 8–12-hour small-account floor, not a prediction for your account. The published fee rules tell us what a buyer may be charged; they do not tell us how many hours any named agency will work. That is where the data runs out, and where the quote and action log have to take over.

Stop asking only, “What percentage?” Ask what you are paying per hour of real decisions, get the page work in writing, and choose a fee structure that does not rise with spend unless the work rises with it.