Arcane Marketing logoArcaneApply
← All posts

Two proposals arrive in the same week for the same scope. One is a flat monthly fee on top of ad spend. The other quotes no retainer at all and a price per qualified lead, pitched as proof the agency believes in its own work. The second reads as the braver offer, and plenty of owners sign it for that reason alone.

Performance based marketing pricing covers at least four shapes, and each pays your agency for a different thing. Commission on media spend pays for spend. A per-lead fee pays for lead count. Revenue share pays on booked revenue, a number your closers move at least as much as your media buyer does. An incentive bolted onto a base fee buys attention on one agreed KPI and ignores the rest. Find the denominator in the contract and you can usually predict the behaviour before the first invoice clears. The standard fee models sit side by side in Marketing Agency Pricing.

Performance based marketing pricing is rarer than the pitch suggests

Fixed fees still run the market. Fixed fee "emerged as the most utilized compensation model" among the 149 agencies surveyed in Q1 2024 by the 4A's and ANA joint task force on compensation, across both project and retainer relationships. Performance incentives went the other way. They fell from 61 percent in 2013 to 15 percent in 2025, the lowest since the ANA started measuring in 1991.

So a founder being pitched rev-share-only is being offered a structure most large advertisers have backed away from. That is a question to ask rather than a reason to walk: what did you see in our numbers that makes you willing to carry this risk, and what happens to your service level if the bet goes against you?

Two caveats before you quote those percentages at anyone. The data is large-advertiser and agency-side (ANA members, 149 agency respondents), so it describes enterprise relationships and says little about what a business spending five or six figures a month actually signs. Dissatisfaction also looks universal regardless of model: 84 percent of marketers say they are satisfied with how they pay agencies, yet 86 percent are revisiting those arrangements within 12 months.

Read the denominator and you know what you bought

Commission on media spend. The 4A's and ANA paper lists among the cons of its own commission model that "the agency could be perceived to be motivated to recommend increasing media spend levels". When the fee is a slice of budget, the fastest path to a raise is a bigger budget, whatever blended CAC does.

Per-lead and per-action fees. The clearest public description of these mechanics sits in a Consumer Financial Protection Bureau circular, which notes that operators are sometimes paid "on a fee-per-action basis ... per click, per application, per conversion, per offer, or per sale", and describes an operator whose payment changes depending on whether a threshold volume of leads is met within a set period, and who uses steering practices to hit it. That circular covered comparison-shopping tools in consumer financial services and was withdrawn in May 2025, so it carries no authority over your agency. The mechanic still travels. Where revenue steps up at a volume threshold, the last week of the period is where lead quality gets tested.

Revenue share. Pays on a number the agency only partly controls, which means they will want influence over your sales process. Decide how much you are granting, and write down what counts as attributed revenue. The clauses people regret skipping are in SLAs and exit terms.

Incentive on top of a base fee. The common shape in the industry data and the one with the fewest surprises, because the base pays for the work while the incentive buys focus on one metric. Most of the risk moves into the choice of metric.

What a flat cost per lead is betting on

Google sells cost-per-outcome bidding, and its eligibility rules read like a list of conditions under which outcome pricing behaves. Pay for conversions in Display needs more than 100 conversions in the last 30 days, it is unavailable when the delay between click and conversion runs more than 7 days, and the target has to be under $200 USD. Even inside those limits, Target CPA is an average: some conversions cost more than the target and some cost less, and the realised figure moves with factors outside Google's control, like changes to your website or ads or increased competition in ad auctions.

Benchmarks say the same thing about stability. The average Google Ads cost per lead was $70.11 in 2025, the dearest industry average was $131.63 in attorneys and legal services, and the all-industry average rose 5.13 percent from $66.69 the year before. A fixed per-lead price is a forecast about an auction, with someone's margin riding on the forecast being right.

Per-lead pricing is fair where volume is high and the lead definition is tight enough to be worth arguing about. Where volume is thin and the sales cycle runs for months, the price turns into a lottery ticket and one of you ends up holding the losing end.

A worked model to re-run on your own numbers

Assumptions, chosen for the arithmetic rather than drawn from anyone's account: $40,000 a month in media at a $100 raw cost per lead, so 400 raw leads. Assume a quarter clear your qualification bar, giving 100 qualified leads. Compare a flat $8,000 fee against a per-lead deal with no base at $110 per qualified lead.

The flat fee costs $8,000. The per-lead deal costs 100 x $110 = $11,000, so it is $3,000 more expensive this month. Break-even sits at 8,000 / 110, about 73 qualified leads. Below that the per-lead deal is cheaper for you, above it the fee is.

Now move the subjective input. Hold media and raw CPL where they are, and loosen the qualification rate from 25 percent to 35 percent. Qualified leads go to 140 and the invoice goes to $15,400, so the agency earns $4,400 more with no change in spend and none in booked revenue. These numbers are a model, so re-run it with your own CPL, your qualification rate and the price you have been quoted.

That is why the definition of a qualified lead deserves more negotiating time than the price per lead. Settle which system holds the record, which field marks qualification, who can change that field and how often you both review it. Reading the resulting report like an operator is a separate skill, and we wrote that one up too.

Conditions that make a performance deal fair

Appetite is not the blocker. 85 percent of respondents expect outcome-based remuneration to grow and 81 percent believe outcome-based models can strengthen client-agency relationships through greater alignment and shared accountability, while 69 percent cite attribution challenges as a major barrier to adoption. On the advertiser side, only 28 percent report adequate visibility into how their agencies make money, and 84 percent cite absent shared measurement as a barrier to change. Settle whose number counts before you settle the percentage.

What the industry bodies say a workable incentive looks like:

  • The metric is one the agency controls or materially influences.
  • Metric, target and measurement methodology are agreed in writing before the deal starts rather than reverse-engineered from the first report.
  • Review runs at least quarterly, with you sharing past, current and projected performance so targets get reset on evidence.
  • Awards come in incremental bands instead of an all-or-nothing cliff.
  • There is enough shared history for a baseline. The guidance points at roughly a year of working together before layering an incentive on top.

One line is worth reading aloud to anyone who wants heavy downside in your deal. The 4A's and ANA paper warns that the risk element and the chance of a downside should be carefully weighed, because an improperly designed program "can backfire with the agency reducing levels of service when a downside is inevitable". A deal that punishes the agency hard enough gets paid for out of your account management.

Nothing in the data above shows that performance pricing produces better marketing than a fee does, and anyone claiming otherwise is selling. What it does buy is alignment and a share of the risk, which is worth paying for when the metric is real and the measurement is joint. Pick the model whose denominator matches the outcome you actually want, agree the measurement before the percentage, and keep enough base fee in the deal that service survives a bad quarter. If you want that read on a proposal sitting in front of you, apply and bring the contract.

Sources

Stuck above the $10k ceiling?

We diagnose your account in 24 hours and tell you straight whether it's a creative, audience, or sales problem.

Talk to us