Performance marketing pricing: the billing models in practice

Performance marketing gets sold like a promise: you only pay when it works. Sounds fair. In practice it’s considerably more complicated — because performance marketing billing isn’t only about formulas and percentages. It’s mostly about one thing nobody likes to say out loud: trust.

My position up front: the moment the invoice hangs on results, every invoice gets argued over — so a genuinely performance-based model only works with trust and real experience in the sector. Anything else is a blind flight, for both sides. In this article I go through what “performance-oriented” actually means, which billing models exist in concrete terms, and why sector experience and trust end up deciding everything.

1. What “performance-oriented” really means — and where the billing question starts

Two terms constantly get thrown into the same pot, even though they mean completely different things:

• Working performance-oriented means data-driven. You measure cleanly, you optimise for real outcomes — revenue, leads, profit — instead of clicks and reach. That’s a way of working.
• Performance-based billing means the agency’s fee hangs on exactly those outcomes. That’s a payment model.

The catch: almost every performance marketing agency calls itself “performance-oriented” today. The vast majority still work on a classic fixed fee — the retainer. And that’s completely fine. “Performance-oriented” is a standard for the work, not an automatic promise that you only pay on success.

The distinction matters because a lot of clients walk into the conversation with the wrong expectation. They hear “performance” and think “I’ll pay them once revenue comes in”. The agency means “we optimise for revenue instead of vanity metrics”. Two different worlds — and the most common reason it blows up later when the invoice arrives.

One more thing belongs to being honest here: billing on success assumes “success” can be measured cleanly in the first place. Without reliable conversion tracking there’s no number to hang a fee on. If your tracking counts micro-actions like “page view” as a conversion, or double-counts leads, then every performance-based invoice is a fight waiting to happen. Data first, model second.

2. The billing models in practice

Now the specifics. In practice almost every arrangement comes down to three base models, plus hybrids. I’ll go through them one by one, honestly, with the upsides and downsides.

Fixed fee + percentage (revenue or profit share)

The most robust performance-based model, and widespread in e-commerce:

• A fixed base amount covers the ongoing work — setup, management, reporting. It means the agency doesn’t start from zero and can keep working through weak months.
• On top comes a percentage of the revenue, or — if tracking is clean — of the profit the campaigns generate.

The upside: the incentives point roughly in the same direction. You sell more, the agency earns more. The fixed portion stops the agency walking away in a weak phase.

The honest downside: revenue isn’t profit. A percentage of revenue can get expensive for you if your margin is thin. Cleaner — but technically more demanding — is a profit share, billed on the actual contribution margin. That requires cost of goods, returns and ad spend to genuinely land in the tracking. This is exactly where a lot of models fall over in reality.

If you want to work out at what ROAS a model like that adds up for you, my ROAS calculator at visnakovs.de/tools/roas-rechner (in German) is a good starting point — it shows how much revenue per advertising euro has to come in before anything is left at the bottom.

Pure performance (per lead, or by lead volume)

Here the fixed portion disappears. You pay exclusively for results — in lead-gen that usually means pay per lead. Two variants:

• Per lead: you pay a fixed amount per qualified lead. €40 per enquiry, €120 per appointment — depending on the sector and what a customer is worth.
• By lead volume: a package gets agreed, say “30 leads a month”, and billed on that.

To clients this sounds like the dream model: no risk, nothing paid up front. In practice everything hangs on one single question: what counts as a lead? An email address? A submitted form? A conversation that actually took place? If that definition isn’t nailed down in detail beforehand, the argument is pre-programmed. The client thinks the leads are too poor, the agency thinks the client is too slow to follow up — and you end up debating every single invoice.

For the agency, pure performance is also a real financial risk. It goes in first: ad budget, work, time — all before the first payment. That only adds up if it knows very precisely what a lead in that niche costs. Anyone who doesn’t know the typical click and lead prices in their sector is calculating blind. The Google Ads calculator at visnakovs.de/tools/roas-rechner#budget (in German) at least helps you get a feel for how many leads a given budget realistically produces — and from there, whether a per-lead price is fair.

Guarantee and risk-reversal models (“no leads, no invoice”)

Guarantee models are pure performance taken to its extreme. Two forms are common:

• Minimum lead guarantee: the agency guarantees a floor — for example “at least 20 leads a month, otherwise we keep working for free”.
• “No leads, no invoice”: full risk reversal. Nothing comes in, you pay nothing.

For you as the client that’s maximum security. But be honest with yourself: every guarantee has its price. Either the per-lead price is higher (the agency has to price in its risk), or it only offers the model when it’s extremely sure of the outcome. And it’s only sure in a sector it already knows.

Which is exactly why a “no leads, no invoice” promise from an agency that has never worked your niche is a warning sign rather than a selling point. Either the price is set high enough to cover its risk — in which case you overpay when it works. Or it has underestimated the risk — in which case it walks after two bad weeks. A serious guarantee model rests on experience, not on nerve.

3. Why trust and sector experience decide everything

Now to the core — my actual position on all of this.

Billing on performance is only really possible for an agency after deep experience in the sector. Not because the formulas are hard, but because the model needs two ingredients you can’t buy: sector knowledge and trust.

The honest reason almost no good agency offers pure performance to an unfamiliar client in an unfamiliar niche:

• Clients often don’t pay, or argue about it. This is the single most common way performance-based models fail in practice. As soon as the invoice hangs on results, every result gets negotiated. Was the lead good enough? Was it really “through the agency”, or would it have come anyway? Was follow-up fast enough? Every one of those questions is a potential refusal to pay.
• Without sector experience the agency calculates blind. If you don’t know the lead prices, the seasonality and the conversion rates of a niche, you can’t make a fair performance-based offer. You’re guessing. And guessing, in a model where you go in first, is the fastest route into the red.
• Trust is the real currency. Performance-based models assume both sides stay fair when things don’t run well — and that nobody bails at the first dip. That doesn’t appear on demand. It builds over time, over promises kept, over clean numbers.

So in practice, pure performance and guarantee models are almost always either the product of a relationship that already exists — you worked together on a fixed fee first and know each other — or they come from an agency that has worked that one niche for years and knows its numbers in its sleep.

What does that mean for you as a client, practically?

• If someone promises you “no leads, no invoice” from a standing start in a sector they clearly don’t know: be careful. Ask for concrete references from your niche.
• A clean fixed fee + percentage is often the more honest and, long-term, the cheaper model compared with an inflated per-lead price that has the agency’s risk baked in.
• The best way into a performance-based model is almost always to start small and with a fixed portion — then raise the success component once both sides can see the collaboration holds and the numbers are right.

And the ground condition for all of it stays the same as in section 1: clean tracking. Without solid numbers neither side can trust the model — and without trust no performance model works.

The bottom line

Performance marketing billing is neither a trick nor a magic formula. It’s a question of picking the model and having the trust. The short version:

• Performance-oriented ≠ paid on performance. One is a way of working, the other a payment model — don’t confuse them.
• The three base models: fixed fee + percentage (revenue or profit share, strong in e-commerce), pure performance (pay per lead or by lead volume), and guarantee models up to “no leads, no invoice”.
• Clean tracking is the entry ticket. Without measurable success there’s nothing to hang a fee on fairly.
• Trust and sector experience decide it. Honest performance models only get offered by people who know the niche — everything else is a blind flight that ends in an argument over every invoice.

Run every model through the numbers yourself before you sign — the Google Ads calculator at visnakovs.de/tools/roas-rechner (in German) shows you in a few minutes whether the maths works for you or only for the agency.

And if you want an honest read on which billing model fits your situation and your sector, write to me at visnakovs@clickspire.de.