The offer is attractive by design: you pay nothing until the rankings arrive, so the risk appears to sit with the agency. In practice the model changes what gets optimised rather than who carries the risk, because whoever defines the result defines the invoice. Some performance-based arrangements are entirely legitimate and well run. Many are structured so that success is nearly automatic on terms that do not produce customers. This page explains how these deals are built, which clauses decide whether one is fair, and what to check before signing anything that promises payment only on results.
How these deals are actually structured
There are three common shapes. Pay per ranking charges a fee for each agreed keyword that reaches a defined position, usually billed monthly for as long as it holds. Pay per lead or per call charges for each enquiry attributed to organic search. Revenue share takes a slice of sales tracked back to the channel. The first is the most commonly advertised and the most easily gamed, because the agency frequently chooses the keyword list, and a list of low-demand phrases will hit page one quickly while producing nothing. The second is closer to the buyer's actual interest but depends entirely on how a lead is defined and who counts it. The third is the fairest in principle and the hardest to administer, because it requires the agency to have visibility of your revenue data and both parties to trust the same attribution. Ask which shape is on offer, then ask who picks the targets.
Why the metric usually favours the agency
Rank is a much weaker measurement than it appears. Results differ by location, by device, by search history and by the moment of the query, so two people searching the same phrase at the same time see different pages, and a vendor reporting from a favourable configuration is not necessarily lying so much as choosing. That ambiguity is why the ranking metric needs to be nailed down in the contract: which tool, whose account, which location setting, which device, checked how often, and averaged over how many days. Google's own guidance on hiring search help warns buyers to be skeptical of anyone guaranteeing a number one ranking, and its local ranking guidance states plainly that there is no way to request or pay for better local placement. A model whose entire pricing depends on a number the platform says cannot be purchased deserves a careful read of the definitions.
The clauses that decide whether it is fair
Five things matter more than the headline. Who selects the keyword list, and can you veto phrases with no commercial value? What is the minimum monthly volume of a qualifying phrase, and is it verified against a source you can both see? How is a result measured, in the terms set out above? What happens to the fees if a ranking is lost, and is there a floor charge regardless? And, most importantly, who owns the work at the end: many pay on results contracts place the pages, the links or even the domain in the agency's control, so ending the arrangement removes the asset you were paying for. A performance deal with an unfavourable ownership clause is not low risk, it is deferred risk with a bigger bill. Ask for the termination and ownership sections before the pricing table.
When performance pricing is genuinely reasonable
It works best where the outcome is unambiguous and cheap to verify: qualified inbound calls with recorded audio, or booked appointments in a shared calendar, in a category with sufficient demand that both parties are confident volume exists. It works badly where the deliverable is a rank on a phrase nobody searches, or where the buyer cannot audit the count. A reasonable middle ground exists and is common among serious firms: a modest base fee that covers real production labour, plus a performance component against an outcome both parties can see. That structure keeps the agency solvent enough to do the underlying work, which in practice is the same technical, on-page and content work any other engagement requires, while still tying part of the fee to something the buyer actually values.
Questions people ask about pay on results seo
Is pay on results SEO a scam?
Not inherently. The model is only as good as the definition of the result and the ownership terms. Deals that let the agency pick the keywords, measure them itself and retain the assets on termination are the ones that reliably disappoint.
Why do most established agencies refuse it?
Because the work requires months of paid labour before any result appears, and because much of what determines the outcome, site changes, approvals, product and pricing, sits with the client. Refusing performance-only pricing is not evidence of low confidence; it is usually cash flow arithmetic.
What is a fairer alternative?
A modest base fee covering production plus a bonus tied to an outcome both sides can audit, such as qualified calls or booked appointments. Define the qualification rules in writing, including who arbitrates a disputed lead.
What is the single most important clause?
Ownership on termination. If the pages, links, tracking or domain do not transfer to you, the low up-front price is buying rented rankings, and the rent rises the moment the results are good.