Outcome-Based Pricing for Talent Attraction: What Counts as a Billable Outcome

Outcome-based pricing for talent attraction means you pay for a defined recruitment result rather than for media time or ad space. The billable result is usually a click, a lead form submission, a completed application, a qualified applicant delivered into your ATS, or in rare cases a hire.

The part that decides whether the model saves you money is not the rate. It is the contract definition of the billable outcome and whose measurement system decides that the outcome occurred. Two vendors can quote the same cost per qualified applicant and deliver results that differ by a factor of three, purely because one of them counts an applicant at the moment a form is submitted and the other counts it only after four conditions are met.

What is outcome-based pricing in talent attraction?

Outcome-based pricing shifts the unit of purchase from exposure to result. Under a fixed-fee or job slot model you buy a posting for a period of time and carry all the performance risk yourself. Under an outcome-based model the vendor carries part of that risk, because they only invoice when the agreed event happens.

Recruitment advertising moved this way because paying for exposure prices badly at volume. Buying clicks does not buy applications, so a click-priced campaign can hit every media target and still miss the hiring target.

The risk transfer is real but partial, because the vendor prices that risk into the rate. A cost per qualified applicant rate is always higher than a cost per click rate for the same role, since the vendor absorbs the conversion rate between the two. You are buying predictability, not a discount.

Which outcomes can you actually buy?

Five billable events are in common use. They are not interchangeable, and the gap between them is where budget leaks.

These are contract terms, not marketing labels, and the usable ones are written down. Wonderkind's terms and conditions define each billable event separately: a lead is a form "submitted by a candidate on the Channel," while a qualified applicant must satisfy four named conditions. Ask any vendor for the equivalent sentence before you compare rates.

Note the word "completes." Some platforms bill on application starts instead. Indeed's terms state that "charges are solely based on Indeed's Apply or Apply Start measurements." An apply start is a candidate who began the flow, not one who finished it. At a 60 percent completion rate, a cost per apply start of 12 dollars is really 20 dollars per completed application. The rate card will not tell you this. The billing definition will.

Who counts the outcome, and why does that matter more than the rate?

In almost every outcome-based recruitment contract, the party that gets paid also operates the measurement system. That is not automatically a problem, but it is the term to read first. Vendors that publish it are easier to evaluate. Wonderkind's terms and conditions state for cost per click that "a click is recorded and counted by the relevant Channel's attribution system. The Channel's count is authoritative." That sentence tells you the vendor is passing through Meta, TikTok, or Google numbers rather than generating its own, which is verifiable against your own channel reporting.

Attribution rules change the count as much as the definition does. A conversion window is, in Google's definition, "the period of time after an ad interaction (such as an ad click or video view) during which a conversion... is recorded," and it is configurable from 1 to 90 days, defaulting to 30 days for click-through conversions. Google states the consequence plainly: "if you set your conversion window to 7 days, any conversion that happens more than 7 days after the ad interaction won't be recorded." Identical traffic billed on a 1-day window and a 30-day window produces two different invoices. Ask for the window length in writing.

Three questions settle this in any vendor conversation. Which system of record produces the number you invoice against? What is the attribution window, and can I see the same number in my own dashboard? If your count and my ATS count disagree, which one governs?

What has to be in the definition of a qualified applicant?

Cost per qualified applicant is the most attractive model for high-volume frontline hiring and the easiest one to write badly. "Qualified" with no definition attached means the vendor decides after the fact.

A usable definition names the conditions and says that all of them must hold. Wonderkind's terms define a qualified applicant as a candidate who "(i) completes the application flow, (ii) passes all knock-out questions as configured by the Customer, (iii) meets the job-description criteria as configured by the Customer, and (iv) is delivered into the Customer's ATS," adding that "all four conditions must be met."

Whether or not you buy from that vendor, that is the shape to insist on. It does four things a loose definition does not:

• It puts the criteria under your control rather than the vendor's, since you configure the knock-out questions and the job description criteria.

• It requires completion, not initiation, so drop-off is the vendor's cost.

• It requires delivery into your ATS, which means an applicant you cannot see is an applicant you do not pay for.

• It uses "all," which closes the gap where a candidate meeting three conditions gets invoiced.

For frontline roles the criteria that actually predict an interview are availability for the shift pattern, travel distance, required certifications, and work eligibility. Those are objective and cheap to check before the application. Wonderkind's product page describes screening flows that "assess availability, intent, and role-specific criteria" with candidates delivered "directly into your ATS."

Keep the gating rules-based rather than scored. A knock-out question with a fixed threshold is auditable and easy to defend. A model that ranks candidates is neither, and it pulls the arrangement into automated decision-making territory you probably did not intend to enter with an advertising contract.

What happens when the count is wrong?

Every measured-outcome model produces some outcomes that are not real. The mature ad channels treat this as routine, and their published policies are the best available template for what your recruitment contract should say.

Google's definition of invalid clicks covers "clicks on ads that aren't the result of genuine user interest, including intentionally fraudulent traffic and accidental or duplicate clicks," and states plainly: "you won't be charged for invalid clicks or impressions as they provide little or no value." The remedy is narrower than most advertisers assume. Google's invalid traffic policy says that "clicks determined to be invalid will result in adjustments or credits, not a refund," with post-invoice detections appearing as credits on later invoices.

The standard behind this sets a floor, not a ceiling. The MRC invalid traffic guidelines separate general invalid traffic, meaning bots, crawlers, and data-center traffic caught by routine filtration, from sophisticated invalid traffic, meaning hijacked devices, incentivized activity, and falsified measurement events. Accredited measurement organizations "must apply GIVT detection processes," while SIVT detection is only "strongly encouraged." The easy fraud is filtered by default. The hard fraud may not be.

So your contract needs a correction mechanism with a clock on it. Wonderkind's terms give the customer "10 business days of the relevant dashboard report" to dispute. Indeed's terms are less generous: refunds are "at the absolute discretion of Indeed and only in the form of credit for Indeed services." Both are legitimate positions. Only one of them is a number you can put in a process.

Ask for three things: a stated dispute window, a stated remedy, and confirmation of whether the remedy is cash or advertising credit. Credit is common and acceptable. Discovering it is credit after you budgeted for a refund is not.

Where does outcome-based pricing stop working?

Outcome-based pricing works where the vendor controls the outcome, and degrades quickly past that point. A vendor can control clicks, form submissions, application completions, and the objective screening gates in front of them. A vendor cannot control your interview scheduling, hiring manager response time, offer competitiveness, or pay rate. Cost per hire contracts push risk onto a party that cannot manage it, and the price reflects that, which is why genuine pay-per-hire advertising remains rare outside staffing.

Market conditions also move the achievable rate underneath you, because an outcome rate is a bet on candidate supply. UK supply loosened through the first half of 2026: ONS figures put unemployment at 4.9 percent for April to June 2026, with payrolled employees down 78,000 over the year and vacancies down to 707,000. Cheap outcomes are easy to promise into a market like that. When vacancies turn back up, that rate becomes a losing one for the vendor, and they will either reprice or quietly throttle delivery.

So set a review point rather than a fixed annual rate, and cap volume per role, or a vendor can fill a good month with marginal applicants who technically meet the definition.

How do you compare two outcome-based offers?

Rates are the last thing to compare, not the first. Work through this order:

1. Get the billable event in writing, including start versus complete.

2. Get the system of record and the attribution window.

3. Get the qualification criteria, and confirm you configure them.

4. Confirm delivery into your ATS is part of the billable definition.

5. Get the dispute window and the remedy, cash or credit.

6. Ask what happens to candidates who fail qualification, since paying twice for the same person under two definitions is a common leak.

7. Normalize both quotes to cost per completed application using your own conversion rates, then compare.

Step seven is where most comparisons flip. A 4 dollar cost per apply start and a 9 dollar cost per qualified applicant are not 5 dollars apart. Once completion and qualification rates are applied they are often within a dollar, and the second arrives pre-screened in your ATS.

FAQ

Is outcome-based pricing cheaper than fixed pricing?
Not inherently. It is more predictable per unit of result. The vendor prices their assumed risk into the rate, so outcome-based pricing usually wins on wasted spend rather than on headline cost.

What is the difference between cost per application and cost per qualified applicant?
Cost per application bills on a completed application regardless of fit. Cost per qualified applicant bills only when the candidate also meets criteria you defined and reaches your ATS. The second shifts screening effort to the vendor and is the reason the rate is higher.

Can I audit an outcome-based invoice?
Only if the contract names a system of record you can also see. Where the ad channel counts the billable event, you can reconcile against your own channel reporting. Where the vendor counts it in their own system, ask for a record-level export with timestamps and the qualification result per candidate.

Does outcome-based pricing work for hard-to-fill roles?
Yes, but expect a much higher rate or a volume cap. Vendors price scarcity, and a role with a thin candidate pool means a high cost per outcome or a refusal to quote on outcomes at all.

The short version

Outcome-based pricing for talent attraction is a measurement contract wearing a pricing label. Compare definitions and systems of record before rates, insist that delivery into your ATS is part of the billable event, and put a number on the dispute window. Wonderkind publishes its billing definitions and dispute terms on its pricing and terms pages, which is the minimum standard to hold any vendor to, including your current one.