Pay affiliates for leads when your own sales team closes the deal and a single referred lead carries enough value to survive a payout. A cost per lead structure pays a fixed bounty, commonly $2 to $10 for a consumer opt-in and $50 to $200 for a B2B demo request, the moment someone completes a defined action. It also removes the credit card as your fraud filter, which means you inherit verification work that pay per sale handled for you.

I get this question from business owners about twice a month, usually phrased the same way. Their affiliates say the offer converts too slowly, or the sales team wants more at-bats, so somebody floats the idea of paying per lead instead of per sale. It’s a fair question. And the answer changes completely depending on who closes your deals.

Cost per lead works beautifully in a handful of situations and destroys margin in the rest. Below is the math I use to figure out which one you’re in, the fraud you take on the day you flip the switch, and the hybrid payout that gets you the volume without handing your budget to somebody running form-fill bots out of a basement.

What a cost per lead affiliate program pays for

A cost per lead affiliate program pays a fixed dollar amount when a referred person completes a non-purchase action you define in advance. The trigger is a form submission, a demo request, a quote request, a webinar registration, a phone call over a set duration, or an email opt-in. No money changes hands from the customer. You pay anyway.

Three terms get mixed up constantly, so pin them down before you write anything into your terms. CPL pays per qualified lead. CPA, in most affiliate software, pays per acquisition, meaning per completed sale. A bounty is a flat dollar amount instead of a percentage, and you can attach one to either event. Paying a $12 flat fee per demo request is a CPL bounty. Paying $150 flat per closed deal is a CPA bounty.

The distinction that governs everything else is where the risk sits. Under pay per sale, your affiliate carries the conversion risk. They only earn if a stranger pulls out a credit card. Under CPL, you carry the conversion risk. Your affiliate earned their money at the form submit, and whether your sales team ever converts that person is your problem. That single change explains most of what follows.

When paying per lead makes more sense than paying per sale

Lead-based payouts make sense in three specific setups, and outside of them the structure works against you.

The first is when your own team closes the sale. If a prospect requests a quote and one of your people gets on the phone to price it, your affiliate can’t influence the close. They control traffic and intent, nothing past that. Paying them on an outcome they can’t move produces exactly what you’d expect: they stop promoting. Home services, insurance, mortgage, legal, staffing, and most consultative B2B fall here. My guide on how to build an affiliate program for a service business covers the attribution problems that come with a human closer.

The second is a long sales cycle. If your median time from first touch to signed contract runs 90 days or more, a pay-per-sale affiliate waits a full quarter to see a dollar. Most partners won’t wait that long on an unproven offer. A lead bounty pays them in weeks and keeps the traffic flowing while your pipeline fills.

The third is when the lead itself has measurable standalone value. If you know a new email subscriber is worth $4.10 over 12 months because you’ve measured it, you can pay $1.25 per opt-in and sleep fine. Course creators and info product sellers running evergreen webinars use this constantly, paying $3 to $8 per registrant and making it back on the back end.

Outside those three, pay per sale wins. Self-serve checkout with no salesperson involved? Pay per sale. Thin margins on physical products? Pay per sale. A brand new program where you have no conversion data at all? Pay per sale until you do, because you can’t price a bounty on a number you’ve never measured.

How to price a lead bounty without guessing

Your maximum lead bounty equals your lead-to-customer conversion rate multiplied by your gross margin per customer, multiplied by the share of that margin you’re willing to hand over. Three numbers, and you need all of them before you publish a bounty. If you don’t have a payout benchmark to start from, my breakdown of what a good affiliate commission rate looks like gives you the industry ranges to convert into flat dollars.

Run it with real figures. Say 8% of your inbound leads become customers and each customer produces $600 in gross margin. Every lead is worth $48 in expected margin. If you’re willing to pay out 30% of margin to acquire a customer, your ceiling is $14.40 per lead. Round to $14 and you have a defensible bounty. Round to $25 because a competitor pays $25 and you’re losing $10.60 on every lead your affiliates send.

Two adjustments before you publish that number. First, affiliate-sourced leads usually convert below your site average, sometimes 30% to 50% below it, because the traffic is colder. Measure your affiliate segment separately rather than borrowing your overall rate. My post on what a good affiliate program conversion rate looks like gives you the benchmarks to compare against.

Second, build in a fraud discount. Assume 5% to 15% of lead volume in an open CPL program turns out to be junk you’ll never convert, and price your bounty against the surviving volume. If you expect to reject 10%, your $14 ceiling becomes $12.60 in practice.

What counts as a qualified lead has to be written down

Every CPL dispute I’ve refereed traces back to the same failure: the program never defined “qualified” in writing, so the affiliate and the merchant each used their own definition. Write the definition into your terms before the first lead comes in.

A workable definition names the fields, the validation, and the exclusions. Something like: a qualified lead includes a deliverable email address that passes syntax and MX validation, a reachable phone number in your service area, a company name for B2B, and an answer to at least one qualifying question. It excludes duplicates within 180 days, existing customers, competitors, leads from outside your geography, and anything flagged by your bot filter.

Then set a review window. Give yourself 7 to 14 days after submission to reject a lead before the bounty locks and becomes payable. Anything longer and your affiliates lose confidence in the program. Anything shorter and your sales team hasn’t finished dialing.

Publish your rejection rate. Affiliates tolerate rejections when they can see the number and it stays stable. They abandon programs where rejections arrive as a surprise at payout time with no explanation. Send a rejection reason with every denial, even a one-word code, so your partners can fix the source instead of guessing.

Lead qualification language belongs in your terms, not in an email you send after the argument starts. Affiliate Terms Wizard is my $49 AI tool trained on more than 1,000 attorney-written agreements, and it writes the qualification, rejection, and reversal clauses in about 10 minutes instead of the hours a lawyer bills for the same pages.

The fraud exposure you take on when you pay for leads

Pay per sale has one fraud control built in that costs you nothing: a real credit card has to clear. Fraudulent volume has to survive payment processing, chargeback review, and your refund window. CPL removes all three gates. A fake lead only has to survive a form validation, and form validations are trivial to beat.

Five mechanisms account for most of the lead fraud I see. Bot form fills submit generated names and disposable email addresses at scale. Incentivized traffic pays real humans a dollar to fill out your form with no purchase intent whatsoever. Aged lead resale takes a database somebody bought in 2021 and drips it through an affiliate link. Co-registration bundles your offer into a checkbox on an unrelated signup page, so the person never knowingly requested contact from you. And duplicate stuffing resubmits the same person with small variations to collect multiple bounties.

The defenses are specific and mostly cheap. Run email addresses through a validation service such as NeverBounce or ZeroBounce before the bounty locks. Validate phone numbers with a lookup API such as Twilio Lookup to catch invented numbers and disconnected lines. Put reCAPTCHA v3 or Cloudflare Turnstile on the form. Deduplicate against your CRM on email, phone, and normalized address. Then watch the two numbers that give fraud away: click-to-lead conversion rate and lead-to-sale conversion rate by affiliate. A partner converting 40% of clicks into leads while closing zero of them is not sending you customers.

Lead fraud and click fraud share most of the same detection stack. Affiliate fraud prevention tools: how to stop paying for fake traffic walks through the specific platforms that catch this, and how to catch and prevent affiliate fraud covers the manual review habits that catch what software misses.

The hybrid structure: a small bounty plus a back-end percentage

The structure I recommend to most business owners pays a small bounty at the lead and a full commission at the sale. It keeps affiliates paid quickly enough to stay motivated and keeps the bulk of your payout tied to revenue you collected.

The ratio that works is roughly 10% to 25% of total expected payout at the lead, with the rest at the close. Two examples of how that looks in practice.

B2B software with a $9,000 annual contract: pay $25 per qualified demo request, then 12% of first-year contract value at close, which is $1,080. Your affiliate collects $25 in week one and $1,105 total on a deal that closes. Your fraud exposure per bad lead caps at $25.

Course creator running an evergreen webinar: pay $3 per registrant, then 35% commission on a $497 course, which is $174. Total on a converting registrant is $177. If your registrant-to-buyer rate is 4%, you’re spending $75 in bounties per 25 registrants and earning one sale, so the bounty has to stay small enough that the math still works. It does at $3 per registrant. It stops working at $9.

The hybrid also solves the motivation problem the pure models each create. Pure CPL gives affiliates no reason to care about lead quality, because they got paid either way. Pure CPA on a slow funnel gives them no reason to keep promoting through a 90-day gap. Splitting the payout across both events puts a piece of their income on each side.

Cap the bounty side. Set a monthly ceiling on lead payouts per affiliate, at least until a partner has a proven close rate. Something like $500 in bounties per month for the first 90 days limits the damage from a bad partner while letting a good one graduate into an uncapped arrangement. The same logic applies to structuring commission tiers and partner types across the rest of your program.

How to track a two-event payout in your software

Most affiliate platforms handle one conversion event per referral cleanly and get confused by two. If you’re paying a bounty at the lead and a commission at the sale, you need software that can fire two separate conversions against the same click, hold the first in a pending state through your review window, and reverse it independently if the lead gets rejected.

Check four things before you commit. Can it record multiple conversion types on one referral? Can it hold a payout pending for a defined number of days? Can it reverse a lead bounty without reversing the sale commission, and the other way around? Can it report close rate by affiliate rather than only click and conversion counts? A platform missing the last one leaves you unable to spot the partner sending volume that never converts.

Attribution windows matter more here than in a pay-per-sale program, because two events happen at different times against one original click. If your lead fires on day one and the sale closes on day 74, a 30-day cookie means the sale commission goes to nobody. Match the window to your measured sales cycle. My explainer on how affiliate tracking works covers the mechanics of stacking events on a single referral.

I built AffiliateHQ because the platforms I used for 20 years fell apart on exactly this kind of structure. Full disclosure, I own it. It handles multiple conversion events per referral, independent holds and reversals on each one, and close-rate reporting by partner, which is the report you need most when you’re paying for leads.

The compliance exposure in lead-based affiliate programs

I’m not an attorney and none of this is legal advice. Talk to one before you launch a lead-gen program, especially in insurance, lending, health, or home services, where regulators pay the most attention.

The exposure that catches business owners off guard is consent. If you or your sales team call or text a lead an affiliate sent, the Telephone Consumer Protection Act governs whether that contact was authorized, and statutory damages run $500 per violation and up to $1,500 per willful violation under 47 U.S.C. section 227(b)(3). Those numbers multiply fast across a list. Your affiliate collected consent on a page you may have never seen, and enforcement lands on the company doing the calling.

Two practices reduce that risk. Require affiliates to capture and store the consent record for every lead, including the page URL, timestamp, IP address, and the exact disclosure language shown, and require them to produce it on request. Then require pre-approval of any landing page or form that collects leads on your behalf, and check the disclosure language yourself before it goes live.

Put both requirements in your terms with a right to audit and a right to withhold payment on leads where the consent record is missing. Writing an affiliate program agreement covers the structure those clauses sit inside.

Payout structure is one chapter of a much bigger system. The Book on Affiliate Management is my 300+ page walkthrough of the framework I used to build a $1 million per month affiliate program in under two years, including how I set commission structures that partners accept and I can afford.

How to test cost per lead without breaking your program

Run it as a 90-day pilot with a small group rather than rewriting your terms for everybody at once. Pick 5 to 10 affiliates who already send traffic that converts, since you have baseline data on them and they have a reason to protect the relationship.

Set the pilot up in five steps. Measure your current lead-to-customer rate and gross margin per customer so you have a defensible ceiling. Write the qualified-lead definition and the review window into a pilot addendum. Set a monthly bounty cap per affiliate. Turn on email and phone validation before the first lead arrives, not after the first bad batch. And track close rate by affiliate weekly, because that number tells you within three weeks whether a partner is sending buyers or filling out forms.

Kill criteria matter as much as success criteria. Decide in advance what ends the pilot: a close rate below half your site average, a rejection rate above 20%, or a cost per acquired customer that runs higher than your paid channels. Write those thresholds down before you start, because in week seven you’ll want to give a bad partner one more chance and the number on paper will stop you.

The comparison that decides whether CPL stays is cost per acquired customer, not cost per lead. Total bounties paid plus total commissions paid, divided by customers closed from affiliate leads. Compare that against your pay-per-sale affiliate segment and against your paid ads. If lead-based costs more per customer than the alternatives, go back to paying on sales. My breakdown of payout schedules and methods and the deeper math in structuring affiliate commissions for a SaaS product both help you run that comparison honestly.

Three things to do next. Calculate your lead-to-customer rate and gross margin per customer this week, because you can’t price a bounty without them. Write the qualified-lead definition and review window into your terms before you accept a single lead. And if you run the pilot, use the hybrid structure with a capped bounty rather than pure CPL, so a partner sending garbage costs you $25 instead of $2,500.

If you want a second set of eyes on the numbers before you commit, grab a free 20-minute coaching call. We’ll look at your current program, your close rate, and whether a lead bounty helps or hurts, and you’ll leave with an action plan for the next 30 to 60 days.