Quick answer: Pay on duration when you want a clock that keeps a call center honest and a payout affiliates can actually be held to. Pay on intent when your network can extract the outcome straight from the call itself and stop waiting on someone else's reporting. Most cost per call campaigns should run on a duration threshold as the default billable-call standard. Intent-based billing belongs in mature programs only, and only when the payout to affiliates scales along with the deeper conversion point – otherwise you are not pricing a call more precisely, you are just moving risk onto the party least able to control it.
Every pay-per-call advertiser wants to pay for what a caller actually wanted. Every affiliate selling that caller wants to be paid for what they can actually control. Those are not the same thing, and the tension between them is the real subject of this piece.
This is not a primer on running a pay-per-call program – this audience already does that daily. It is about the one decision inside that program that quietly decides who eats the cost when the two sides disagree about what a call was worth: which side of the call gets billed for what.
Why do advertisers and affiliates want to be billed differently for the same call?
An advertiser wants to pay for intent: a caller who actually wanted the thing being sold, confirmed on the call itself. That is the conversion they are trying to buy, not a proxy for it.
An affiliate wants to be paid on duration, or something close to it, and they are right to want that. Once a call is delivered and connected, the affiliate has no further control over it. What happens next – how the call center handles it, how fast the salesperson qualifies, whether the appointment actually gets booked – belongs to somebody else's team.
Holding an affiliate financially responsible for an outcome they cannot influence past the point of handoff is not a pricing model. It is cost-shifting dressed up as one.
So you have two legitimate, opposite preferences sitting on either side of the same transaction. Somebody has to stand between them. That somebody is the network, and standing there is not a footnote to the business – it is the business. The network carries the mismatch as a risk position: when the two sides value the same call differently, that gap does not disappear on its own. The network prices that risk into what it charges the advertiser and what it pays the affiliate, which is what makes carrying the mismatch a business the network is actually built to run, not a favor it does for free. That gap has to live somewhere until an industry-wide standard closes it, and until then, the network is the party pricing it, holding it, and getting paid to hold it.
Why is duration-based billing still the default in pay-per-call?
Duration has an underrated virtue that gets lost whenever this debate turns into "intent good, duration old-fashioned." Duration is a clock on the sales team.
A duration threshold forces a call center to make a fast decision. Either the caller is worth continuing to qualify, or the rep needs to disposition the call and move to the next one. Take the clock away and there is far less pressure to figure out quickly whether a lead is worth the time. Nobody is under any obligation to move fast when nothing is measuring how long they took.
That is also why duration remains the right default answer in most instances, not a legacy holdover waiting to be replaced. It is the most directly measurable outcome that an affiliate can rightly be held to. A call either ran long enough to demonstrate a real conversation or it did not. Everyone can check that against the same timestamp, and nobody has to trust anyone else's downstream reporting to settle the bill.
Duration's problems are real. It can be gamed, and it depends on tight coordination between marketing and the call center – creative that sets the wrong expectation produces calls that run long for the wrong reasons, or short for the wrong reasons, and neither side notices until payouts start looking strange. None of that erases what duration does well. It disciplines urgency in a way intent-based billing, by itself, does not.
What actually makes intent-based billing possible now?
Attribution independence is what actually makes intent-based billing possible, and it has very little to do with billing philosophy and everything to do with who is stuck waiting on whom.
If a lead generator has to rely on the advertiser's own records to know whether a call converted, that lead generator is exposed to every delay in the advertiser's reporting cycle. Clients are often slow to hand back outcome data, and a network billing off someone else's delayed records is a network that cannot close its own books on time or defend a disputed payout with confidence.
Intent-based billing becomes attractive the moment a network can extract the outcome itself, straight from the call, without waiting on the client at all. That is the actual case for intent-based billing. It is not that intent is a more virtuous way to price a phone call. It is that not having to depend on someone else's reporting is a real operational advantage, and it is worth pursuing on that basis alone.
What makes that extraction possible is real-time disposition built on AI transcription. A system that listens to the call and does fact extraction against it can determine, call by call, what actually happened – whether the conversation produced the outcome the advertiser is trying to buy – and do it fast enough to price against.
This is the role Aria plays in our own network. It is an intelligence and disposition layer: it listens, extracts, and tells you what happened on a call. It is not a router and it does not enforce a billing model on anyone. What a network or an advertiser chooses to do with that disposition – bill on it, use it to QA a duration model, or hold it as a diagnostic before changing anything – is still a business decision made on top of what Aria surfaces, not one Aria makes.
Who carries the risk when one side pays duration and the other pays intent?
This is where the network's position gets uncomfortable.
If a network pays its affiliates on duration but gets paid by the advertiser on intent, the two sides of that transaction are not pointed the same direction. The advertiser has no clock on them. There is no duration threshold forcing a fast decision, so the advertiser can take as long as it wants evaluating whether a given call actually produced the outcome it is willing to pay for. Meanwhile the affiliate has already been paid, or is owed payment, on a duration basis regardless of how that evaluation eventually lands.
That split creates a real risk, and the network is the party left holding it. Not the advertiser, who is not on a clock. Not the affiliate, who was never on the hook for anything past delivery. The network sits in the middle of a mismatch it did not create and cannot fully control, absorbing the difference every time the two pricing models disagree about the same call.
There is a genuine counterweight here, and it points the other way: an advertiser taking its time evaluating a call is not purely a downside. Slower, more careful evaluation of a lead also seems, in the accounts we manage, more likely to produce the outcome the advertiser actually wants. That is a pattern we observe, not a measured finding, but it is real enough that both things can be true at once, and neither cancels the other out.
How do you define caller intent for a specific vertical?
Intent is not a setting you flip on a dashboard. It is a scoping conversation, and it looks different for every advertiser, even within the same vertical.
Take a roofing contractor. For that business, intent typically means a homeowner has agreed to an in-home appointment with a salesperson – a specific, checkable event, not a general sense that the caller sounded interested. But that definition is not universal even across other roofing contractors; what counts as the right outcome for one contractor will not match another's definition of it, because no two contracting businesses run their qualification the same way. AI can only extract the right disposition from a call once someone has done the work of defining, in detail, what "the right outcome" looks like for that specific business. Skip that step and the AI has nothing correct to extract toward.
That scoping burden is visible in our own network data. Across our network's Pay Per Call category, over a recent three-month window, we logged roughly 18,000 post-conversion events against roughly 426,000 qualified calls. That is not a qualification rate, and it should not be read as one – it is a count of how often attribute data gets passed back to us across a mixed set of campaigns. Almost all of those 18,000 events are caller-attribute signals: things like Medicare qualification status, age bracket, residency, or account status, plus a marker that simply confirms a call was initiated. What is almost entirely absent from that count is the other half of the picture – an appointment actually booked, a policy actually issued, a deal actually closed.
That gap says something useful: the infrastructure to capture who a caller is already exists at meaningful scale. The infrastructure to capture what ultimately happened to that caller, vertical by vertical and advertiser by advertiser, still largely has to be built. That is exactly why intent-based billing cannot launch as a universal feature. It has to be scoped one relationship at a time.
Does intent-based billing solve fraud?
No. Intent-based billing is not a fraud fix, and treating it as one is a mistake.
Duration and intent get gamed by the exact same mechanism: the coached call. A coached call is nothing more than someone claiming to be a person they are not, or interested in something they are not, in order to trigger whatever the payout is keyed to – whether that trigger is a call running past a length threshold or a false appointment booked to satisfy an intent signal. It is the number one complaint in this category, and switching the billing basis does not make it go away. It just relocates where the fraud has to aim.
Automated scrubbing catches a shrinking share of it, at least based on our own scrubbing across our own network over the three months we tracked. Scrub rates ran roughly 0.31% in the earliest of those three months, down to about 0.15% the following month, and about 0.12% the month after that. That decline is not evidence that coached calls are disappearing. It is evidence that automated scrubbing is a narrow tool, built to catch mechanical fraud patterns, not a person convincingly performing a role on a live phone call. The mechanism that actually catches a coached call is closer inspection of the conversation itself, not a scrub rule running against metadata.
Moving to intent-based billing does not remove the incentive to fake an outcome. It just moves the target from "make this call last longer" to "make this call sound like it produced the right disposition." Any network considering intent-based terms needs a coached-call defense built for that reality from day one, not an assumption that better billing automatically means less fraud.
When should you actually move the conversion point deeper into the funnel?
This is the question that decides whether intent-based billing is a genuine upgrade or just a way to quietly shift risk onto whoever has the least power to refuse it.
Duration is still the right answer in most instances. It is the most directly measurable outcome affiliates can rightly be held to, and that should remain the default assumption for any new cost per call campaign.
Moving the conversion point deeper into the funnel – toward a booked appointment surfaced through conversion intelligence, rather than a duration threshold – is available only inside a mature relationship between a buyer and a network, or a buyer and an affiliate directly. And it is available only under one condition: the payout has to scale so that affiliates make more money on the program, not less, as a result of the deeper conversion point.
State that plainly, because it is the hinge the whole model turns on. Moving the billable event deeper into the funnel without moving the payout with it is not a pricing improvement. It is asking the affiliate to accept more risk – judged now on an outcome further from anything they control – for the same money, or worse. That is not intent-based billing done well. That is duration-based economics with an intent-based excuse attached to it.
Done correctly, the sequence runs the other way. The payout increases first, in proportion to the added risk and the added distance from the affiliate's control. Only then does it make sense to move the conversion point deeper. Any advertiser or network proposing the reverse order – deeper conversion point first, payout conversation later – is not proposing a better billing model. It is proposing a worse deal.
What does this mean for your next cost per call campaign?
If you are an advertiser evaluating cost per call advertising terms, the practical takeaway is not "switch to intent-based billing." It is "understand what duration is actually buying you before you try to replace it." If you want to move toward intent, expect to invest in a real scoping conversation about what intent means for your specific business, not a generic industry definition – because there is no such thing as a generic definition that will hold up.
If you are an affiliate and a network or advertiser proposes moving your payout basis deeper into the funnel, the only reasonable response is: show me the payout increase first. A mature relationship can support that conversation. An immature one usually cannot, and pretending otherwise costs the affiliate money it cannot get back.
Both sides should expect this to keep being a work in progress. Bringing more advertisers onto an intent-based model is not a solved problem for anyone running pay-per-call at scale right now, us included. Both duration and intent are viable; the right choice depends on the specific advertiser and vertical, and scaling intent-based billing responsibly takes longer than flipping a setting, because it requires the scoping, the disposition infrastructure, and the payout math all to be in place before the first call gets billed on it.
Frequently Asked Questions
What is a qualified call in pay-per-call advertising? A qualified call is a call that meets the criteria a campaign has defined for a real prospect – commonly a duration threshold, sometimes paired with caller-attribute criteria like age bracket, residency, or account status depending on the vertical. "Qualified" describes whether the call meets the campaign's stated bar, not whether it ultimately produced a sale.
What is a billable call? A billable call is any qualified call that triggers payment under the campaign's billing model. In a duration-based campaign, that means the call cleared the duration threshold. In an intent-based campaign, it means the disposition extracted from the call matched the outcome the advertiser defined as billable – an appointment set, for example, rather than simply a long conversation.
What is a duration threshold? A duration threshold is the minimum length a connected call must run before it counts as billable. There is no industry-standard number here – the right threshold is set per campaign, based on what a real qualifying conversation actually requires in that specific vertical, not an industry rule of thumb. It is the mechanism behind duration-based billing and remains the standard trigger across most pay-per-call campaigns today.
Where does billing actually happen in the pay-per-call flow? An affiliate drives a caller to a tracking number, the call connects to the advertiser or its call center, and the campaign's billing rule determines whether that specific call is billable – either because it cleared a duration threshold or because it demonstrated the caller intent the advertiser defined as a conversion. Billing sits at that last step, after the call has already happened, which is exactly why the rule governing it is worth getting right before the first call runs, not after a dispute.
Should affiliates ever agree to intent-based payouts? Only when the payout scales to reflect the added risk, and only inside a relationship mature enough to have that conversation honestly. Moving the conversion point deeper without moving the payout is not a pricing improvement – it is asking the affiliate to carry more risk for the same money, or worse. Affiliates should treat any proposal to move the conversion point deeper as a payout negotiation first, not a billing-mechanics update to accept on faith.
Does moving to intent-based billing reduce call fraud? No. Duration and intent are gamed by the same mechanism – the coached call, where someone claims to be a person or an interest they are not in order to trigger a payout. Switching the billing basis moves the target for that fraud; it does not remove the incentive behind it.
If you are weighing duration, intent, or a mature path toward moving your conversion point deeper on your next cost per call campaign, talk to our team about what a qualified call and a billable call should actually mean for your vertical – and what payout structure has to be in place before either side agrees to change either one.