How to Evaluate a Pay-Per-Call Network: 12 Questions Every Advertiser Should Ask Before Signing

The duration trap

Pay-per-call has priced on call duration since the industry began. Sixty or ninety seconds of conversation equals "qualified." Qualified equals billable. The advertiser pays. The network keeps its margin. The next call rings.

Duration is supposed to be a proxy for value. In practice, it's a bad one. A ninety-second conversation isn't necessarily a prospect – it can be a customer service call, an existing customer trying to reach a different department, a curious caller who never had purchase intent. Advertisers have spent two decades paying for those calls because the alternative – paying on actual outcome – wasn't technically possible at scale.

The process of determining whether a call worked has always been the same. The advertiser's sales team speaks with the caller. They figure out whether the appointment got booked, the policy got sold, the deal closed. That information eventually makes its way back to marketing. What's been broken isn't the process – it's the timing. The outcome data lived in the CRM, recorded hours or days after the call. By the time the advertiser could act on it, the campaign window was closed and the next week's bids were already set on duration.

Real-time AI dispositioning at the moment a call ends, pushed back to the network's call routing platform within seconds, makes outcome-based billing technically feasible for the first time. Combined with a feed from the advertiser's CRM – sales-team disposition data flowing back into the analytics layer alongside the AI signal – outcome becomes the basis for billing decisions rather than duration.

Most advertisers don't know this is the case. They're still negotiating duration-based terms because that's the only structure on the table. The most important question an advertiser can ask a pay-per-call network is whether they're willing to work on outcome-based terms – and whether the technology and the network itself can support it.

What follows is the checklist we'd hand a marketing leader on the advertiser side sitting down to evaluate pay-per-call networks. Twelve questions across four categories, written from the operator's seat.

Pricing & Network Scale (Q1–3)

Q1: Are you willing to price on outcome instead of duration – and do you have the technology to support it?

Why ask: This is the question most advertisers never ask, and the one that changes the most. For two decades, pay-per-call has billed on duration because real-time outcome data didn't exist. Now it does. Advertisers who don't ask the question keep paying on duration by default – even when their campaign math doesn't support it.

Outcome-based pricing isn't always the right answer. The economics have to work for both sides. A network earning a comfortable margin on duration-qualified calls may not accept the risk profile of outcome-based billing, even if the headline rate is much higher. But when there's a supply-and-demand mismatch – when duration-based campaigns have stopped producing the close rates an advertiser needs, and the relationship is about to fall apart – outcome-based terms are sometimes the only structure that keeps the program going. That option doesn't exist if the technology isn't in place.

The technical requirement is real-time disposition feedback. The call ends. AI scores and dispositions the conversation within seconds. The disposition is pushed back to the network's call routing platform immediately. Alongside that, the network is integrated with the advertiser's CRM so the sales team's actual disposition flows back into the same analytics layer for verification. That's what makes outcome billing real rather than aspirational.

Good answer: "Yes, with the right data flow. Here's how we'd structure it." A network that's done outcome-based deals will walk you through the disposition feedback loop, the timing requirements, the CRM integration. A network with the technology in place – or one willing to integrate with platforms like Aria that provide it – will have specifics on what dispositions count as outcomes, how the data flows in both directions, and how disputes get resolved.

Bad answer: "We bill on duration. That's the industry standard." It is the industry standard. That doesn't make it the right structure for your campaign.

Q2: How large is your network, and how diverse is the traffic within it?

Why ask: Network size matters in pay-per-call in ways most advertisers underestimate. A larger affiliate network gives you more diverse traffic sources, more redundancy when an individual affiliate pauses or gets removed, more competitive pressure on quality, and more data signal. The network effect is structural: the larger and more diverse the network, the better the optimization room across geographies, verticals, time windows, and audience segments – even compared to a smaller network claiming specialized vertical focus.

The instinct to pick a small specialized network over a large generalist is usually wrong. Specialization comes from operational expertise – knowing the vertical, the compliance posture, the close-rate benchmarks, the publishers that consistently work in that space. That expertise lives in the network's operating team, not in a small affiliate base. A large network with vertical operational depth gives you both: the scale of affiliate inventory plus the specialization. A small network without that depth gives you neither.

Scale also produces a compounding data advantage. A network running hundreds of millions of calls a year sees patterns smaller networks can't see – which sources convert in which geos for which verticals at which times of day. That pattern recognition becomes a quality moat, and the gap doesn't close quickly. Time in business and total volume are leading indicators of how well a network can match traffic to your business.

Good answer: Specific numbers on network size, monthly call volume, time in business, vertical coverage, and third-party rankings (mThink's Blue Book is the industry standard). Concrete examples of how the network's scale translates to better optimization in your vertical.

Bad answer: Vague claims of "premium publishers" or "specialized focus" without scale numbers. New entrants without a multi-year track record. Networks competing on size or specialization alone, rather than both.

Q3: Is this pure performance pricing, or are there technology fees on top – and what do those fees unlock?

Why ask: Traditional pay-per-call is purely performance-priced. You pay per call (on duration) or per outcome (if you've negotiated Q1 well). That structure is clean – but it also constrains what the network can deliver. If you want real-time dispositioning, CRM-integrated outcome billing, cross-platform analytics, or other technology that goes beyond standard call delivery, a network may layer a technology licensing fee on top of the per-call price.

That fee isn't inherently bad. It's the price of getting capabilities that pure performance pricing can't fund. What matters is whether the capability is worth the fee – whether the outcome billing it enables, or the analytics it provides, or the cross-network normalization it delivers, produces enough value to justify the structure.

Good answer: Transparent breakdown of what's performance-priced versus what's technology-licensed. A clear explanation of what the technology fee delivers. Modeling support for fully-loaded CPA including all fees.

Bad answer: Fees buried in the contract with no clear capability tied to them. Or, conversely, the absence of any technology layer at all – pure performance pricing means you're stuck with whatever the network's standard delivery looks like, with no path to outcome-based billing or cross-platform analytics.

Operations & Quality (Q4–6)

Q4: How quickly can we ramp up, scale down, or pause if performance isn't where we need it?

Why ask: Pay-per-call is performance-priced. There are no real three, six, or twelve-month commitments to worry about – you pay per call (or per outcome) and that's it. What matters operationally isn't lock-in; it's responsiveness. How quickly can the network ramp volume to meet a budget push? How fast can a campaign pause when something breaks? How responsive is the account team when the numbers don't look right?

The networks that earn long-term advertiser relationships compete on operational responsiveness, not on contract terms. There are no contract terms worth competing on.

Good answer: Dedicated account team. Defined response SLAs. Same-day or next-day adjustments on volume, geography, and vertical filters. Active campaign management with regular check-ins.

Bad answer: Slow ramp times. Account managers juggling dozens of advertisers with no dedicated coverage. Adjustments that take a week to propagate through the network.

Q5: How do you determine a call was qualified – and does your judgment match my sales team's?

Why ask: The definition of "qualified" hasn't fundamentally changed. A qualified call has always meant a call that produces the outcome the advertiser cares about – an appointment booked, a policy sold, a deal closed. The advertiser's sales team has always determined this, recorded it in the CRM, and eventually fed that information back to marketing.

What's changed is whether the network can know this in real time, while the campaign is still running. Some networks rely entirely on duration. Others apply AI dispositioning at the moment of the call. The serious networks build a full pay-per-call outcome intelligence layer – AI dispositioning combined with a feed from the advertiser's CRM so the network's judgment is cross-checked against the sales team's actual outcome. Disagreements between the two systems become the basis for billing disputes, model improvement, or both.

Good answer: A documented disposition taxonomy aligned to your business. AI conversation analysis at the moment of the call. CRM integration that brings your sales team's actual outcomes back into the network's analytics. The network's view of "qualified" should match yours within a known margin of error.

Bad answer: Duration-only qualification. No CRM integration. The network has its own definition of qualified that doesn't reference yours.

Q6: How is your call quality scoring validated against actual sales outcomes?

Why ask: AI dispositioning is only as accurate as its ground truth. Internal benchmarks we've seen range from around 70% accuracy on the low end to over 90% on the high end. At scale, that gap is enormous. A network billing you on 70%-accurate dispositioning is generating a 30% noise rate on your billing – paying for "qualified" calls that were actually misclassified, missing genuinely good calls the model got wrong.

The only way to know real accuracy is to validate against actual sales outcomes. That requires integration with the advertiser's CRM so the network sees the sales team's eventual disposition alongside its AI judgment. Networks that do this can quote you an accuracy number with confidence. Networks that don't are guessing.

Good answer: A published or shareable accuracy rate, validated against CRM-recorded sales outcomes, not just internal human QA. The network knows the number and can explain how it's measured.

Bad answer: "We don't measure that." Or accuracy claims based purely on internal review with no external validation loop.

Visibility & Data (Q7–9)

Q7: What visibility do I get into the network – and what stays proprietary?

Why ask: Advertisers working with a pay-per-call network get more transparency than most realize. You see every call. You hear every call recording. You see every creative an affiliate is running on your behalf. You see your conversion data, your call quality data, your disposition data. You see compliance verification – Jornaya, TrustedForm, source-page validation – confirming the lead was generated from a compliant page with proper consent.

What you don't see is the names of the individual affiliates behind the traffic. That's the proprietary piece. The network's affiliate roster – built and vetted over years, sustained through reputation, kept active through ongoing relationship management – is a meaningful part of what makes the network valuable. If advertisers could see and direct-source the affiliates, the network's value would disappear. The reason advertisers work with the network in the first place is the inverse: that the operational layer is doing the work advertisers don't want to do.

The trade is fair when the network earns it. You give up affiliate identity. You get scale (Q2), comprehensive visibility into every call and creative running on your behalf, compliance verification, fraud mitigation, workflow management, and an account team that pays attention to your program. That's the value the network creates.

Good answer: Comprehensive visibility into every call, recording, creative, disposition, and compliance check. Clear articulation of what stays proprietary and why. Concrete reputation markers – time in business, third-party rankings, references in your vertical.

Bad answer: A network that offers limited visibility on calls, recordings, or creatives. Or vague claims of "premium publishers" without operational specifics. Limited visibility almost always signals that the network can't stand behind what's actually happening on its publisher side.

Q8: How do you vet publishers, and how do payment terms protect me from quality issues during ramp-up?

Why ask: One bad publisher in a network can pollute your call volume – non-compliant traffic, repeat dialers, low-intent leads, recycled lists. The network's vetting and payment processes are your protection layer.

The strongest networks operate a thorough know-your-affiliate vetting process – understanding who the publisher actually is, where their traffic comes from, what their compliance posture looks like. That vetting is proprietary by design, but the network should be able to describe it in general terms.

Equally important: payment terms structured to protect the advertiser. A new publisher shouldn't create immediate financial exposure. Strong networks pay new publishers on a delay long enough to verify call quality – checking both the network's own quality signals and the advertiser's CRM outcomes before payment accelerates. Publishers who consistently deliver good calls get faster payment terms over time, which earns the network more volume from them. Publishers whose down-funnel performance drops get slower payment terms or get paused entirely. The financial model and the quality model reinforce each other.

Good answer: Described know-your-affiliate vetting process. Tiered payment terms that protect the advertiser during ramp. Down-funnel performance monitoring with actual outcome data feeding the decisions to scale, pay faster, or pause a publisher. Examples of action taken.

Bad answer: "We work with verified publishers" with no details on how verification happens or what changes when a publisher's quality drops.

Q9: Can you push outcome data back to my analytics or CRM?

Why ask: You need the outcome of every call flowing into your own systems so you can model lifetime value, run cohort analysis, and make portfolio decisions across your full marketing stack.

The data direction that matters most is outbound from the network to you – call records, dispositions, outcome signals, attribution data – landed in your CRM or analytics layer. The reverse direction matters too: your CRM dispositions flowing back to the network for verification and outcome billing (see Q5 and Q1).

Good answer: API access. Webhook delivery. Native integrations to Salesforce, HubSpot, vertical-specific platforms. Real-time data delivery. No additional fee for exports.

Bad answer: Data accessible only in the network's dashboard. Exports require a premium tier. Bidirectional integration not supported.

Compliance & Disputes (Q10–12)

Q10: What's your TCPA compliance posture – operationally, not just contractually?

Why ask: Telephone Consumer Protection Act exposure is one of the most material risks in pay-per-call. The legal framework is well-established. What varies enormously between networks is the operational posture – what actually happens day-to-day to keep your company from absorbing publisher violations.

Most networks rely on third-party outbound dialing affiliates. That model carries real risk: outbound dialers based in foreign markets aren't subject to US laws, and the further the dialer sits from a direct relationship with the advertiser, the less you can trust their compliance posture. You can't audit what you can't see.

The strongest networks operate their own internal call centers – based in North America, with agents under direct management – which means TCPA compliance is enforced in-house rather than through downstream contractual obligations. For traffic that does come from third-party affiliates, verification tools like Jornaya and TrustedForm capture proof that the lead was generated from a compliant page with proper consent. The best networks layer in real-time technology that automatically disconnects calls and flags traffic sources failing those checks, so non-compliant leads never reach the advertiser.

Good answer: An in-house North American call center. Third-party verification tools (Jornaya, TrustedForm) on affiliate traffic. Automated disconnect and source-flagging technology. Written contractual indemnification.

Bad answer: "Publishers are responsible for their own compliance." That phrasing means: when your company gets sued, the network isn't on the hook.

Q11: What happens when I've paid for unqualified leads?

Why ask: No matter how well a network is run, some calls won't be what they should have been. The question isn't whether disagreements on quality will happen – they will. The question is what the process looks like when they do, and how the network minimizes them in the first place.

Strong networks build avoidance into their operation upstream – quality control, publisher vetting, real-time compliance checks, accurate dispositioning. When disagreements do surface, the process should be straightforward and bilateral. The advertiser provides full rationale for the disputed call. The network reviews against its own data and, where available, the advertiser's CRM disposition. Verified misqualifications get credited or refunded within a defined timeline. The network has parallel terms with its publishers, so a confirmed bad call doesn't leave the network paying its publisher for traffic the advertiser correctly disputed.

The networks that struggle here have adversarial processes – long delays, no clear timeline, decisions made arbitrarily – or, on the other end, networks that absorb every dispute without question (which usually means the publisher contracts aren't structured properly, and the network will recover those losses elsewhere in your relationship).

CRM integration changes this category materially. When the advertiser's actual sales-team disposition is flowing back into the network's analytics, disputes get resolved by data rather than negotiation. The objective record of what the sales team actually wrote in the CRM settles disagreements before they become contentious.

Good answer: Defined dispute process in the contract. Required rationale from the advertiser. Timely review using CRM verification where available. Bilateral terms with affiliates. Communication cadence built into the account relationship to surface issues before they become formal disputes.

Bad answer: No defined process. Decisions made arbitrarily. The network refuses all disputes (suggesting they're not standing behind quality) or absorbs every dispute without verification (suggesting the publisher relationships aren't structured properly).

Q12: Can you give me outcome intelligence across every call I buy – not just the calls you deliver?

Why ask: Most network analytics show only the calls that network delivered. That's billing reconciliation in a dashboard, not analytics. What you actually need to optimize spend is pay-per-call outcome intelligence – the analytical layer above call delivery that ties every call you buy to its actual business outcome, normalized across every source, network, and call-tracking platform you use. Without that layer, you can't compare networks on level ground, you can't optimize spend across your portfolio, and outcome-based pricing under Q1 only changes one supplier relationship instead of your full program economics.

Outcome intelligence is a category, not a feature. Source-agnostic ingestion of every call you buy. AI dispositioning on each conversation. CRM validation against actual sales outcomes. One normalized portfolio view of cost-per-outcome by source. The networks that recognize this is where pay-per-call is moving are the ones positioned to define the next decade of the category.

Good answer: A platform-agnostic outcome intelligence layer – built by the network or recommended as a separate system. Aria is purpose-built for this. It sits above whichever call-tracking platform you or your networks use (Ringba, Invoca, Phonexa, CallRail, Retreaver, or network-hosted analytics) and gives advertisers one normalized outcome view across every call they buy.

Bad answer: "Our analytics show all your calls." Translation: all your calls from us. Single-network lock-in by data architecture.

How to use the checklist

Three practical recommendations:

1. Send the questions ahead of the demo. The network's response time and answer quality is itself diagnostic signal. Networks who give you thoughtful written answers before the sales call are showing you how they'll treat you as a customer.
2. Build a comparison matrix. Score each network across the four categories. Look for networks strong across the board, not brilliant in one area. A network with great pricing but no compliance program is not a good network.
3. Get the six most important answers in writing. Questions 1, 2, 5, 7, 10, and 12 – the highest-leverage ones – should be answered in writing and attached to the master services agreement.

Red flags by question