Pay-Per-Call Marketing in 2026: A Buyer-Side Operator’s Guide to Floors, Waterfalls, and Agency Selection

Questions this article answers:

• What is pay-per-call marketing, and how is it different from PPC or lead generation?
• How do I set a duration floor against my buyer’s billable threshold?
• What is a buyer waterfall, and how do I build one?
• How has the FCC one-to-one consent rule changed pay-per-call sourcing in 2026?
• How do I tell a real pay-per-call agency from a network reseller?
• When does pay-per-call not make sense as a channel?

Most pay-per-call content online is written by networks selling inventory. That is why every guide reads the same: a definition, a list of benefits, a pitch. None of it teaches you how to actually configure a campaign.

This one is written from the buyer side. If you spend real money on paid media and you are considering pay-per-call, or you are already running it and your payout numbers never quite reconcile, the gap is almost always in three places: your duration floor, your buyer waterfall, and how your call qualification step writes back to the bid layer. Get those right and the channel works. Get them wrong and you spend the quarter arguing about shaved minutes that were never shaved.

We will walk through the mechanics, the compliance state of play in 2026, and an eight-question diagnostic for telling a real agency from a reseller in a nicer hoodie.

What Is Pay-Per-Call Marketing, and How Is It Different From PPC or Lead Generation?

Pay-per-call marketing is a performance advertising model where the advertiser pays for a qualified inbound phone call instead of a click, an impression, or a form-fill. “Qualified” is the load-bearing word. A call only counts when it crosses two thresholds: a minimum duration (the billable threshold) and a disposition that matches what the buyer is paying for.

That is the part most guides describe. What they skip is who is on each side of the trade.

Buyer-side vs. publisher-side: who owns which risk

Pay-per-call has two sides, and the economics are different on each.

Buyer-side is the advertiser, or an agency representing the advertiser, paying a network or a direct source for qualified inbound calls routed to their sales floor. The buyer’s risk is paying for calls that look qualified on paper but do not close. Their job is to set billable thresholds, dispositions, and caps that protect their cost-per-sale.

Publisher-side is the media buyer, or an agency, sourcing traffic (usually through Google Search, LSA, Meta, or native) and getting paid by a buyer or a network when calls clear the threshold. The publisher’s risk is spending real ad dollars on calls that get scrubbed, disputed, or never cross the floor.

The agency-as-buyer-rep role is invisible on the search results because every page that ranks for this query is written by an inventory seller. A real pay-per-call agency sits on the buyer side. They own the call qualification flow, they configure the waterfall, they negotiate buyer terms, and they reconcile the settlement reports. They do not just hand you a tracking number and a payout report.

Where pay-per-call fits next to PPC, LSA, and form-fill lead gen

Pay-per-call works when intent decays fast. A homeowner with a flooded basement is not filling out a form and waiting for a callback. A Medicare-eligible caller who just heard a Q4 ad wants to talk now. Form-fill lead gen tolerates a longer call-back window. Live-transfer calls do not.

This is why duration floors exist at all. Click data tells you someone arrived. Call duration tells you they engaged. The longer the call, the more likely the intent was real.

How Do I Set a Duration Floor Against My Buyer’s Billable Threshold?

Your duration floor should equal your buyer’s billable threshold minus your average call-qualification traversal time minus a 10-second buffer. That single configuration change closes the gap between what your call tracking platform says you delivered and what your buyer says they will pay for.

Most operators inherit a default floor from their call tracking platform and never touch it. That default is a vendor convenience, not a unit-economics decision. It was not set against your buyer’s contract because your buyer’s contract did not exist when the default was written.

The reconciliation formula

Publisher duration floor = Buyer billable threshold − Average call-qualification traversal time − 10-second buffer

Three pieces:

• Buyer billable threshold. Whatever they pay against. Get it in writing.
• Average call-qualification traversal time. How long the caller spends inside your qualification flow before reaching the buyer’s billable connection point. This is the IVR, the live fronter, or the AI voice agent, depending on your setup. Measure it from your own call logs. Do not guess. This varies with the number of qualification questions you ask, which is why measuring it per campaign matters more than importing a benchmark.
• 10-second buffer. Network latency, human hesitation, and the fact that timestamps do not agree perfectly across systems.

Worked example: a 120-second buyer threshold

Say your Medicare buyer pays on 120 seconds of connected-to-agent time. Your own logs show qualification runs about 25 seconds before the transfer. Your publisher-side floor should be 120 minus 25 minus 10, which is 85 seconds.

Set it at 85. Now every call you count as delivered will, with very few exceptions, also clear the buyer’s threshold. Your delivered count and your billable count converge. Your settlement reconciles within a few points instead of fighting about it monthly.

Leave it at a generic platform default and you will deliver a meaningful share of calls that die between that default and the buyer’s actual threshold. The buyer marks them unbillable. You spend the rest of the quarter wondering whether the network is shaving you. They are not. Your floor was set against the wrong number.

Duration floors vary by vertical

Intent decays at different speeds in different verticals, and qualification flows take different amounts of time. The right floor for your campaign is whatever the formula above produces against your actual buyer contract and your actual measured qualification time. Do not import a benchmark from a network deck. Pull the numbers from your own logs and your buyer’s term sheet.

What is consistent across verticals as a directional rule of thumb: higher-trust sales (final expense, mass tort intake) take longer qualification, so both the buyer threshold and your floor sit higher. Faster transactional intent (HVAC emergency dispatch) clears faster on both sides.

This single decision fixes more pay-per-call P&Ls than any creative test. We have written more on the IVR disposition side of this loop and on how publisher buffer time creates incentives upstream if you want to keep going.

What Is a Buyer Waterfall, and How Do I Build One?

A buyer waterfall is the routing logic that decides which buyer gets a given call first, second, and third. It is the routing brain of a pay-per-call campaign, and it is missing from every guide that ranks for this term.

Four real variables drive it:

1. Priority. The strategic ranking you give a buyer. Sometimes your highest payer is not your highest priority because they are slow to pay, they dispute aggressively, or they are testing the relationship. Priority is your call, not the math’s call.
2. Payout. Gross dollars per billable call. Easy to measure. Easy to over-weight.
3. Cap remaining. How much of the buyer’s daily or hourly cap is still available. A buyer who has burned through most of today’s cap should be treated very differently from one that has barely started.
4. Concurrency. How many of the buyer’s agent seats are live right now. Routing a call to a buyer whose floor is full means a hold time, a hang-up, and a wasted call.

How to weight the variables by traffic intent

The weighting changes with your traffic mix.

High-intent traffic (branded search, LSA, direct-response Search) should weight payout and priority. The call is precious. Send it to the buyer who pays the most and treats you well. You can afford to be picky.

Broad-intent traffic (Meta, native, broad Search) should weight concurrency and cap remaining. Contact rate is the constraint. The fastest pick-up wins. A long hold time on a Meta-sourced caller is a dead call.

This is why a campaign that performs on Search collapses on Meta even with an identical qualification flow. The waterfall logic that protects margin on high-intent traffic strangles contact rate on broad-intent traffic.

Ping-tree vs. direct-route routing

Ping-tree routing asks multiple buyers in parallel whether they will take a call and at what price. Best for live-transfer models where buyers actively bid on the call. Direct-route sends to a single buyer based on your waterfall rules. Cheaper, faster, and better for most inbound pay-per-call campaigns. Use ping-tree only when your buyers genuinely compete on price for the same call type.

Why daily caps strangle yield when set wrong

Daily caps are a buyer-side cost control, but they create a routing problem you have to solve. A buyer who hits their cap mid-morning dies for the rest of the day. If your waterfall is not watching cap utilization in real time, your second-priority buyer does not pick up the slack until you notice manually. By then you have lost the afternoon.

The fix is pacing-aware routing, not a static daily cap. The agent watches the buyer’s spend against their stated cap and the hours remaining in the day. Calculate cap-utilization pacing as: calls delivered into the cap so far, divided by the buyer’s stated daily cap, compared against the share of the buying day elapsed. If the buyer is pacing to exhaust the cap before close of business, demote them in the waterfall now instead of waiting for the hard stop. Same logic for the inverse. A buyer pacing slow should get promoted before you leave volume on the table. (Concurrency belongs as a separate real-time check on whether seats are open right now, not inside the pacing math.)

If you are evaluating routing platforms, we wrote a practical comparison of the major call routing tools that covers the actual feature differences. There is also a Ringba vs TrackDrive breakdown for buyer-side operators specifically.

How Calls Should Write Back Into the Bid Layer

Most operators treat the call qualification step as a gate. Sophisticated buyer-side operators treat it as a feedback channel that writes back into Google Ads, Meta, and LSA.

A working disposition taxonomy looks something like this:

• Billable-connected (transferred, agent connected, duration cleared)
• Qualified-not-billable (qualified caller, transfer failed or dropped before threshold)
• Wrong-vertical (intent mismatch)
• Hang-up pre-threshold (caller dropped during qualification)
• Transfer declined (buyer rejected during whisper)
• DNQ (failed eligibility, wrong state, wrong age, etc.)

Each disposition gets pushed back to the bid layer via offline conversion imports. Google Ads accepts these through the API or Data Manager (per the Google Ads offline conversions documentation), Meta accepts them through the Conversions API, and LSA accepts them through dispute and feedback flags. The point is that the bid platform learns from the disposition, not from raw call counts. Google’s call conversion measurement guidance describes feeding call outcomes into bidding rather than raw call volume.

This is why call qualification rate (qualified calls divided by total delivered calls) is your leading indicator of traffic-source health. CPL moves last. Qualification rate moves first. When your qualification rate on Meta-sourced calls drops two points week over week, your creative is fatiguing or your audience is contaminating. CPL will not catch up to that signal for another week.

The metric you actually manage to is cost-per-billable-call, not CPL. Cost-per-billable-call is total campaign cost divided by calls that actually crossed your floor and the buyer’s threshold. That is the number that ties to revenue.

How Has the FCC One-to-One Consent Rule Changed Pay-Per-Call Sourcing in 2026?

The FCC one-to-one consent rule was vacated by the 11th Circuit in January 2025, but most buyers in regulated verticals still price one-to-one risk into their contracts. Sourcing calls in 2026 looks a lot like sourcing them under the rule, even though the rule is not in force.

Quick recap. The FCC’s 2023 order required prior express written consent for each individual seller a consumer would receive marketing calls from, replacing the older shared-list practice. The rule was set to take effect in early 2025. In January 2025 the 11th Circuit vacated it (per the Wiley law-firm analysis). The FCC withdrew enforcement.

What did not change: lead-buyer general counsels. Most reputable Medicare, ACA, final expense, and legal buyers continued to require one-to-one-style consent in their buying contracts through 2025 and into 2026, because the underlying TCPA exposure on misrouted marketing calls did not go anywhere. Their compliance team would rather lose volume than absorb a class action.

What this means for sourcing in 2026:

• If you are publisher-side, assume buyers will continue to require named-seller consent on the lead-capture page that produced the click that produced the call. Generic “partners may contact you” language will get your calls rejected by the better buyers.
• If you are buyer-side, your vendor agreements should still spell out consent capture, recording retention, and replay rights, even though the federal rule is not enforced. We wrote about the practical setup we recommend in more detail.
• Two-party-consent states still require recording disclosure. The greeting that handles this needs to be in every campaign.
• The SMS-to-call funnel (text first, call second) carries its own TCPA exposure on the transferred call. Do not assume consent at the text step covers the call step.

The practical takeaway: source as if the rule is in force. The cost is small. The downside of doing it the other way is large.

How Do I Tell a Real Pay-Per-Call Agency From a Network Reseller?

Eight questions, asked in a pitch meeting, surface this faster than any reference check. The pattern is consistent. A real agency owns the mechanics. A reseller owns a relationship with someone who owns the mechanics.

1. Can you show me a buyer waterfall you have configured? Real answer: they pull up a screenshot or walk through one on screen, naming the four variables and how they weighted them for a specific buyer mix. Reseller answer: “Our network handles the routing.”

2. Do you own the qualification flow, or are you using a network-provided one? Real answer: “We own it. Here is our disposition taxonomy.” Reseller answer: “We use the network’s standard setup.”

3. Will you share disposition-level data, not just settlement totals? Real answer: yes, weekly or on demand, with the disposition codes mapped to your buyer’s terms. Reseller answer: “We share the network’s report.”

4. How do you set duration floors per buyer? Real answer: some version of the reconciliation formula above, with logic for why each buyer’s floor is where it is. Reseller answer: a single global number inherited from the platform default.

5. Do you price by traffic source or by call type? Real answer: by call type, with traffic source as a quality lens behind the scenes. Reseller answer: a flat rate that does not change with what you are buying.

6. Walk me through your last settlement dispute. Real answer: a specific story with a specific resolution and what they changed afterward. Reseller answer: vague, or “we do not really get disputes.”

7. What is your concurrency cap logic? Real answer: real-time agent availability factored into routing, with a fallback when seats fill. Reseller answer: blank stare or “the network handles it.”

8. Who owns the call recordings? Real answer: you do, with a defined retention policy. Reseller answer: “The network stores them.”

If five or more answers come back in the reseller column, you are not hiring an agency. You are paying agency fees for network-broker access.

Reading a settlement report: the four line items that matter

When the report lands, audit these four numbers monthly:

• Scrub rate. Calls rejected for quality reasons by the buyer. Trending up over weeks signals a traffic-source problem. A sudden spike signals a buyer policy change you were not told about.
• Dispute rate. Calls credited back after settlement. A modest baseline is normal. A sustained run higher than that on a single buyer signals their floor is moving without your knowledge.
• Shaved-minute flags. Calls where your delivered duration and the buyer’s reported duration differ materially. A few seconds is timing drift. Twenty-plus seconds, consistently, is the conversation to have.
• Disposition-rejection drift. Dispositions that used to qualify and now do not. Often the first sign that a buyer is tightening criteria under cap pressure.

If you cannot audit these because your vendor does not share them at this granularity, that is itself the answer to question 3 above.

When Pay-Per-Call Does Not Make Sense

Pay-per-call works when four conditions hold:

• Sales close on the phone, not async.
• Customer LTV supports a meaningful cost-per-billable-call.
• You have agent capacity or routing flexibility to honor contact rate.
• The vertical has live intent decay that rewards inbound speed.

It does not work when:

• AOV is low and the call cost cannot be absorbed.
• The sales motion is async (email, scheduled demo, multi-touch enterprise).
• You do not have a phone team or a routing partner.
• The buyer-side billable threshold in your vertical has compressed below what publisher economics can deliver. We see this in some over-saturated verticals where buyers have pushed thresholds long enough that publisher acquisition costs no longer pencil.

A healthy pay-per-call P&L looks the same at small spend and at large spend. Qualification rate inside the band your traffic source supports. Cost-per-billable-call inside your contribution margin. Contribution margin left over after operating cost. At higher spend you have more leverage with buyers, which means better caps, better feedback loops, and tighter floors. The fundamentals do not change with scale. The negotiating power does.

Frequently Asked Questions

What is pay-per-call marketing, and how is it different from PPC or lead generation?

Pay-per-call marketing is a performance advertising model where the advertiser pays for a qualified inbound phone call rather than a click, an impression, or a form-fill. Qualification is defined by call duration (a billable threshold) and disposition (the call reason and outcome matching what the buyer pays for). It differs from PPC because the trigger is a phone conversation, not a click, and it differs from form-fill lead gen because the intent decays in minutes, not hours.

How do I set a duration floor against my buyer’s billable threshold?

Take the buyer’s billable threshold, subtract your average call-qualification traversal time, then subtract a 10-second buffer for network latency. If the buyer pays at 120 seconds and your qualification runs 25 seconds, your floor should be 85 seconds. This closes the gap between your delivered call count and the buyer’s billable count, which is the gap that gets misdiagnosed as shaved minutes every quarter.

What is a buyer waterfall, and how do I build one?

A buyer waterfall is the routing logic that decides which buyer receives a given call first, ranked by four variables: priority, payout, cap remaining, and concurrency. Build it by ranking buyers on those four variables and weighting them differently for high-intent versus broad-intent traffic. High-intent traffic should weight payout and priority. Broad-intent traffic should weight concurrency to protect contact rate.

How has the FCC one-to-one consent rule changed pay-per-call sourcing in 2026?

The FCC’s one-to-one consent rule was vacated by the 11th Circuit in January 2025, but most reputable buyers in regulated verticals still price that risk into their contracts. Sourcing calls in 2026 looks a lot like sourcing them under the rule, because lead-buyer general counsels are slower to relax than regulators. Generic consent language will get your calls rejected by the better buyers regardless of federal enforcement.

How do I tell a real pay-per-call agency from a network reseller?

A real agency owns the qualification flow, the buyer waterfall, the disposition taxonomy, and the duration-floor logic per buyer. A reseller owns a relationship with a network that owns those things. Ask to see a configured waterfall, a disposition taxonomy mapped to specific buyers, and a real settlement dispute they resolved. Vague answers on three or more of these means you are paying agency fees for broker access.

When does pay-per-call not make sense as a channel?

Pay-per-call does not work when your sales motion is async, your AOV cannot absorb a meaningful cost-per-billable-call, you do not have agent capacity, or the vertical’s buyer thresholds have compressed below what publisher economics can deliver. It works best when intent decays fast, LTV is meaningful, and a phone conversation is how the sale actually happens. If three of those four are missing, run a form-fill or click-based motion instead.

We are media buyers and pay-per-call operators sharing what we see in the field. This is not legal advice. TCPA and state consent rules are genuinely complicated and vary by state and vertical, so talk to an actual attorney before changing your consent flows or vendor contracts.

If you want a second set of eyes on your current pay-per-call setup, whether that is auditing a network settlement, redesigning a buyer waterfall, or running the eight questions above against an agency pitch sitting on your desk, book a free consultation and we will walk through it with you.