- The biggest driver of dispute volume and CPA blowouts in pay-per-call isn’t payout or routing. It’s the billable-call duration threshold, which most buyers leave at the 60- or 90-second platform default.
- The right buffer sits at the disqualifier gate: the moment the agent confirms the one thing they can’t fix. That’s the health-question block on final expense, the SOA recording on Medicare, and the service-area-plus-homeowner check on HVAC.
- Set the buffer below the gate and you bill for calls that were already dead. Set it above the gate and you train publishers to stall-pad until the meter runs out.
- Pay per call buffer time settings by vertical should be negotiated explicitly in publisher agreements, not inherited from a Ringba, Invoca, or Retreaver UI default.
- Validate the setting by reading rejected-call clustering. If rejections bunch tightly just past your buffer, the gate is inside your billable window and the number is wrong.
The 60-Second Default Exists Because of a UI Pre-Fill, Not Because of Economics
Most buyers running real pay-per-call spend inherited their billable threshold from a dropdown. Sixty seconds on one platform, ninety on another, the same number across every campaign in the account. It looks like an industry standard. It isn’t. It’s a default field nobody changed.
That single laziness drives more rejected-call disputes, publisher churn, and cost-per-acquisition blowouts than any payout negotiation you’ll have this year. The buffer (the minimum call duration that triggers billing) is the one setting that tells your publishers what kind of call you’ll pay for. Set it wrong and you’re either paying for calls that already failed or paying publishers to stretch calls that should have ended.
This piece maps pay per call buffer time settings by vertical to the actual structure of each sales call. Final expense, Medicare, ACA, HVAC, plumbing, roofing, solar. The framework is the same in every vertical. The right number is not.
The Right Buffer Sits at the Disqualifier Gate, Not the Average Call Duration
The disqualifier gate is the first moment in the call where the agent confirms the one thing they cannot fix. Health status on a final expense call. Eligibility on Medicare. Service area and homeowner status on HVAC. If that confirmation fails, the call is dead no matter how skilled the agent is. Your buffer should sit just past that moment.
Most competing guides tell you to set buffer to average call duration. That’s wrong on its face. Average duration mixes qualified calls, disqualified calls, and stall-padded calls into one number. Optimizing to it bakes the failure modes into your floor.
Two failure modes, one bad setting:
- Buffer below the gate. You’re billing for calls that were going to disqualify anyway. Dispute rate climbs. You burn time on credits and clawbacks. Publishers stop trusting your dispute window.
- Buffer above the gate. You’re paying publishers to slow-walk the script. They will. Their margin improves with every extra second of hold music, and they will rationally optimize for it.
How to find the disqualifier gate when you don’t already know it
Pull thirty recent calls that closed and thirty that didn’t. Listen for the moment in the closed calls where the agent confirms the one unfixable thing. Then check whether that timestamp falls before or after the disqualified calls were lost. The clean separation point is your gate. It’s almost never the network default.
IVR length shifts this linearly. Every fifteen seconds of pre-agent menu pushes the gate fifteen seconds later. If your traffic source runs a long IVR, your buffer climbs with it. The gate is measured from call connect, not from agent connect.
Insurance Verticals Cluster Tight: Final Expense Short, Medicare Long, ACA in the Middle
Insurance gates are underwriting questions. They resolve fast because they’re yes-or-no. After auditing call recordings across these verticals, the gates fall in fairly predictable bands.
Final expense. Roughly 50 to 60 seconds. The health-question block is the whole call until it clears. Tobacco, major conditions, recent hospitalizations. If the prospect fails the knockout questions, the agent already knows the policy can’t be written. Setting buffer at 90 seconds on final expense is how you end up paying publishers to keep dead prospects on the phone for thirty extra seconds. (We dug into this dynamic in our final expense CPL writeup.)
Medicare. Roughly 110 to 130 seconds, and longer during AEP. Medicare’s gate is heavier because the Scope of Appointment requirement under CMS marketing rules sits inside the qualifying flow. SOA confirmation plus effective-date and zip-eligibility checks take the better part of two minutes before the agent knows whether the call is workable. Buyers who default to 90 seconds on Medicare are billing on calls that haven’t even finished the eligibility step.
ACA. Roughly 90 to 110 seconds. The subsidy and household-income block is the binder. Family size, projected income, current coverage status. Until those three are confirmed the agent can’t quote. More on the 2026 mix in our ACA open enrollment piece.
Under-65 major medical. 75 to 90 seconds. Pre-existing condition plus state availability. State availability alone has killed more campaigns than any single underwriting factor.
Term life. 60 to 75 seconds. Age band, tobacco, target coverage. Closer to final expense than to Medicare on the buffer math.
The final expense dispute pattern that proves the default is wrong
Audit a final expense pay-per-call book running on a 60-second default and the rejected-call distribution tells a consistent story. A sharp cluster of rejections lands in the 60-to-80-second window. That cluster is the gate. The calls are billing right before the agent finishes the health questions, then getting disputed when the prospect fails. Move the buffer to 65 or 70 seconds and the disputes collapse, because now you’re only billing on calls that survived the gate.
Why Medicare during AEP is the exception
AEP volume is so concentrated that some buyers run a slightly tighter buffer (closer to 100 seconds) to clear inventory. That’s a defensible trade only if your publisher is sophisticated enough to honor an explicit dispute window past the buffer. If they aren’t, you’ll eat the dispute volume in January. We’ve covered the AEP cycle separately in our Medicare Advantage AEP playbook.
Home Services Run Longer Because the Gate Is Operational, Not Underwriting
Home services gates aren’t yes-or-no questions. They’re operational confirmations: service area, same-day availability, homeowner status, project scope. They take longer because they’re multi-part. Treating home services with insurance buffers is the most common mistake we see when a buyer expands from one to the other.
HVAC. Roughly 90 to 150 seconds. Service area plus same-day availability plus homeowner. The wide range is real, and it’s driven by dispatch. In-house dispatch clears the gate faster. Outsourced dispatch adds a transfer step that pushes the gate later. If you’re running both, they need different buffers, full stop. For the broader cost framework here, see our HVAC lead generation benchmarks.
Plumbing. Split the vertical. Emergency plumbing clears at 45 to 60 seconds because the only meaningful gate is are you in our service area right now. The prospect already wants the truck. Scheduled service runs 90 seconds or more because the gate includes scope, scheduling, and pricing range. Running both on the same campaign with the same buffer is how plumbing buyers end up underpaying for emergency leads and overpaying for scheduled ones. Our plumbing LSA versus Google Search piece covers the routing implications.
Roofing. Split storm versus retail. Storm roofing has an insurance-claim-status gate at 75 to 120 seconds (carrier, claim filed, adjuster scheduled). Retail roofing is closer to solar’s profile, 90 to 120 seconds for homeowner, project type, and timeline.
Solar. 90 to 120 seconds. Homeowner status, credit indicator, roof condition. All three need to clear before the appointment-setter can book a site survey. Background on the post-ITC economics is in our solar CPL writeup.
Pest control. 60 to 75 seconds. Service area plus property type plus pest type. Closer to plumbing emergency than to HVAC. See the pest control LSA breakdown for booked-job math.
The emergency-vs-scheduled split inside plumbing
Most plumbing buyers we audit run one campaign for both call types. Their average call duration looks fine in aggregate. Underneath it, emergency calls are billing at 90 seconds when they should bill at 50, and scheduled calls are billing at 90 when they should bill at 110. The net looks reasonable. The per-segment economics are wrong in both directions.
The fix is two campaigns, two buffers, and ideally two routing trees. If your network’s UI doesn’t let you split cleanly, you’re using the wrong network.
Set the Buffer Above the Gate and Your Publishers Will Slow-Walk the Script
Publishers are rational economic actors. If your buffer is 120 seconds and their gate clears at 90, every call between 90 and 120 seconds is unpaid work for them. Their margin improves by stretching it. They will stretch it.
Stall-padding has a sound. Extended hold music after the prospect connects. Agent-prompted redundant questions (“let me confirm your zip code one more time”). Deliberate slow pacing. Loop-backs through the script. None of this is the publisher being lazy. It’s the publisher responding to the incentive you set.
The cost isn’t just inflated billing. Prospect intent decays the longer they’re on hold. A caller who came in hot at second zero is lukewarm at second 90 and irritated at second 130. The lead the agent finally gets is a worse lead than the one the publisher received. You’re not just paying more per call. You’re paying more for a call that converts at a lower rate.
The call-recording signatures of a stall-padded call
The patterns are consistent on playback. Two or more rounds of hold music after the initial connect. The agent asking for information the IVR already captured. A pause in the script right before the buffer triggers. “I want to make sure I have this right” repeated more than once. If you hear these on calls that bill at exactly your buffer plus three to five seconds, your buffer is above the gate.
Negotiating the Buffer in Publisher Agreements Without Breaking the Relationship
The operational fix is straightforward once you’ve mapped your gate. The relationship fix is harder. Publishers built their margin models on your current buffer. Changing it is a real economic move on their side, not just a config tweak.
Here’s the sequence we recommend.
The A/B test that doesn’t blow up the publisher relationship
Start on one source or one publisher. Hold payout constant. Run the new buffer for two to three weeks. Measure dispute rate, qualification rate, and booked-appointment rate (or whatever your downstream conversion signal is). If the numbers move in your direction, expand. Telling a publisher you’re testing is fine. Changing the buffer without warning across the whole account is how you lose your top three sources in a week.
When to raise payout instead of raising the buffer
If the gate genuinely sits late in your vertical, you may need to compensate. Medicare during AEP is the cleanest example. The buffer needs to sit past 120 seconds because the gate is there, but your publishers were modeling on 90. Raising payout by 10 to 15 percent alongside the buffer change is often cheaper than absorbing the publisher churn you’d get from raising the buffer alone.
Reading rejected-call clustering to validate your setting
Pull a 30-day rejected-call report. Bucket rejections by call duration in five-second bins. If you see a clear spike in any bin between your buffer and your buffer-plus-30-seconds, your gate is inside the billable window. Move the buffer to the right of the spike. If rejections are evenly distributed across all durations, the issue is traffic quality, not buffer setting. That’s a different conversation with the publisher.
The formulas worth watching as you make these changes:
| Metric | Formula | Whose View |
|---|---|---|
| Cost per qualified call | total campaign cost ÷ qualified calls | Buyer |
| Effective payout per converting call | payout × (1 − post-bill rejection rate) | Publisher |
| Buffer-adjusted CPA | (payout × calls billed) ÷ closed sales | Buyer |
The publisher’s view and yours should both improve when the buffer sits at the gate. If one improves and the other gets worse, you’ve shifted the problem rather than solved it.
Language to put in publisher agreements
The clauses we push hardest on when we rewrite buyer-side terms: explicit buffer setting (the number, not “network default”), explicit dispute window, and a definition of qualified call that ties to the gate, not just to duration. “Call lasting X seconds with confirmed service area, homeowner status, and same-day availability” is a workable definition. “Call lasting X seconds” is not.
We’ve covered adjacent terms in the pay-per-call insurance buyer pricing tiers writeup and the DID rotation piece. Both interact with the buffer in ways most buyers don’t model.
FAQ
What should the minimum call duration be on a final expense pay-per-call campaign?
The gate usually clears in the 50 to 60 second range. That’s where the health-question block resolves and the unfixable disqualifier is confirmed. Setting it higher trains publishers to stretch calls past the point where the agent already knows the policy can’t be written.
Why is the right Medicare buffer so much longer than the right final expense buffer?
Medicare’s qualifying flow includes the Scope of Appointment requirement plus an eligibility and effective-date check, none of which clear before about two minutes. Final expense underwriting questions resolve in under a minute. Same product category, fundamentally different gate timing.
How do I tell stall-padding apart from a legitimately slow agent on a recording?
Look for patterns that repeat across calls from the same publisher: hold music inserted after the IVR, redundant questions the IVR already captured, pauses right before your buffer triggers, and bill durations that cluster tightly at buffer-plus-three-to-five seconds. A slow agent is randomly distributed. Stall-padding clusters.
If I raise my buffer from 60 to 120 seconds, what happens to volume?
In the short term, publisher volume usually drops because their margin model assumed the lower number. Within two to three weeks you should see qualification rate climb and dispute rate fall on the remaining volume. If volume craters and stays cratered, the publisher was relying on calls that wouldn’t have survived the longer threshold anyway.
Should plumbing emergency and scheduled service use the same buffer?
No. Emergency plumbing qualifies in 45 to 60 seconds because the only gate is service area. Scheduled service needs 90-plus because the gate includes scope and scheduling. Running them on one campaign with one buffer over-bills emergency calls and under-bills scheduled ones simultaneously.
How does IVR length affect the right buffer setting?
Linearly. Every 15 seconds of pre-agent IVR pushes the disqualifier gate 15 seconds later from call connect. If your traffic source runs a 30-second IVR and the underlying agent-side gate is 90 seconds, your real buffer sits closer to 120.
What’s the fastest way to validate that my current buffer is wrong?
Pull 30 days of rejected calls and bucket them by call duration. If rejections cluster sharply between your buffer and your buffer-plus-30-seconds, the disqualifier gate sits inside your billable window. The buffer is too low. If rejections are spread evenly across durations, your buffer is fine and the issue is traffic quality.
Talk to Elevarus About Buffer Settings for Your Vertical
If you’re spending meaningfully on pay-per-call and you’re still on whatever number the platform pre-filled, the buffer is almost certainly costing you more than any payout negotiation could recover. The right setting is vertical-specific, dispatch-specific, and IVR-specific. The default is none of those.
If you want a second set of eyes on your campaign, book a free strategy call and ask about buffer mapping for your specific vertical. We’ll walk through your current setting, your dispute clustering, and what a defensible buffer looks like for the volume you’re running.
