Back to Blog
Guide11 min read

How Pay-Per-Call Billing Works: Billable Duration, Caps, and Disputes

Billable-second thresholds, buyer caps, duplicate windows, dispute flows — the billing mechanics that separate pay-per-call operators who know their margins from those who don't.

Picture a bad month for a mid-size plumbing operator: $4,200 for 60 calls, thirty-eight of them under 90 seconds, fourteen landing in voicemail, three repeat callers within a week. None of that shows up until someone actually pulls the recordings.

That's the billing literacy gap in pay-per-call. The mechanics aren't complicated — billable thresholds, duplicate windows, caps, credits — but if you don't understand them, you're handing over cash for calls that never had a chance to convert. Networks aren't going to explain this unprompted. They're not obligated to. The invoice goes out, you pay it, life goes on.

This guide covers how pay-per-call billing actually works. No jargon soup. Real numbers where they matter. When you finish reading, you'll know exactly which levers to pull before you pay your next invoice.

What and Why

Pay-per-call billing determines when a phone call triggers a payment from buyer to publisher. Simple on the surface. A mess underneath.

Why does this matter right now? The industry has matured past "I pay you per call" into conditional logic that would make a tax attorney wince: duration thresholds, IVR completion requirements, geographic filters, time-of-day rules, duplicate suppression, and real-time bidding. New buyers enter pay-per-call from digital lead gen assuming it works like CPL web forms. It doesn't. Not even close. A web lead is either submitted or not. Binary. A phone call has a dozen variables that determine whether payment triggers — and another dozen that determine whether you can dispute it after the fact. Plenty of operators run profitable campaigns for months without ever working out why some calls bill and others don't. That's luck, not strategy. Luck runs out.

Core Concepts

Before we get into process, five concepts you need cold.

Billable Duration vs. Call Duration. Call duration is total time on the line. Billable duration is the portion that counts. If your contract says "90-second minimum," a 2-minute call clears it and earns one flat payout in per-call models. In per-minute models it bills either 2 minutes or 30 seconds, depending on whether your contract counts the time before the minimum. The gap between these two numbers is where margin hides.

Qualification Threshold. The minimum bar a call must clear to be billable. Usually time-based (90 seconds, 120 seconds), but sometimes includes IVR completion, verbal confirmation, or geographic match. "Qualified call" means something different on every rate card. Read yours.

Duplicate Window. How long a caller ID is suppressed after billing. A 30-day window means if the same phone number calls again within 30 days, the second call doesn't trigger payment. Windows range from 7 days (home services — repeat calls are often real customers) to 45 days (mass tort — leads are one-time events). Mismatched windows inflate or deflate your true CPL.

Caps. Limits on exposure. Daily call caps. Daily spend caps. Concurrency caps (max simultaneous calls). Caps exist to protect buyers from traffic spikes. Without them, a rogue publisher can blow your monthly budget in a weekend — and it doesn't take bad intent, just one unfiltered traffic source having a good day. Set them before you need them, not after.

Dispute/Credit Flow. The process for contesting a billable call after the fact. Pull recording, document the issue (voicemail, out-of-area, wrong intent, short duration), submit ticket, wait. Networks vary wildly in how they handle disputes. Some credit automatically on recording evidence. Others deny everything and make you escalate. Know your network's policy before you need it.

The Complete Billing Process

Step by step, from ringing phone to invoice line item.

Step 1: Call Arrives

A caller dials your tracking number. The platform captures ANI (caller ID), time, and — if you're running DNI — the traffic source that triggered the call. This metadata travels with the call through the entire billing flow.

Step 2: Routing Decision

Your ring tree (some networks call it a ping tree) evaluates which buyer should receive the call. Real-time bidding collects offers in a fraction of a second. The highest bid above your floor wins. No bid above floor? Call goes to default routing or hangup messaging.

The bid that wins isn't always the payout you receive. Some networks take a spread. Some take a flat cut. Read your IO. Actually read it. The whole thing. Yes, it's tedious — but the three paragraphs you skip are the ones that bite you.

Step 3: IVR and Pre-Qualification

Before the call reaches a live agent, most setups run the caller through an IVR: a menu asking them to press 1 for service or 2 for sales, or a geographic check asking whether they're calling about a property in Texas.

IVR completion affects billing. Some contracts require the caller to pass IVR to trigger payment. Others bill on call start regardless. If you're a buyer, you want IVR completion required — it filters out pocket dials and wrong numbers before you pay for them.

Step 4: Agent Connect and Timer Start

The call transfers to a live agent. In most contracts, the billable timer starts here — at agent connect, not at call start. The difference matters. If your IVR is 45 seconds and the billable clock starts at agent connect, that IVR time doesn't count toward the minimum threshold.

Watch for contracts that start the timer at call start. Those are worse for buyers.

Step 5: Duration Monitoring

The platform watches the call clock. When the call crosses your qualification threshold (say, 90 seconds), it flips from "pending" to "qualified" in real time. Most platforms expose this in dashboards as it happens. If you're running capacity-based routing with fallback, you want to know the moment qualification hits so you can manage concurrent call limits.

Step 6: Call Ends

Caller or agent hangs up. Total call duration locks. The platform calculates billable duration under your contract's rules — for example, leaving out IVR time or the time before the threshold if the contract says so.

Here's where duplicate suppression runs. The platform checks the caller ID against recent calls. If the same ANI triggered a billable event within your duplicate window, this call is marked non-billable. Good platforms show you the prior call it matched against. Bad platforms just mark "duplicate" with no context.

Step 7: Billing Record Created

A billable event — qualified call exceeding threshold, passing duplicate check, matching geographic and time-of-day rules — generates a billing record. This includes:

  • Call ID
  • Timestamp (start, connect, end)
  • ANI and destination
  • Duration (total and billable)
  • Source/publisher ID
  • Buyer bid and payout
  • Recording URL (if enabled)

This record is what you reconcile against when the invoice arrives.

Step 8: Invoice and Reconciliation

Networks batch billing records into invoices — daily, weekly, or monthly depending on volume. Your job at invoice time:

  1. Pull the billing detail report.
  2. Spot-check 10-20 calls against recordings.
  3. Flag anything that shouldn't have billed (voicemails, short durations, duplicates, out-of-area).
  4. Submit disputes before the dispute window closes. Most networks give you 7-14 days. Miss the window, lose the credit.

For attribution on the web side of these calls, JustAnalytics handles source-level tracking if you're running multi-channel campaigns.

Advanced Tips

Negotiate threshold and window separately. Most buyers negotiate CPL but accept default threshold and duplicate window. Bad move. A 120-second threshold vs. a 90-second one changes how many calls bill, and with it your effective CPL. A 14-day duplicate window vs. a 30-day one changes your duplicate rate, most of all in high-repeat verticals. Negotiate all three.

Log your own call times. Don't trust the network's timestamp exclusively. Run a parallel log on your side — call start, agent connect, call end — and compare against invoice detail. Drift happens, and a few seconds shaved off hundreds of calls adds up fast. Annoying to set up. Worth it at scale. Treating the network's clock as gospel is an easy habit to fall into — and an expensive one to keep.

Cap concurrency, not just daily volume. A daily cap of 100 calls doesn't help if 40 of them arrive in the same 10-minute window and your agents can't handle the load. Unanswered calls that hit voicemail after a 45-second hold still bill in some contracts. Set concurrency caps (e.g., max 8 simultaneous calls) to match your actual staffing.

Track dispute outcomes by publisher. Some publishers generate 2% dispute rates. Others generate 18%. The invoice looks the same until you slice it. Publishers with sustained 10%+ dispute rates aren't fixable with credits — they're burning your team's time. Cut them. It's math. A spreadsheet with publisher ID, call count, and dispute count is enough to see the pattern — no special tooling required.

Common Mistakes

Paying for voicemail. Contracts that bill on "call answered" include voicemail pickup. If your buyer's line rolls to voicemail after 30 seconds and the caller leaves a message, that's a billable call. Fix: require live-agent connect for billing, not just call answered.

Ignoring duplicate windows. Default windows are often 14 days. In legal verticals where the same claimant calls five firms, that's way too short — you're paying for the same lead multiple times. In HVAC where a homeowner might call back to schedule, 14 days might suppress legitimate repeat business. Match your window to your vertical's sales cycle.

Missing dispute deadlines. Networks love this. Invoice arrives Friday, dispute window closes in 7 days, your team doesn't review until Monday week two. Credits denied, window closed, money gone. Put invoice review on a calendar trigger.

Trusting network QA tags. "Quality Score: High" means nothing if you haven't listened to the recording. A 94-second voicemail can carry a "High" quality tag just as easily as a real conversation can. Frustrating? Deeply. QA tags are the network's opinion — often generated by an algorithm that's optimizing for something other than your ROI. Your ears are the source of truth. For more on this, see our pay-per-call bleeding money guide.

No caps on new publishers. First week with a new publisher? Set aggressive caps. 20 calls/day max. If quality holds, scale. If quality tanks, you've lost $800, not $8,000. Generous caps on an unproven source are how operators end up with a $6K bill to a feed that's 70% VOIP spoofs — an expensive way to find out.

Frequently Asked Questions

What's the difference between call duration and billable duration?

Call duration is total time from connection to hangup. Billable duration is the portion that counts toward payout, and it depends on a minimum threshold (often 90-120 seconds). In per-call models, a call that clears the threshold earns one flat payout. In per-minute models, contracts differ: a 4-minute call against a 90-second threshold bills either all 4 minutes or 2.5, depending on whether the first 90 seconds count. Buyers use the threshold to avoid paying for calls that hang up before qualification.

How do duplicate windows work in pay-per-call?

Duplicate windows define how long a caller ID is suppressed after a billable call. A 30-day window means if the same number calls again within 30 days, the second call isn't billed. Windows typically run 7-30 days depending on vertical. Legal uses longer windows (leads are one-time). Home services uses shorter windows (repeat service calls are legitimate). Match your window to the buyer's sales cycle.

What should I do if a network denies my dispute?

Pull the call recording and transcript. Document the specific issue — out-of-area, voicemail, wrong intent, sub-threshold duration. Escalate with evidence attached. If the network still denies, note the pattern and reduce volume to that source. Persistent unjustified denials are a signal to exit the relationship. Track dispute rates by publisher — anything over 8% sustained is a red flag.

How do buyer caps protect against overspend?

Caps limit exposure: daily call caps (e.g., 50 calls/day), daily spend caps (e.g., $2,000/day), and concurrency caps (e.g., 5 simultaneous calls). When caps hit, calls route to the next buyer in the ring tree. Without caps, a traffic spike from a new publisher can blow your weekly budget in hours. Set caps 20% above your target run rate and review weekly.

Resources


Try VeloCalls for Your Vertical

Pay-per-call routing platform built for HVAC, plumbing, roofing, PI lawyers, Medicare brokers, and insurance. Smart routing, real-time bidding, visual IVR builder, AI conversation intelligence (transcription, sentiment, summarization). Per-minute pricing — Managed starts at 4¢/min, BYOC at 2¢/min, both drop as you scale.

See pricing →

pay-per-call-billingbillable-durationcall-trackinglead-generationdispute-resolutionbuildinpublicsaasstudioaiworkforcebuildwithclaude
Share

Ready to try VeloCalls?

Set up intelligent call tracking and routing in minutes. No credit card required.

Get Started Free

Stay Updated

Get the latest articles and industry insights delivered to your inbox.

No spam. Unsubscribe anytime.

Related Articles