Fixing a High Call Return Rate: Why Buyers Reject Your Pay-Per-Call Leads

Three weeks ago I pulled the return report for a plumbing campaign and nearly closed my laptop. 23% call return rate. One in four calls coming back rejected, clawbacks hitting within 72 hours, and the buyer threatening to pause the campaign entirely.

Traffic looked fine. Calls connected. Durations averaged 4 minutes — well above the 90-second billing threshold. But the buyer kept sending calls back with disposition codes I'd never seen before: "out of area," "renter not homeowner," "no actual leak," "caller hung up on agent."

I stared at those codes for twenty minutes before it clicked. The problem wasn't traffic quality. It was everything that happened after the call connected.

A high call return rate is a publisher's most expensive leak. You've already paid for the click, already paid for the call minutes, already counted that revenue — and then it disappears. Worse, high return rates get you capped, paused, or dropped from buyer networks entirely. I've watched publishers lose $15,000/month relationships because they couldn't get their return rate under control.

This guide covers how to diagnose what's causing returns and how to fix it. The four culprits: weak qualification, geographic mismatches, short effective duration, and intent problems. We'll go through each one. By the end, you'll have a checklist and a triage order.

Understanding Why Buyers Return Calls

Buyers don't return calls for fun. Every return costs them admin time. They return calls because the call didn't convert — and they believe it shouldn't have been billed in the first place.

The most common return reasons break into four categories:

Qualification failures. The caller didn't meet basic criteria. For home services, that's usually "not the homeowner" or "commercial property, not residential." For legal, it's "no injury" or "accident outside statute of limitations." For Medicare, it's "not enrolled in Parts A and B" or "already has a broker." These are callers who should have been filtered before they ever reached a buyer.

Geographic mismatches. The caller's location doesn't match the buyer's service area. A roofing contractor in Dallas gets a call from someone in El Paso. The call "qualified" — homeowner, storm damage, ready to schedule — but the buyer can't service that zip code. Returned.

Duration technicalities. The call hit the billing threshold (usually 60-120 seconds) but didn't represent genuine engagement. The caller stayed on hold for 90 seconds, then hung up. Or the IVR took 70 seconds, the caller reached an agent, and immediately disconnected. The clock ran, but nothing real happened.

Intent problems. The caller wasn't actually in-market. They were price shopping with no intent to buy. Or they wanted information, not service. Or they called the wrong number and stayed on long enough to trigger billing before realizing the mistake. These are the hardest to filter because the caller often sounds qualified.

Different return reason codes require different fixes. Don't treat returns as one problem. Break them apart. (I learned this the hard way after trying to fix "returns" as a single category for two months straight. Waste of time.)

Step 1: Pull Your Return Data and Categorize It

Before you fix anything, understand what you're fixing.

Export your returns from the past 30 days. Every platform — VeloCalls, Ringba, Invoca, whatever you're running — should give you return reason codes. If your buyers aren't providing codes, ask them to. "Call returned" without context is useless.

Group returns by reason:

Calculate the percentage each category represents. If 50% of your returns are geographic, that's your first fix. If 30% are qualification failures, that's your lever. Don't try to fix everything at once. Attack the biggest bucket first.

(I once spent two weeks optimizing IVR questions when 60% of my returns were geographic mismatches that a simple zip code filter would have caught. Embarrassing. Don't repeat my mistakes.)

Step 2: Fix Qualification Gaps

If buyers are returning calls for qualification failures, your IVR or pre-screen isn't filtering properly.

Audit your current qualification questions. What are you asking? What order? A surprising number of publishers run zero qualification — calls connect straight to buyers with no filter. That's fine if you're buying perfect traffic. It's not fine if you're buying from networks or running broad paid search.

Add disqualifiers at the front. The question that filters out the most non-buyers should come first. For HVAC and plumbing: "Are you the homeowner, or are you calling on behalf of the homeowner?" If no, exit. For personal injury: "Were you or a family member injured in this accident?" If no, exit. See our call qualification script guide for the full structure.

Don't rely on caller honesty alone. Some callers lie. Some callers don't understand the question. "Are you the homeowner?" gets "yes" from renters who think it means "are you the person who lives here." Consider phrasing: "Do you own this property, or do you rent?" More explicit, harder to misunderstand.

Match qualification to buyer requirements. If your buyer requires "actively leaking water" for emergency plumbing, ask that. If they require "accident within 2 years," ask that. Your IVR should mirror buyer intake criteria exactly. When those drift apart, returns spike.

I've seen publishers add a single question — "Is the homeowner available to speak with us today?" — and drop returns 8 percentage points. One question. Eight points. Low-hanging fruit exists if you look for it.

Step 3: Fix Geographic Mismatches

Geographic returns are the most fixable. You know where buyers operate. You know where callers are. Match them.

Implement zip code verification in IVR. Ask callers for their zip code. Yes, it adds 10 seconds. Yes, completion rates drop slightly. The trade-off is worth it when 20% of your calls are getting returned for geography.

Voice prompt: "To connect you with a local specialist, please enter your five-digit zip code." In VeloCalls, you map the response against your buyer coverage matrix and route accordingly. If no buyer covers that zip, route to national overflow or exit gracefully.

Don't trust ANI-based geo alone. A caller's phone area code doesn't mean they're in that area. Mobile numbers are portable. Someone with a 312 Chicago number might be calling from Phoenix about a Phoenix property. ANI lookup is a hint, not a fact. Combine it with explicit zip code confirmation.

Build a coverage matrix. List every buyer and every zip code (or zip3) they'll accept. Store this somewhere accessible. When you add a new buyer or a buyer changes coverage, update the matrix. Route against the matrix, not against assumptions.

Example setup in routing logic:

IF caller_zip IN buyer_a_coverage
  ROUTE to buyer_a
ELIF caller_zip IN buyer_b_coverage
  ROUTE to buyer_b
ELIF caller_zip IN national_overflow_coverage
  ROUTE to overflow_buyer (lower payout, accepts all)
ELSE
  PLAY "We don't currently have a partner in your area. Thank you for calling."
  EXIT

This is basic. Most publishers skip it anyway.

Don't be most publishers.

For traffic source filtering before calls even happen, ClickzProtect can block clicks from geographic regions where you have no buyer coverage. Why pay for a click from Wyoming if none of your buyers serve Wyoming?

Step 4: Fix Duration Problems

Duration-based returns are tricky because the call technically hit your threshold. The buyer is saying: "Yes, it was 2 minutes. No, nothing happened in those 2 minutes that counts."

Audit what's consuming duration. Pull a sample of returned calls and listen. Where did the time go?

Common findings:

• Long hold times before agent pickup (counts toward duration but isn't conversation)
• Extended IVR with lots of prompts (caller on IVR for 90 seconds, then 10 seconds with agent)
• Caller hung up during transfer (warm transfers take 15-30 seconds, cold transfers faster)
• Agent couldn't help and caller disconnected quickly after connection

Tighten your IVR. If your IVR is eating 60+ seconds of a 90-second call, the agent barely gets 30 seconds of conversation. Not enough to qualify and schedule. Compress IVR time. Fewer questions asked faster. Under 45 seconds for standard campaigns — yes, I know that sounds aggressive, but I've tested it and the data backs it up.

Track time-to-agent. Most platforms separate IVR time from conversation time. If callers spend 80 seconds navigating IVR and 15 seconds with an agent, that's a problem your return rate will reflect. See our IVR abandonment study for benchmarks by vertical.

Consider conversation-start billing. Some buyer agreements bill from when the agent answers, not when the call connects. If you can negotiate that, duration gaming becomes harder and both sides align on what "qualified" means. It's a harder sell to buyers, but worth proposing.

Flag transfer failures. If a caller connects, gets transferred, and the transfer fails (buyer doesn't answer, busy signal, wrong extension), that shouldn't count as a completed call. Build logic to detect failed transfers and exclude those from billing. Most platforms track transfer success separately — use it.

Step 5: Address Intent Problems

Intent-based returns are the hardest to fix because they require filtering at the traffic source, not just the IVR.

Understand what "low intent" looks like in your vertical.

In home services: "Just getting prices," "Not ready to schedule anything," "Just curious about costs."

In legal: "Wondering if I have a case," "Haven't decided if I want a lawyer," "My accident was years ago and I'm just now looking into it."

In insurance: "Comparing quotes but not switching anytime soon," "My renewal isn't for six months."

None of these are necessarily bad leads. But they're not leads that convert in a single phone call to an intake team expecting decision-ready callers.

Align intent with buyer expectations. If your buyers want "ready to book this week" and you're sending "shopping for next month," returns will spike. Either find buyers who accept early-funnel leads (usually at lower payouts) or filter harder for purchase intent.

Add intent questions to IVR — carefully. "Are you looking to schedule service in the next two weeks, or are you just getting information for later?" This sorts callers. But it's also where completion rates drop. Callers don't like admitting they're not ready to buy. Many will just hang up rather than self-select as "not serious."

Filter upstream. If 30% of your calls are low-intent and you can't fix it with IVR, your traffic source is the problem. Certain keywords attract researchers, not buyers. Certain placements (GDN, Facebook browse, content arbitrage) skew to curiosity-driven clicks that don't convert. JustAnalytics can correlate traffic source to downstream call outcomes, so you can see which publishers or keywords generate "qualified on paper, returned in practice" calls.

The ugly truth: some intent problems can't be fixed post-click. You have to buy better traffic. That's a budget and sourcing conversation, not an IVR optimization.

Nobody wants to hear that. I certainly didn't when I was running $8K/month on content arb traffic and wondering why nothing converted. But there it is.

Step 6: Build Monitoring for Ongoing Health

Fixing returns once doesn't mean they stay fixed. Monitor weekly.

Track return rate by buyer. Different buyers have different standards. A 5% return rate with Buyer A and a 22% return rate with Buyer B means the problem is either Buyer B's pickiness or your fit for Buyer B's criteria. Either renegotiate expectations or stop sending calls there. Building dashboards in JustAnalytics makes this buyer-level tracking automatic.

Track return rate by traffic source. Certain publishers or campaigns may produce calls that look fine but consistently return. If Source X has a 25% return rate while your average is 10%, cut Source X. The CPL savings aren't real if a quarter of the calls claw back.

Track return rate by day-of-week and time-of-day. I've seen return rates spike on weekends when buyer staffing is thinner and callers reach overwhelmed agents who can't properly intake. If your Saturday return rate is 3x your Tuesday rate, you have a buyer capacity problem, not a lead quality problem.

Set alerts. If return rate exceeds 15% for any rolling 7-day window, investigate immediately. Don't wait for the monthly review. By then you've lost money you'll never recover.

Honestly? Most of us learn this by getting burned once. Maybe twice. Fine. Just don't get burned a third time.

Common Patterns and Quick Fixes

Pattern: Returns spike after adding a new traffic source.

Check: That source isn't converting. Pause it, analyze the call recordings, identify the gap. Maybe it's geo, maybe it's intent, maybe it's bots. If it's bots, ClickzProtect should be filtering upstream.

Pattern: Returns cluster around specific times.

Check: Buyer availability. If calls at 4pm get returned at 2x the rate of 10am calls, buyers are probably at capacity or wrapping up their day. Route afternoon calls to buyers with better late-day coverage.

Pattern: Returns mostly come with "no conversation occurred."

Check: Transfer success rate. Your calls are connecting to IVR but not completing transfer to agents. The agent endpoint might be wrong, busy, or the transfer is timing out.

Pattern: Same caller appears in multiple returns.

Check: You're routing duplicates. Same person calling multiple tracking numbers (maybe from different ads) and reaching the same buyer pool. Dedupe by ANI before routing.

When to Walk Away from a Buyer

Sometimes the return rate isn't your fault. Some buyers return everything. They're using clawbacks as a negotiating tactic, hoping you'll accept lower payouts to "improve quality."

Red flags:

• Return reasons are vague ("not qualified") with no actionable detail
• Returns come in batches at end of billing period
• Return rate increases when your volume increases (capacity issue disguised as quality issue)
• You've fixed every identifiable problem and returns stay constant

If a buyer's return rate is 20%+ and they can't explain what they want differently, move on. Find buyers who'll work with you on dispositions and give you data to improve. The best buyer relationships include return reason codes, call recordings of returned calls (so you can audit), and monthly calibration calls.

I've fired buyers with 25% return rates and replaced them with buyers at slightly lower payouts but 6% return rates. The math always favors the lower-return buyer. Always.

Strong opinion: if a buyer can't give you return reason codes within 48 hours, they're not worth working with. Period. I don't care what payout they're dangling.

Next Steps

You've diagnosed your return categories, tightened qualification, fixed geo routing, compressed IVR duration, and addressed intent filtering. Here's what to do next.

Implement weekly return audits. Every Monday, pull returns from the prior week, categorize by reason, compare to the week before. Treat it like a weekly standup for your call quality.

Build a buyer scorecard. Track each buyer's return rate, average payout, payment terms, and communication quality. The best buyer isn't the one with the highest payout — it's the one with the best net payout after returns.

Test qualification changes in isolation. When you add a new IVR question or change routing logic, measure the impact on both completion rate and return rate. A question that drops completion 10% but drops returns 15% is probably worth it. A question that drops completion 15% and doesn't move returns is just losing you money.

For more on optimizing the upstream traffic layer, see our junk call filtering guide. And if your answer rates are the issue before calls even get to qualification, check our answer rate fix guide.

Returns are feedback. Expensive, annoying feedback — but feedback.

Listen to what buyers are telling you with their disposition codes. Fix the gaps. The publishers who run under 5% return rates aren't doing magic. They're just paying attention. And yeah, they probably made all the same mistakes we did before they figured it out.

Frequently Asked Questions

What is a good call return rate in pay-per-call?

Under 10% is solid. Under 5% is excellent. If buyers are returning more than 15% of your calls, something's broken — weak qualification, geographic mismatches, or intent problems that your IVR isn't catching. Industry benchmarks show top publishers maintain 3-7% return rates while bottom-tier publishers run 20%+ and wonder why buyers cap their volume.

Why do buyers return calls that met the minimum duration?

Duration alone doesn't equal quality. A 3-minute call where the caller was confused, gave fake information, or wasn't decision-ready still gets returned. Buyers evaluate disposition — did the caller have genuine intent, correct contact info, and authority to purchase? Many publishers hit duration thresholds with callers who were just asking questions or price shopping with no intent to buy. The call "qualified" on paper but failed in reality.

How do I reduce geographic mismatches in my call routing?

Verify caller location before routing. Use ANI lookup for area code, but don't trust it completely — mobile numbers travel. Add an IVR prompt asking for zip code, then validate against your buyer coverage map. Route mismatches to a national overflow buyer or exit gracefully. Some platforms like VeloCalls support geo-filtering at the routing level so mismatched calls never reach mismatched buyers.

Should I add more IVR questions to reduce returns?

Sometimes, but carefully. Each additional question drops your completion rate 5-12%. The goal isn't more questions — it's better questions positioned earlier. If 40% of your returns are non-homeowners, add that disqualifier at position 1, not position 6. If returns are mostly "not ready to buy," you probably can't fix that with IVR questions — you're buying bad traffic. Filter upstream instead.

Try VeloCalls for Your Vertical

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

See pricing → · Book a demo