Why Your Pay-Per-Call Campaigns Are Bleeding Money (And How to Fix It)

You've set up the campaigns. The calls are coming in. Your call volume dashboard looks healthy — 200, maybe 300 calls a day. The numbers feel good.

Then you check conversions. Flat. Or worse, declining while your call volume climbs.

Sound familiar?

We've worked with pay-per-call advertisers for long enough to know that this pattern shows up constantly. And it almost always comes down to three fixable problems that nobody talks about until you've already lost the money.

Here's what we've learned from building VeloCalls — call tracking and fraud detection for pay-per-call — and from watching campaigns go wrong in very predictable ways.

The Three Money Pits

1. Your Call Routing Is Dumber Than You Think

Most pay-per-call setups route calls based on one thing: geography. Caller's in Texas? Send them to the Texas buyer. Simple.

Except it's not.

We've audited dozens of pay-per-call accounts, and here's what keeps showing up: about 30-40% of call spend goes to connections that can't convert. Not because the leads are bad — because the routing is.

The most common waste? After-hours calls. A caller in Phoenix at 6pm reaches a plumbing company that closed at 5pm. The call goes to voicemail. The lead is dead. But the advertiser still paid for it.

Then there's skill mismatches. Someone calling about a gas line issue gets routed to a plumber who doesn't handle gas work. Three minutes of confusion, then a hang-up. Paid for that one too.

Geographic routing is the bare minimum. If that's all you're doing, you're leaving a third of your budget on the table. We've seen this across home services, insurance, legal — it's not industry-specific. It's a routing problem.

2. You Don't Have Fraud Detection (Or It's Not Enough)

This one's personal for us because it's why VeloCalls exists.

Pay-per-call fraud is different from click fraud, and it's uglier. (For the click-side version of this, our sister product wrote up the real cost of click fraud in 2026 — same fraud economy, different surface.) The common pay-per-call scams:

Repeat callers. Someone (or a bot using VoIP spoofing) calls the same tracking number 40 times a day. Each call lasts just past the minimum duration threshold — usually 90 seconds — and the advertiser pays every time. We've seen campaigns where a single phone number generated over 100 "qualified" calls in a month. All from the same source. That's thousands in payouts for nothing.

Call-back loops. A fraudster calls your tracking number, gets connected, immediately hangs up and calls again. Some older call tracking platforms count each connection as a separate billable call. This pattern can rack up thousands in charges over a couple months before anyone notices.

Duration stuffing. The caller reaches the buyer, says nothing (or plays a recording of background noise), and stays on the line past the duration threshold. The call "qualifies" based on length, the affiliate gets paid, and the buyer gets zero value.

Most call tracking platforms offer basic duplicate detection — they'll flag if the same number calls twice in 24 hours. That catches maybe 20% of fraud. The VoIP spoofing, duration stuffing, and sophisticated repeat patterns? Those sail right through.

Here's the industry context: according to the Communications Fraud Control Association (CFCA), global telecom fraud losses hit an estimated $38.95 billion in 2023. Pay-per-call is a small slice of that, but the same techniques apply — and the fraud operators know pay-per-call campaigns are softer targets than carrier-level infrastructure.

3. You're in the Wrong Verticals (Or the Right Verticals at the Wrong Price)

Contrarian take: most pay-per-call advice tells you to chase high-payout verticals. Insurance. Legal. Rehab. The calls pay $50-$200 each.

Here's what that advice leaves out: everyone else is chasing those verticals too.

The competition drives up traffic acquisition costs. In insurance pay-per-call, cost-per-call can hit $38-$45 for a payout of $55. That's a $10-$17 margin on a good call — and not every call converts. If your conversion rate is 25% (which is solid for insurance), you might actually be losing money once you factor in all costs.

Meanwhile, lower-profile verticals tell a different story. Pest control, for example: cost-per-call often runs $4-$7. Payouts are $18-$22. Conversion rates tend to be higher — people calling about a rat problem are extremely motivated buyers. The margin per call is smaller in absolute terms, but the ROI math works out significantly better.

The headline payout number is not the metric that matters. Effective margin is.

The Fix: What Actually Works

Enough problems. Here's what to do about them.

Fix your routing in layers. Geography is layer one. Add these:

• Time-of-day routing. Don't send calls to closed businesses. This sounds painfully obvious, and it is, but we audit accounts every month where it isn't set up. Build a routing schedule that matches buyer availability, including lunch hours and holidays.
• Skill-based routing. Tag your buyers by service type, not just location. A plumber who handles emergencies gets the "burst pipe" calls. A plumber who does remodels gets the "bathroom renovation" calls. We've seen this single change push conversion rates from the low teens to the low twenties.
• Capacity-based routing. If a buyer's already on a call, route to the next available buyer instead of sending to voicemail. Most platforms support this, but it's off by default.
• Weighted routing based on conversion history. Track which buyers actually convert calls into customers, and send more volume to the ones who close. One buyer converting at 35% should get significantly more calls than a buyer converting at 12%. This is probably the single highest-ROI routing change you can make.

Layer your fraud detection.

Single-point fraud detection doesn't work anymore. You need multiple signals:

• Duplicate number detection (table stakes — not enough alone)
• Call velocity monitoring — flag numbers generating more than 3 calls in 24 hours
• Duration pattern analysis — flag calls that consistently end within 5 seconds of the qualification threshold
• VoIP detection — a high percentage of VoIP-originated calls is a red flag
• Geographic anomaly detection — caller area code doesn't match the IP of the device that clicked the ad? Flag it
• Caller behavior scoring — first-time callers who ask specific service questions score high; repeat callers who go silent for 60 seconds score low

Stack these together and you catch 80-85% of fraud instead of 20%. The ROI on this is immediate — most accounts we onboard see a 20-40% reduction in wasted spend within the first 30 days. That's our data, not a third-party study, so take it for what it's worth. But the pattern is consistent.

If you want to attribute calls back to specific traffic sources and ad creatives, this is also where wiring your call data into a privacy-first analytics tool like JustAnalytics helps — you can spot which sources drive qualifying calls vs. which mostly drive fraud, and reallocate budget on real margin instead of vanity volume.

Pick verticals like an investor, not a gambler.

Stop chasing the highest payouts. Start calculating effective margins. The verticals where we've seen the strongest ROI for pay-per-call (based on what our users report — not a formal study):

1. Pest control — low competition, high caller intent, strong conversion rates
2. Garage door repair — $15-$20 payouts, but very low cost-per-call and high conversion
3. Locksmith — emergency intent drives conversion, though fraud is higher in this vertical (lots of VoIP spoofing)
4. Water damage restoration — seasonal but extremely profitable during peak months
5. HVAC — competitive but still good margins if your routing is tight

Insurance and legal can work. They do work for some operators running serious volume. But they require scale, strong fraud detection, and routing that most people don't have. If you're under $20K/month in spend, start with mid-payout verticals and build your systems there.

A Simple Evaluation Framework

We use this internally when evaluating any new vertical or campaign change. It's dead simple:

• Payout per qualified call: $X
• Cost to generate the call: $Y
• Qualification rate (% of calls meeting duration/intent thresholds): Z%
• Effective margin: ($X * Z%) - $Y

If the effective margin is below $5, don't touch it. We don't care if the payout is $200.

Start Here — Right Now

If you take one thing from this and go do it today:

• Pull your last 30 days of call data. Look for phone numbers that appear more than 3 times. Calculate what you paid for those repeat calls. It's probably more than you expect.
• Check your after-hours call volume. If more than 10% of your calls are hitting outside business hours, you're burning money on voicemails nobody returns.
• Calculate your effective margin per vertical. Not payout — margin. If any vertical is below $5 per call, either fix the routing or drop it.
• Set up fraud alerts for call velocity spikes. Any number generating more than 5 calls in a week should be flagged and reviewed.

These four things don't require new tools or a budget increase. They require attention. And they'll probably save you 20-30% of your current spend.

If you want the deeper technical setup — buyer scoring models, VoIP fingerprinting, multi-signal fraud stacking — that's what VeloCalls was built for. (And if you also run paid search, ClickzProtect handles the click-side fraud problem from the same playbook.) But honestly, start with the basics above. They're free, they're fast, and they work. More on this on the VeloCalls blog.

Frequently Asked Questions

Why do pay-per-call campaigns lose 30-40% of spend?

Three big leaks: (1) bad call routing — wrong vertical or wrong-state agents pick up, callers hang up; (2) call fraud and short-call abuse — call farms or duplicates trigger pay events; (3) wrong-source attribution — campaigns paying for calls that came from organic or another vertical. Fixing routing alone usually recovers 12-18% of spend in the first month.

What's the difference between call tracking and call routing?

Call tracking attributes a call to a source (campaign, keyword, ad). Call routing decides where the call goes once placed (which agent, which vertical, which buyer). Most campaigns get tracking right and routing wrong — money walks out the door because the routed agent is unprepared, in the wrong time zone, or never picks up.

Are pay-per-call leads TCPA compliant by default?

No. TCPA compliance depends on the consent capture upstream (the publisher or affiliate generating the call) and how the call is dialed. Express written consent is required for autodialed calls to mobile numbers in the US. If you're buying calls through an affiliate network, you need consent records on file or you're carrying TCPA liability. Audit your top 5 sources first.

How do I detect call fraud in pay-per-call?

Watch for these signals: calls under 30 seconds with no transfer, calls from VOIP-only numbers, repeated calls from the same number across different campaigns, and unusual hour-of-day distribution (fraud spikes at 2-5am ET). VeloCalls flags these automatically; if you're on Ringba or CallRail, you'll need to build the rules yourself in their reporting.