How Publishers Can Forecast Pay-Per-Call Revenue Before Scaling Spend

Last month I watched a publisher dump $14,000 into Google Ads for an HVAC campaign without running a single projection model. His logic: "The payout is $55 per call, I'm getting calls at $22, that's 2.5x return, let's go."

He didn't account for billable rate. Or answer rate. Or the buyer's 90-second duration threshold that only 61% of his calls hit. I've made this exact mistake myself — twice, actually — so I'm not judging.

Three weeks later? His actual RPC was $28.40 against a $22 cost per call. A 1.29x return instead of 2.5x. Still profitable — barely — but nowhere close to what he'd budgeted for. The "scale" turned into a painful lesson in unit economics.

This is the mistake publishers keep making. They see payout rates, subtract cost per call, and assume the margin is real.

It's not.

The margin only exists after answer rates, billable thresholds, conversion rates, and payout reliability filter through. If you're bleeding money on pay-per-call campaigns without knowing why, this is usually where the leak starts.

Here's the forecasting model that lets you project actual revenue before you 5x your ad budget and learn the hard way. (I wish someone had shown me this four years ago.)

The Four Variables That Actually Determine Your Revenue

Forget "payout minus cost per call." That math is wrong. Here's what actually happens to every dollar you spend on traffic:

Projected Revenue = Clicks × Call Rate × Answer Rate × Billable Rate × Payout

Or simplified to the per-call level:

Revenue Per Call (RPC) = Payout × Billable Rate × Answer Rate Adjustment

Let me break down each variable, because most publishers only track one or two of them.

1. Answer Rate — What percentage of calls actually connect to a human? Industry range is 55-70% depending on vertical and caller ID reputation. If your tracking numbers are spam-flagged, you're running 40%. (And probably don't know it.)

2. Billable Rate — What percentage of answered calls meet the buyer's qualification threshold? Most buyers require 90-120 seconds of talk time. Some require IVR qualification. Some require geographic match. A "billable rate" of 65% means 35% of your answered calls pay you nothing.

3. Conversion Rate — Of the calls that hit billable threshold, what percentage actually get paid? Disputes happen. Chargebacks happen. Buyers who claim calls "didn't meet criteria" after the fact. For established buyer relationships, this is 95-100%. For new buyers, budget for 85-90% until trust is built.

4. Payout — The headline number everyone focuses on. But $55/call means nothing if only 40% of your calls hit billable and 15% of those get disputed. I think this is where most publishers go wrong — they anchor on payout like it's gospel.

The model that matters:

True RPC = Payout × Billable Rate × Conversion Rate
Projected Weekly Revenue = Weekly Calls × Answer Rate × True RPC

If you're driving 200 calls/week at $55 payout, 64% billable rate, and 95% conversion, your True RPC is $33.44. At 62% answer rate, 124 calls connect, and you're projecting $4,146/week in actual payouts. Not $11,000.

That's the gap between naive math and reality. Honestly, it's embarrassing how long I ran campaigns without understanding this.

Building Your Baseline: What to Track Before You Scale

You can't forecast without data. Here's the baseline you need before any scaling decision.

Pull from your call tracking platform (VeloCalls, Ringba, CallRail, etc.):

• Total calls last 14 days
• Answer rate (calls that connected / total calls)
• Average call duration
• Billable calls (calls meeting buyer threshold)
• Billable rate (billable / answered)

Pull from your buyer reports or payout dashboard:

• Total payouts received
• Disputes or chargebacks
• Conversion rate (payouts / billable calls submitted)

Calculate from your ad platform:

• Total ad spend
• Cost per call (spend / total calls)

Now you have the inputs. Let's build the model.

Example baseline:

True RPC = $48 × 0.61 × 0.94 = $27.53

Gross margin per call = $27.53 - $19 = $8.53

Gross margin % = 31%

That's your actual unit economics. Not $48 - $19 = $29. The real margin is $8.53. And yes, I know that's frustrating to hear when you thought you were making $29/call.

If you want tighter click-level attribution before calls even happen, JustAnalytics tracks sessions without the GDPR consent headaches — useful for knowing which landing page variants drive calls that actually bill.

The Projection Model: What Happens When You Scale

Now the useful part. You've got baseline metrics. What happens if you 3x your ad budget?

Assumption 1: Cost per call holds steady. This is optimistic. In practice, CPCs rise as you scale because you're exhausting the most efficient auction inventory first. Budget a 10-20% CPC increase at 3x scale.

Assumption 2: Answer rate drops slightly. More volume means more strain on caller ID reputation. Your tracking numbers get more active. Spam algorithms notice. Budget a 3-5 point answer rate drop at 3x volume if you're not actively managing number hygiene. Our IVR abandonment rate study found that menu depth compounds the problem — every extra IVR level costs 6 points of callers before they even reach an agent.

Assumption 3: Billable rate stays constant. This one usually holds unless your traffic source quality degrades. Same IVR, same buyer thresholds, same billable rate.

Assumption 4: Conversion rate stays constant. Buyer payout reliability shouldn't change with your volume — unless you're overwhelming their intake capacity. Check buyer caps before scaling.

Projection at 3x spend ($2,850 → $8,550):

Wait. Net margin dropped from $509 to $260 at 3x spend?

Yeah.

That's what happens when CPC rises 10% and answer rate drops 4 points. The unit economics got squeezed from both sides. This is the whole point of forecasting — you catch this before you spend the $8,550, not after. (Ask me how I know.)

The fix: Either maintain answer rate through number hygiene (warming new tracking numbers, rotating out flagged ones) or accept lower margin at scale. Or find traffic sources where CPC doesn't scale linearly.

Scenario Modeling: Pessimistic, Realistic, Optimistic

Don't run one projection. Run three.

Pessimistic: CPC +20%, answer rate -6 points, billable rate -3 points. Everything that can degrade, does.

Realistic: CPC +10%, answer rate -4 points, billable rate holds. Based on typical scaling patterns.

Optimistic: CPC +5%, answer rate -2 points, billable rate +2 points. You've optimized well and traffic quality improves slightly.

If the pessimistic case still shows positive margin, scale. If pessimistic goes negative but realistic is strong, scale cautiously and monitor weekly. If realistic is barely positive, don't scale — fix your baseline first.

Here's what the three scenarios look like for our example at 3x spend:

The pessimistic case loses money. That's a red flag.

This campaign isn't ready for 3x scale. The margin is too thin at baseline — $8.53/call doesn't survive efficiency degradation. I'm not saying don't scale. I'm saying don't scale yet.

What to do instead: Improve baseline metrics before scaling. Get billable rate from 61% to 70% by tightening IVR qualification. Get answer rate from 64% to 68% through caller ID hygiene. Then re-run the projection. You'll need baseline margin of $12-15/call before 3x scale makes sense on these CPCs.

Vertical-Specific Adjustments

The model works across verticals, but the inputs differ. Here are reference ranges — use your actual data, but if you're entering a new vertical, these give you starting points.

Home Services (HVAC, plumbing, roofing):

• Payouts: $35-65 per qualified call
• Billable rate: 55-70% (highly dependent on IVR setup)
• Answer rate: 58-68% (local presence matters)
• Typical True RPC range: $14-32

Personal Injury Legal (PI auto, mass tort):

• Payouts: $150-400+ depending on case type
• Billable rate: 40-60% (strict qualification, longer calls)
• Answer rate: 55-65% (some spam issues in legal verticals)
• Typical True RPC range: $35-150

Medicare/Insurance:

• Payouts: $40-80 (AEP/OEP periods higher)
• Billable rate: 50-65% (compliance requirements add friction)
• Answer rate: 50-60% (heavy robocall association in category)
• Typical True RPC range: $12-32

For legal, the high payout partially compensates for lower billable rates — but the unit economics are volatile. One good mass tort week can be worth a month of HVAC. One bad week can wipe out two months. Forecast conservatively. (In my opinion, legal is not for beginners. Your mileage may vary, but I'd start with home services.)

If you're evaluating which vertical to enter, our home services pay-per-call playbook covers HVAC, plumbing, and roofing with real CPL ranges and publisher names.

Weekly Tracking: The Numbers That Predict Revenue Shifts

Once you're scaled, forecast weekly to catch degradation before it costs you.

Track every Monday:

• Trailing 7-day answer rate (vs. prior week and 30-day average)
• Trailing 7-day billable rate (same comparisons)
• True RPC (recalculate weekly)
• Cost per call trend
• Gross margin per call

Set alerts for:

• Answer rate drop > 5 points week-over-week → investigate caller ID reputation
• Billable rate drop > 4 points → audit IVR or buyer threshold changes
• True RPC drop > 10% → something broke; diagnose immediately
• CPC increase > 15% → traffic source is getting competitive; evaluate alternatives

The operators who scale without blowing up are the ones who watch these numbers weekly — not monthly. Monthly reviews mean you've already lost $5K before you notice the trend. Been there. It hurts.

ClickzProtect catches bot traffic before it becomes garbage calls. If your cost per call is rising but call quality is dropping, you might be paying for fake clicks that generate fake calls. Filter upstream.

Forecasting for New Buyer Relationships

New buyers are wildcards. You don't have conversion rate history. You don't know their dispute patterns. Here's how to model conservatively.

Assume 85% conversion rate until you have 60+ days of payout data. Some new buyers dispute 20%+ of calls during "evaluation." Build that into the model.

Start with small volume — 20-30 calls/week for the first month. This isn't about being conservative; it's about getting clean data. You need actual billable rate and conversion rate before you can forecast.

Watch payout timing. If a buyer pays net-30 and you're projecting weekly revenue, your cash flow model is wrong. You won't see that money for a month. Some buyers pay net-45. Some "forget" invoices. Factor this in. If you're considering AI voice qualification to improve billable rates on new buyer flows, run the CPL math first — it doesn't always help.

Build buyer reliability scores. After 90 days, each buyer should have a reliability score: (actual payouts received) / (billable calls submitted). A 94% score is solid. An 82% score means you're losing 18% to disputes and you should either negotiate tighter terms or replace the buyer. No sentimentality — bad buyers get cut.

When to Re-Forecast

Your baseline model isn't permanent. Re-run the full projection when:

Traffic source changes. New traffic source = new cost per call = new answer rate (different caller demographics) = new model.

Buyer changes. New buyer = new payout = new billable threshold = new conversion rate = new model.

Vertical expansion. New vertical = entirely new model. Don't assume your HVAC numbers transfer to legal.

Significant volume change. 3x+ scale justifies a new model with scaling adjustments.

Quarterly check. Even if nothing changed, validate your assumptions against trailing 90-day actuals. Models drift.

I've seen publishers run the same projection model for eight months while their actual metrics shifted 20%. They couldn't figure out why margin kept shrinking. The model was stale.

Don't be that publisher.

The Pre-Scale Checklist

Before you increase ad spend, run through this:

• Do I have 30+ days of baseline data for this traffic source and buyer?
• Is my baseline True RPC at least 1.5x my cost per call?
• Have I run pessimistic, realistic, and optimistic scenarios?
• Does the pessimistic scenario stay profitable (or at worst, break-even)?
• Have I confirmed buyer capacity for increased volume? (See TCPA one-to-one consent requirements before scaling with new publishers.)
• Am I tracking answer rate and billable rate weekly to catch degradation?
• Do I have new tracking numbers ready to rotate if current ones get flagged?

If you can't check all seven, you're not ready to scale. Fix the gaps first.

For the ROI tracking that tells you whether your projections matched reality, track ROI by publisher covers the attribution layer.

Frequently Asked Questions

What's the difference between revenue forecasting and break-even analysis in pay-per-call?

Break-even analysis tells you the minimum performance needed to not lose money — it's backward-looking and defensive. Revenue forecasting is forward-looking: given your current metrics, what will your revenue be if you scale spend by 3x or 5x? Break-even asks 'can I survive?' Forecasting asks 'how much will I make?' Both matter, but publishers who only run break-even math never know when they're leaving money on the table by not scaling harder.

How do I calculate Revenue Per Call (RPC) for forecasting?

RPC = (Total Payout × Billable Rate × Conversion Rate). If your buyer pays $45 per qualified call, 68% of your calls meet the billable duration threshold, and 100% of billable calls actually get paid out (no disputes), your RPC is $45 × 0.68 × 1.0 = $30.60. Track this weekly. If your RPC drops below your cost per call, you're underwater. If it's 2x your cost per call, you have room to scale.

What answer rate should publishers use in revenue forecasts?

Use your trailing 14-day answer rate from your call tracking platform, not industry averages. Industry benchmarks say 55-65% for home services and 60-70% for legal, but your specific traffic sources, caller ID reputation, and buyer quality can push you 15 points in either direction. Pull your actual number. If you don't have 14 days of data, start with 55% and adjust weekly as real data comes in.

How do I forecast revenue when scaling to a new vertical?

You can't forecast precisely — you're guessing. But you can build a range. Research payout rates for the vertical (HVAC runs $35-65, PI auto runs $150-400, Medicare AEP runs $40-80). Assume a conservative billable rate (55-60% for a new vertical where you haven't optimized IVR yet). Assume answer rates 10% below your best-performing current vertical. Run the model at pessimistic, realistic, and optimistic scenarios. If the pessimistic case still beats your current vertical's RPC, it's worth testing.

Try VeloCalls for Your Vertical

Intelligent call tracking and 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 analysis, summarization). Per-minute pricing — Managed starts at 4¢/min, BYOC at 2¢/min, drops to 2¢/min and 0.5¢/min respectively at Enterprise tier.

See pricing → · Book a demo