A Medicare broker in Houston told me he was running a "profitable" campaign because his call volume hit 400/week during AEP. I asked about his revenue per call. Blank stare. I asked about his return rate. Longer pause. "I track calls," he said.
He didn't know he was losing $2,800/week.
His close rate had dropped from 22% to 11% over six weeks — buyers who couldn't handle the volume were rushing through calls and fumbling the close. His dashboard showed green because the calls kept coming. The money was walking out the back door.
I've made this exact mistake. Embarrassingly often, actually. Watching the call counter tick up like it meant something while the bank account told a different story.
Here's the thing most operators get wrong: call volume is a vanity metric. It feels good to see 500 calls/week. But if 180 of those go to voicemail, 90 convert, and 45 of those conversions cancel or no-show, you're running an expensive charity for your publishers.
This is the call analytics best practices list I wish someone had handed me three years ago. Not the 30-metric dashboard tour — just the numbers that actually matter for lead buyers, what good looks like, and what to do when they go red.
1. Answer Rate: Your First Funnel Check
Answer rate is the percentage of routed calls that actually get picked up by a human. Not by voicemail. Not by AMD false positives. A human being saying "hello."
What good looks like:
- Dedicated intake team (5+ agents): 85-90%
- Small buyer operation (1-2 people): 75-82%
- Solo operator running their own calls: 65-75%
Below 70%? Something is broken. Either buyer capacity is maxed, routing is sending calls to closed offices, or your caller IDs got spam-flagged. (For a deep dive on fixing low answer rates, see our answer rate troubleshooting guide.)
What to check daily:
- Answer rate by hour. If 12pm-2pm tanks, you've got a lunch coverage gap.
- Answer rate by buyer. One underperformer can drag your aggregate down 10+ points.
- Answer rate vs. yesterday, same time. Sudden drops signal capacity problems.
Here's my unpopular opinion: most "traffic quality" complaints are actually answer rate problems in disguise. Publishers get blamed for "bad leads" when the real issue is calls routing to buyers who aren't picking up. Check answer rate first. Blame traffic second.
2. Revenue Per Call (RPC): The Only Number That Pays Bills
RPC is total revenue generated divided by total calls received. Not connected calls. Not qualified calls. Total calls — because you're paying for all of them.
Formula:
RPC = Closed Revenue / Total Calls
If you bought 1,000 calls at $45 each ($45,000 spend) and closed $68,000 in revenue, your RPC is $68. That's a 1.51x return. Workable, but not great.
Vertical benchmarks (these vary — trust your own data over mine):
- HVAC emergency: $18-28 RPC
- Plumbing general: $14-22 RPC
- Roofing: $22-35 RPC (higher ticket, lower volume)
- Personal injury auto: $55-85 RPC
- Medicare AEP: $35-55 RPC
- Mass tort intake: $80-140 RPC
What to track daily:
- Rolling 7-day RPC by traffic source. If one source is running 40% below average, cut it before it eats your margin.
- RPC by day of week. Some verticals close better on Mondays (people calling after weekend emergencies). Others tank on Fridays (prospects "thinking about it" over the weekend and never calling back).
- RPC trend line. Flat or climbing = good. Declining for 3+ consecutive days = investigate immediately.
I've watched operators optimize for cost per call while ignoring RPC. They celebrate getting CPL down to $38 while RPC fell from $52 to $41. Congrats, you saved $7/call and lost $11/call in revenue. The spreadsheet looked better. The bank account looked worse.
(I did this for an entire quarter once. Built a beautiful Looker dashboard tracking CPL by source. Felt like a genius. Lost $14K because I wasn't tracking what actually mattered.)
This is why tracking ROI by publisher matters more than tracking volume.
3. Conversion Rate: Where Calls Become Money
Conversion rate is answered calls that result in your desired outcome — sale, booked appointment, qualified lead, whatever your buyers get paid for.
Formula:
Conversion Rate = Converted Calls / Answered Calls × 100
Benchmarks by vertical:
- HVAC (emergency repair): 35-50% (high urgency = high conversion)
- HVAC (maintenance): 18-28%
- Legal intake (PI): 12-22% (lots of non-qualifying cases)
- Medicare: 15-25% (depends heavily on agent skill)
- Home services general: 22-35%
The daily check:
- Conversion rate by buyer. Your best closer might convert at 38% while another buyer limps along at 14%. Route more volume to your closers. This sounds obvious. Most operators don't do it.
- Conversion rate by time of day. Morning callers often convert higher — they're starting their day with intent. 4pm callers are often "just getting quotes" and convert 30% lower.
- Conversion rate by call duration. Plot conversion rate against call length. You'll usually find a sweet spot — calls under 90 seconds rarely convert, calls over 8 minutes often do. This tells you minimum billable thresholds.
Red flag: If conversion rate drops while answer rate stays flat, your sales process is breaking down. Listen to call recordings. Are buyers rushing? Are they getting objections they can't handle? Is the IVR pre-qualification too loose, sending unqualified calls to buyers who then waste time on no-hope conversations?
And look — this is the frustrating part — conversion rate is where buyer skill matters more than anything you can control as an operator. You can optimize routing, filter junk traffic with ClickzProtect, tune your IVR until it's perfect. Doesn't matter. If the buyer can't close, conversion tanks. Know which buyers deserve your best calls. Fire the ones who don't. Harsh? Maybe. But I've watched too many operators coddle underperformers while their margins evaporate.
4. Billable Percentage: The Hidden Margin Killer
Billable percentage is the share of answered calls that meet your billing threshold — usually a duration cutoff like 90 seconds or 120 seconds.
Formula:
Billable % = (Calls Meeting Duration Threshold / Answered Calls) × 100
Benchmarks:
- 90-second threshold: 70-80% billable
- 120-second threshold: 60-70% billable
- 180-second threshold: 50-60% billable
Below your vertical's floor? You're paying publishers for calls that generate zero revenue.
Where calls drop early (and what to do):
- 15-45 seconds: IVR friction. Too many menu options. Confusing prompts. Fix the IVR. Check our IVR abandonment study for benchmarks.
- 45-90 seconds: Caller-buyer mismatch. The caller's problem doesn't match what the buyer offers. Tighten IVR qualification or route by service type.
- 80-90 seconds (just under threshold): Buyer handling. They're rushing through calls without building rapport. Train better or route away from that buyer.
One operator I worked with had 58% billable on a 90-second threshold. That's 42% of answered calls generating zero revenue while still costing payout. We found the IVR was playing 35 seconds of "please wait" music before transfer. Cut the music to 8 seconds. Billable jumped to 71% in a week. Sometimes the fix is embarrassingly simple.
5. Return Rate: The Repeat Caller Problem
Return rate is the percentage of calls from numbers that have already called your campaign before.
Formula:
Return Rate = (Repeat Caller Calls / Total Calls) × 100
Healthy range: 5-12%
Some returns are fine — a customer calling back to book after getting a quote, someone who got disconnected. But return rates above 15% signal problems.
What high return rate usually means:
- Callers didn't get resolution. They're trying again.
- Misroutes. The first buyer couldn't help, so they're hoping for someone else.
- Publishers recycling leads. That "fresh" call already went through another buyer's funnel last week.
- System errors. Calls dropping mid-conversation and callers redialing.
Daily action: Pull your top 20 repeat callers weekly. Listen to their original call and the return call. What happened? If it's "buyer said they'd call back and never did," you have a sales process problem. If it's "called from a different tracking number," your publisher is double-dipping.
Return rate gets ignored because it doesn't sound urgent. Honestly? I ignored it for two years. It was on my dashboard. I just... didn't look at it. Then I ran the math on a campaign that hit 19% returns and realized I'd overpaid by $8K in a single month. Now I check it daily. If you're seeing high return rates from specific sources, it may be duplicate lead detection catching recycled traffic.
Not glamorous to track. Matters anyway.
6. Average Call Duration: Quality Signal (When Read Right)
Average call duration tells you how long calls last before ending. But the number alone is almost meaningless — context matters.
What the segments tell you:
- 0-30 seconds: IVR abandonments, wrong numbers, pocket dials. This bucket should be under 15% of calls.
- 30-90 seconds: Quick qualifications that didn't fit, or buyers who rushed the call. Review a sample.
- 90-180 seconds: Short qualified calls. Some verticals (emergency services) close fast.
- 3-6 minutes: Standard qualified call length for most verticals.
- 8+ minutes: Either complex cases (legal, insurance) or buyers who ramble. Not always good.
The real insight: Distribution matters more than average. A campaign with 4:30 average could be all calls in the 4-5 minute range (consistent, healthy) or half at 45 seconds and half at 8+ minutes (bimodal, worth investigating).
Pull the distribution chart. If you see two peaks with a valley in between, figure out what's happening in each bucket. JustAnalytics can correlate web session behavior with downstream call patterns — useful if you want to know whether short calls come from mobile sessions, organic traffic, or specific landing pages.
7. Cost Per Acquisition (CPA): What Closed Revenue Actually Costs
CPA is your total spend (publisher payouts + platform fees + ad spend) divided by closed sales. This is where campaign budget optimization becomes critical — knowing your CPA ceiling lets you set hard caps.
Formula:
CPA = Total Campaign Costs / Number of Closed Sales
Why this beats CPL: Cost per lead ignores that 60% of leads don't close. CPA tells you what a closed customer actually cost. If you're paying $45/call, closing 15% of answered calls, and answering 80% of routed calls, your effective CPA is $375 ($45 ÷ 0.80 ÷ 0.15). Does that make sense for your average ticket?
Daily CPA tracking:
- By traffic source. Some sources deliver $280 CPA, others $520 CPA on identical payouts. The CPL is the same; the CPA tells you which actually works.
- By buyer. Your closer running 28% conversion has a $200 CPA. Your underperformer at 11% is running $515 CPA. Route accordingly. This is where real-time call dashboards pay for themselves.
- By day of week. Campaign-level CPA might look stable while Tuesday runs 40% below Thursday. Bid more on your good days.
8. Time to Answer: Operations Speed Check
Time to answer (TTA) is how long the phone rings before someone picks up. This is different from answer rate — TTA measures speed, not whether anyone answered at all.
Benchmarks:
- Under 15 seconds: Excellent. Dedicated intake doing their job.
- 15-25 seconds: Acceptable. Normal for small operations.
- 25-40 seconds: Concerning. Callers start abandoning around 20 seconds.
- 40+ seconds: Broken. You're losing 15-20% of calls to ring-out.
What slow TTA signals:
- Buyer understaffed. More calls than agents.
- Concurrency cap hit. Buyer can only handle 2 calls and you're routing 4.
- Wrong ring timeout settings. Your system waits 45 seconds before escalating; the caller hung up at 25.
Track TTA by buyer. If Buyer A averages 12 seconds and Buyer B averages 34 seconds, Buyer B is dragging your funnel. Either fix their staffing or deprioritize them in routing.
Honorable Mentions
First Call Resolution (FCR): Did the caller's issue get solved in one call? Matters for retention campaigns and support lines. Less critical for lead-gen.
Sentiment Score: AI-driven analysis of call tone. Useful for QA at scale, but don't obsess over it daily. VeloCalls offers sentiment analysis at 5¢/use — worth running on a sample, not every call. (I'll be honest: I still don't fully trust sentiment AI. It flags "frustrated" when someone's just talking fast. Use it as a filter to surface calls worth reviewing, not as ground truth.)
Lead-to-Appointment Rate: For verticals that book appointments before closing. Tracks how many leads actually show up. High appointment rate with low show rate means your confirmation process is broken.
Source Attribution Accuracy: What percentage of calls have complete sub-ID tracking? If you're running 30% unknown source, you can't optimize. Fix the attribution gaps before trusting any per-source metrics.
Quick Verdict
If you only track three metrics from this list, track these:
- Revenue Per Call — because it's the number that actually measures profitability
- Answer Rate — because everything downstream fails if nobody picks up
- Conversion Rate by Buyer — because your closers deserve more calls and your underperformers deserve fewer
Track daily. React in hours, not weeks.
Here's my controversial take: the campaigns that win aren't the ones with the best traffic. They're the ones run by operators who obsess over their numbers while competitors are "too busy" to check dashboards. Speed kills. Hesitation bleeds money.
Set your alerts: answer rate below 70%, RPC dropping 15%+ vs. 7-day average, conversion rate by buyer dropping 5+ points. When the alerts fire, stop what you're doing and investigate. A 15-minute diagnosis can save you $4,000 in wasted spend. I've done this math too many times on my own campaigns.
And if you're still manually checking dashboards twice a day... you're too slow. The market moved while you were in a meeting.
Frequently Asked Questions
What's a good revenue per call (RPC) benchmark for pay-per-call campaigns?
RPC varies wildly by vertical. HVAC emergency averages $18-28 RPC after accounting for close rates and ticket values. Legal intake runs $45-120 RPC depending on case type — personal injury auto sits around $65, while mass tort clears $90+. Medicare AEP campaigns hit $35-55 RPC during enrollment windows. If you're below your vertical's floor, either your close rate is suffering or your traffic quality needs auditing.
How often should lead buyers check their call analytics dashboards?
For campaigns running 100+ calls/day, check hourly during peak hours. For 50-100 calls/day, check 3-4 times during business hours. Set threshold alerts so you don't have to watch constantly — get notified when answer rate drops below 70%, conversion rate drops below 15%, or RPC falls 20% below your 7-day average. React in minutes, not days.
What's the difference between answer rate and conversion rate in call analytics?
Answer rate measures how many routed calls actually get picked up by a human — it's an operations metric. Conversion rate measures how many answered calls result in a sale, appointment, or qualified lead — it's a sales metric. You can have 90% answer rate with 8% conversion (great operations, weak sales handling) or 65% answer rate with 28% conversion (poor routing, but strong closers). Track both to know where your funnel leaks.
What causes return rate to spike in pay-per-call campaigns?
Return rate — repeat callers who've already been through your funnel — spikes from three sources: callers who didn't get resolution on the first call and are trying again, misrouted calls where the buyer couldn't help and the caller is retrying for a different agent, or publishers recycling leads they've already sold elsewhere. If return rate jumps above 12-15%, audit your call recordings to figure out which pattern you're seeing.
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.