Most performance marketers drown in data but starve for insights. Your CallMatrix dashboard provides hundreds of data points, but scaling requires focus on the 'Vital Few.' These 10 KPIs are the pulse of your marketplace efficiency.
The difference between operators who scale smoothly and operators who plateau isn't the size of their dashboard—it's the discipline of checking the right five numbers every morning. Everything below is table stakes. Build them into a standing Monday report. When any one of them drifts more than 10% week-over-week, that's the one you investigate before you touch anything else.
1. Yield Per Call (YpC)
The ultimate north star. Total Revenue / Total Inbound Calls. This tells you the gross value of every 'Ring.' If your YpC is lower than your CPA, your business model is unsustainable.
Track YpC at the campaign level, not just the marketplace level. A blended YpC can hide 3 or 4 campaigns that are bleeding cash while another 3 or 4 carry the whole operation. Segment it by source (Google Ads, Meta, organic), by vertical, and by landing page. The first time you do this, expect to find one campaign doing 2 × the blended YpC — scale it. And one campaign doing 0.4 × — pause it.
2. Average RPC (Revenue Per Connected Call)
Unlike YpC, RPC only counts calls that reached a buyer. This helps you understand the quality of your buyers' bidding and the effectiveness of your auction engine.
The gap between YpC and RPC is your Connection Rate expressed as revenue. If YpC is 35 and RPC is 50, you're connecting 70% of inbound calls. If that ratio drops to 55%, your buyer capacity is the constraint—add overflow buyers, raise caps, or recruit more. If the ratio rises above 85%, your filters are too loose and you're connecting calls that shouldn't pass QA.
3. Bail Rate (The 'Dead Air' Metric)
The percentage of callers who hang up during the IVR or while being bridged. A Bail Rate over 5% indicates a technical latency issue or an overly complex IVR flow.
Break bail rate down by phase: IVR-bail (hung up during qualification) vs. bridge-bail (hung up during handoff to buyer). High IVR-bail means your questions are too long or too many. High bridge-bail means your routing is too slow or you're putting callers on hold while the auction clears. The remedies are different; the diagnosis is the same metric split two ways.
4. Average Bid Count per Call
How many buyers are competing for each call. Below 3 bidders per auction, your prices are essentially set by one or two buyers—that's a risky concentration. Above 8 bidders, your auction is healthy and you can push prices up by tightening acceptance criteria. The metric tells you whether your buyer roster is over- or under-saturated for your current traffic mix.
5. Agent Talk Time (ATT)
Average minutes of conversation per connected call on the buyer's side. Low ATT (under 90 seconds) usually means the buyer's agent is disqualifying calls quickly—good for the buyer's P&L but bad for yours because short calls mean low close rate. High ATT (over 8 minutes) means engaged conversations; your traffic quality is strong.
6. Disposition Accuracy
How often the buyer's reported outcome matches the call's actual signal profile. Buyers who systematically under-report 'Qualified' to avoid higher-tier pricing will show up here. Run a quarterly reconciliation: for your top 5 buyers, audit 50 random calls against their reported dispositions. Dispositional drift above 15% is a contract conversation.
7. DNI Matching Rate
The percentage of inbound calls that successfully matched back to a DNI session. Healthy is 95%+. Anything lower means your pool is undersized, your session cookie is getting wiped prematurely, or your selector is failing silently on some templates. DNI match rate is the canary for attribution integrity across your whole reporting stack.
8. Ping Latency (p50 / p95)
How fast the ping-post auction clears. Track both median and 95th percentile. Median above 400ms means the caller is waiting perceptibly. p95 above 1 second means a meaningful slice of your calls are bailing during the auction itself. Latency improvements pay back on both sides: lower bail rate and higher buyer acceptance rate.
9. Return Rate
Percentage of accepted calls that buyers credit back within the scrub window. Every return is lost revenue plus a reputation hit. Break return rate down by buyer and by root cause: 'wrong number' returns point to attribution issues; 'disqualified' returns point to IVR gaps; 'duplicate' returns point to dedup window mis-tuning. Target sub-5% blended return rate; above 10% and the marketplace is bleeding margin.
10. LTV Proxy
Projected long-term value of a caller, inferred from call signals before you have actual LTV data. Useful for campaigns where the sales cycle is too long to wait for closed-won feedback—insurance policies, legal retainers, enterprise software. Build a proxy from three inputs: buyer bid price, IVR disposition, and call duration on the buyer side. Rank new traffic sources by LTV proxy before committing ad spend scaling decisions.
Monitoring these 10 metrics daily allows you to move from reactive management to predictive scaling. In the modern marketplace, the fastest to react to data is the one who wins the auction.