Custom Metrics Worth Building for a Call-Driven Business

Custom metrics are where a call-driven business’s reporting stops looking like an ecommerce dashboard with the labels changed. Ad platforms give you cost per lead and ROAS. Your CRM gives you deal counts. Neither one tells you what it costs to put a qualified prospect in front of a closer, or how much revenue each shown call is worth, or whether your setters are booking calls that actually close. Those numbers you have to build.

Below are the custom metrics worth building if you sell on calls: coaching, consulting, info products with a sales team, agencies, high-ticket services. Each comes with a formula, a worked example and the trap that makes it misleading if you’re careless. Numbers are illustrative.

Why default metrics fall short for call funnels

Ad platforms optimize toward events they can see, and the events they see best are clicks and form fills. Your business doesn’t make money on form fills. Between the lead and the payment there are at least three steps that decide everything: the booking, the show, and the close. Each one can vary wildly by campaign.

Default dashboards also tend to count events instead of people. If a lead submits your form twice, books, reschedules and books again, an event-based report might show two leads and two bookings for one human. When you divide spend by those inflated counts, your costs look better than they are.

So two principles apply to every metric below:

1. Count people, not events, wherever the question is about people. Leads, buyers and shown prospects should be unique.
2. Build metrics around the call stages, because that’s where campaigns actually differ.

Cost per shown call

Formula: Ad spend ÷ shown calls

If you track only one custom metric, make it this one. Cost per booked call is popular, but bookings are cheap to inflate. A campaign that books lots of tire-kickers will look great on cost per booking and terrible once you see who shows up.

Example: Campaign A spends $6,000 and books 60 calls, 33 of which show. Campaign B spends $6,000 and books 40 calls, 32 of which show. Cost per booked call says A is better ($100 vs $150). Cost per shown call says they’re nearly identical ($182 vs $188), and B is easier on your closers’ calendars.

Trap: make sure “shown” means the prospect attended, not that the call wasn’t cancelled. If reps mark outcomes inconsistently, this metric inherits their mess.

Cost per qualified call

Formula: Ad spend ÷ calls where the prospect met your qualification criteria

Show rate tells you whether people turn up. Qualification tells you whether they should have been booked. Define qualified narrowly (budget confirmed, decision-maker present, real problem you solve) and track it per campaign.

This is often the best early predictor of revenue, because it shows up within days rather than waiting for payments. It’s also a strong candidate for a conversion event you send back to Meta or Google. See pixel conditioning explained for why mid-funnel events often train better than purchases.

Trap: qualification is a judgment call. If closers decide it, they may mark lost deals unqualified to protect their close rate. Base it on observable criteria from the call, not the outcome.

A variant: cost per application

Formula: Ad spend ÷ completed applications

For application funnels, this sits between cost per lead and cost per booking. It’s useful because applications usually have qualifying questions, so you can build variants: cost per application with budget above a threshold, cost per application from business owners, and so on.

Trap: don’t compare cost per application across funnels with different application lengths. A three-question form and a twelve-question form aren’t measuring the same commitment.

Close rate on shown calls

Formula: Unique buyers ÷ shown calls

Not closes divided by booked calls, and not payments divided by anything. Shown calls are the denominator because a closer can’t close a no-show, and unique buyers are the numerator because a buyer on a payment plan is one sale, not three.

Trap: installments. If your sales metric counts payments, every rebill raises your close rate. Use a metric that counts each buyer once. In Callix, the Unique Sales metric counts sales once per prospect, and Callix Leads counts prospects rather than events, so a lead who submits twice is one lead. Building close rate from those avoids the double counting. How to attribute revenue when buyers pay in installments covers the payment-plan side in more detail.

Revenue per shown call

Formula: Cash collected (or first-payment revenue) ÷ shown calls

This metric combines close rate and deal size into one number that answers a very practical question: what is one more shown call worth to us right now?

Example: in September you had 140 shown calls and collected $210,000 in first payments. Revenue per shown call is $1,500. If your cost per shown call is $190, every shown call returns roughly eight times its cost. That’s the ratio you defend when someone wants to cut ad spend in a slow month.

Trap: timing. Revenue from September’s calls keeps landing into October. Compare periods over the same lag, or use a cohort view.

Setter-to-closer quality

Formula: Close rate on calls set by each setter

Most teams measure setters on bookings. Bookings are what setters control most, so it’s tempting. But a setter’s real job is to put good conversations on a closer’s calendar.

Build this by splitting close rate on shown calls by the setter who booked the call. Pair it with show rate by setter. A setter with 30% fewer bookings and double the close rate is your best setter, and the default leaderboard will tell you otherwise.

Trap: small samples. A setter with eight shown calls in a month can swing from 10% to 40% close rate on one deal. Look at rolling 60- or 90-day windows.

Time from booking to close

Formula: Median time between booked call and first payment

Median, not average, because a few deals that close after three months will drag an average around. This metric tells you how long your cash cycle really is, which affects how you judge new campaigns. If the median is 9 days, a campaign that’s two weeks old has shown you most of its revenue. If it’s 35 days, it hasn’t.

In Callix, Time Between Events measures how long leads take to get from one event to the next, like Booked Call to Onboarding Call, as a median or an average, and you can add it to any metric, report or table widget. Times read in days and hours, like 2d 4h.

Formula metrics: when one part needs its own filter

Some of the most useful custom metrics combine two numbers that need different scopes. Take cost per sale of your flagship program. The spend should be all spend on the campaigns you’re looking at, but the sales should be only flagship sales, not the low-ticket offer or the renewal.

If your reporting tool applies one filter to the whole formula, you can’t build this. Filter the widget to the flagship product and spend may disappear or be miscounted. Leave it unfiltered and the sales count includes everything.

The solution is a formula where each part can carry its own filter. In Callix, Create custom metric combines metrics with a formula, and the + on a formula metric limits that metric alone, for example Unique Sales limited to three products, while the rest of the formula stays on the widget’s own filters. So “Ad Spend ÷ Unique Sales (flagship products only)” works as a single column.

Other formulas that need this:

• Cost per qualified call for a single call type, such as closer calls only, while spend stays account-wide.
• Upsell rate: unique buyers of the back-end offer ÷ unique buyers of the front-end offer.
• Cost per sale by offer when multiple offers share the same campaigns.

A starter dashboard

If you’re building from scratch, this set covers most of the decisions a call-driven business makes weekly:

Metric	Formula	Decision it supports
Cost per shown call	Spend ÷ shown calls	Which campaigns to scale
Show rate	Shown ÷ booked	Confirmation and reminder process; lead quality
Close rate on shown	Unique buyers ÷ shown	Sales performance; offer-market fit
Revenue per shown call	First-payment revenue ÷ shown	How much you can pay for a call
Cost per qualified call	Spend ÷ qualified calls	Early read on new campaigns
Setter close rate	Close rate by setter	Setter coaching and comp
Booking-to-close time	Median days booked → paid	How long to wait before judging a campaign

How to keep custom metrics trustworthy

Custom metrics lose trust fast if two people get different answers. A few habits help:

• Write the definition next to the number. Hover text, a footnote or a glossary page. “Shown” and “qualified” need explicit definitions.
• Calculate totals from totals. The total cost per shown call across campaigns is total spend ÷ total shown calls, never an average of each campaign’s ratio. Averages of ratios overweight small campaigns.
• Review definitions quarterly. As offers and processes change, so should what counts as qualified.
• Keep the count of metrics small. If your dashboard has 40 custom metrics, nobody reads any of them. Seven that drive decisions beat forty that describe activity.

For guidance on presenting these to leadership, see building a marketing dashboard executives will actually read. The point of custom metrics isn’t more numbers. It’s the handful of numbers that match how your business actually makes money.