The 10-Metric Playbook for Evaluating Call Center Customers Before They Cost You

Most call centers look good on paper. High volumes, steady traffic, and a solid rate per minute, what's not to love? Plenty, as it turns out. The carriers who get burned are the ones chasing volume without reading the room.

Behind every promising call center customer is a traffic profile that tells the real story: routing efficiency, failure rates, billing behavior, and whether their team can tell a SIP 503 from a 404.

The margin-killers don't announce themselves. They hide in your ASR stats, flood your NOC with tickets, and dispute invoices like it's their hobby.

This guide gives you the dual-lens framework, financial and operational. Let’s help you separate the genuinely profitable call center customers from the ones quietly costing you more than they're worth.

Why Traffic Profiles Matter More Than Volume?

Volume is vanity. Profitability is sanity.

A call center pushing 10 million minutes a month sounds impressive. That is until you notice the ASR is under 15%, the failed calls are burning your signaling infrastructure, and their support team emails at 2 a.m. with "urgent" tickets that lack a single timestamp.

The real cost of a bad call center customer isn't just in the margin per minute. It's in the routing inefficiencies that inflate costs, the support overhead that drains your NOC, and the billing disputes that tie up your finance team.

Volume without context is just noise. Traffic profiles add the signal.

The smart approach? A dual-lens evaluation: one lens on financial metrics, one on operational sustainability. Both matter. Neither alone tells the whole story.

Now let's get into the numbers because 10 metrics stand between a great customer and a very expensive headache.

The 10 Metrics That Define a Good vs. Bad Call Center Customer

Think of these 10 metrics as a report card split into two halves. The first five are your revenue reality check. The second five are about whether working with this customer will slowly drive your operations team to quit.

Part 1 - Profitability Metrics (Revenue Reality Check)

1. Average Call Duration (ACD)

ACD is a deceptively simple metric. Industry benchmarks suggest that an ACD of 4–5 minutes is acceptable, while anything above 6 minutes indicates excellent traffic quality.

A very short ACD, think under 60 seconds, is a red flag. It often signals spam traffic, autodialers churning through number lists, or poor audio quality causing early hang-ups.

Watch out especially for False Answer Supervision (FAS), a fraud pattern where calls are marked as answered before a real connection exists, artificially skewing your ACD and costing you revenue on calls that were never real.

For outbound call centers using predictive dialers, a somewhat lower ACD is expected, but it should still be consistent and explainable.

2. Answer-Seizure Ratio (ASR)

ASR is probably the metric you'll argue about most with call center customers. And for good reason. ASR directly determines how much of your infrastructure is working for revenue versus burning in silence.

As defined by the ITU E.411 standard, ASR measures the ratio of answered calls to total call attempts. Outbound call centers using auto-dialers typically produce an ASR below 20%; meaning the vast majority of attempts fail to connect.

A healthy wholesale VoIP route sits at 40–60%+. Anything below 20% means your infrastructure is mostly burning capacity for zero revenue.

The rule of thumb: high ASR = cleaner traffic = better margins. Low ASR = wasted capacity and a conversation you need to have with your customer.

If you're seeing sustained low ASR paired with high call volumes, you may also be carrying robocall or spam traffic. This is a separate regulatory risk in itself.

3. Post Dial Delay (PDD)

Nobody likes dead air. Most carriers consider anything under 7 seconds acceptable for PDD, a threshold also recognised by the ITU. Best-in-class VoIP teams target under 2–3 seconds.

High PDD directly erodes ASR. Impatient callers hang up before the ringback even starts, which shows up as a failed call; dead bandwidth, no revenue.

For call centers making thousands of attempts per hour, even a 2-second improvement in PDD can move the margin needle meaningfully.

If a customer consistently reports high PDD, the culprit is usually too many routing hops, slow carrier response times, or inefficient DNS resolution. Either way, it becomes your problem to debug.

4. Call Completion Rate & Failed Call Ratio

Every failed call generates signaling traffic. SIP messages flying back and forth, infrastructure processing a transaction that earns you nothing.

At scale, a high failed call ratio is like running a factory where half the machines are running but producing no output.

High failure rates also strain vendor relationships. If your upstream carriers start seeing degraded performance on routes you're using for this customer, it's your reputation on the line, not theirs.

5. Margin Per Minute / Effective Yield

This is the metric that actually pays your bills. Margin per minute, or effective yield, is what's left after you subtract routing costs, failure-related overhead, dispute credits, and support time from the revenue generated.

A customer paying a premium rate with a 15% ASR and weekly billing disputes might yield less per minute than a lower-rate customer with clean traffic and zero tickets.

Run the real numbers. Don't let rate cards fool you. If your routing isn't optimised, you may also be leaving margin on the table. Least Cost Routing is one lever worth pulling to improve effective yield across your customer base.

Revenue metrics tell you what you're earning. The next five tell you what it's actually costing you to earn it.

Part 2 - Quality-of-Life Metrics (The Hidden Cost Layer)

6. Ticket Volume per Traffic Unit

Some customers generate more tickets than traffic. If a call center is raising support requests at a rate disproportionate to their minutes, that's a signal. Either their traffic quality is poor, or they expect your NOC to solve problems that originate on their end.

High ticket frequency is an operational tax. Every ticket consumes engineer time, delays other customers, and quietly inflates your true cost-to-serve for this account.

7. Technical Competence of Customer

Can they read a SIP ladder? Do they know what a 486 Busy Here or 4XX series error means, or do they just forward every anomaly to you with "calls not working" as the subject line?

A technically competent call center customer is worth their weight in reduced support overhead. They can isolate issues on their side before escalating, provide pcaps when asked, and understand that not every failed call is a carrier problem.

The ones who can't, or won't, will consistently consume more of your team's time than their margin justifies.

8. Troubleshooting Collaboration Score

When something breaks, how does your customer show up? Do they provide call samples, timestamps, and destination prefixes? Or do they send a vague complaint and expect you to reverse-engineer the issue from CDRs?

Good troubleshooting partners make your NOC faster and more effective. Poor ones create friction, extend resolution times, and if they're aggressive about it, damage team morale. This is harder to quantify than ASR but very real in its impact.

9. Dispute Frequency & Billing Behavior

Billing disputes are a fact of life in telecom. Habitual disputes are a red flag. There's a difference between a customer who raises a legitimate discrepancy once a quarter and one who disputes invoices reflexively, without data, as a negotiating tactic.

The second type consumes finance resources, delays revenue recognition, and signals a relationship built on mistrust. Late payments and aggressive chargebacks compound the problem. If this is a pattern, factor it into your effective yield calculation because the real cost of this customer just got higher.

10. Escalation Behaviour & Communication Quality

How a customer escalates reveals a lot about the partnership. Professional escalation looks like: clear issue description, specific examples, a collaborative tone.

Reactive escalation looks like: executive CC chains at midnight, threats disguised as feedback, and "we'll move our traffic" as an opener.

The former builds long-term partnerships. The latter is exhausting to manage and often signals the relationship won't improve over time. Life is too short, and your NOCs too valuable, for customers who escalate emotionally rather than constructively.

Now that you have your 10 metrics, the next move is turning them into a decision, not just a feeling.

How to Combine These Metrics into a Decision Framework

Stop making customer decisions on gut feel. Use a simple quadrant:

• High profit + High QoL → Ideal customer. Invest in the relationship. Offer better rates, priority support, dedicated routes.
• High profit + Low QoL → Short-term gain, long-term risk. Set clear operational expectations or start planning your exit.
• Low profit + High QoL → Optimization opportunity. Work together to improve traffic quality or renegotiate terms.
• Low profit + Low QoL → Exit strategy required. These customers cost more than they contribute. Act accordingly.

The quadrant keeps decisions grounded in data rather than relationship inertia, which is the real enemy of margin in this industry. Before you build a scoring model, know what disqualifies a customer before they even get in the door.

Red Flags That Should Immediately Disqualify a Call Center

Some things aren't optimization problems, they're exit criteria.

Extremely low ASR with high volume: Auto-dialers running unchecked traffic through your network drain capacity and degrade route quality for everyone else.

Persistent disputes without data: If a customer disputes invoices without providing supporting CDRs or call logs, they're not engaging in good faith.

No technical ownership from their team: If every issue, regardless of origin, is escalated to your NOC, you're effectively subsidizing their operations team.

Excessive dependence on your infrastructure: A customer who has outsourced all technical diagnosis to you is a permanent support liability.

Recognizing these patterns early saves you from months of strained operations and eroding margins. Some customers are simply not the right fit. Once you know the red flags, build a scoring model that makes this evaluation systematic rather than situational.

How to Score and Rank Call Center Customers

Keep it simple. Score each of the 10 metrics from 0 to 10. Apply the following weights:

• Profitability Metrics (5 metrics): 60% of total score
• Quality-of-Life Metrics (5 metrics): 40% of total score

A customer scoring below 5 overall warrants a review conversation. Below 3 is an exit conversation. Above 7 is a relationship worth investing in.

Run this assessment at onboarding and repeat it quarterly. Traffic profiles change. A customer who was marginal six months ago may have improved significantly, or a top-tier account may be slowly degrading.

The model only works if you use it consistently. Numbers are the foundation. But the conclusion is the real insight.

Final Thoughts: Profit Is Not Just Revenue, It's Operational Efficiency

The best call center customers don't just send traffic. They send clean traffic, communicate like professionals, and treat your network as the shared asset it is.

Volume will always be tempting. But the carriers who build sustainable, high-margin businesses are the ones who've learned to read between the CDR lines, to see not just what a customer generates, but what they cost.

Evaluate the profile. Run the numbers. Trust the framework. The most profitable relationships in VoIP wholesale aren't the loudest or the largest. They're the ones where clean traffic meets professional collaboration, and everyone's margin reflects it.