Your Contact Center Already Knows What’s Broken

We build contact center software, so I spend a lot of my week with the people who run contact centers. In a first meeting, one goal comes up more than any other, and it comes up early: they want fewer calls.

I get it. Every call costs money, and every minute an agent spends on the phone is a minute somebody has to budget for. For years, my pitch was the one they wanted to hear: shorter handle times, and a bot that answers the easy questions before they ever reach an agent.

I still think those things matter. But I’ve changed my mind about what the goal should be. Fewer calls is a fine result. As a target, though, it can quietly teach a company to stop listening to the people who are telling it, in detail and for free, what’s wrong.

Nobody Calls to Say Everything’s Fine

Think about the last time you contacted a company. Odds are you weren’t calling to pay them a compliment. Maybe a charge on your card didn’t make sense. Maybe the app logged you out for the third time that week and wouldn’t let you back in.

That’s true of most of the conversations that come through a contact center. Somewhere upstream, a product or a piece of copy didn’t do its job, and a customer had to stop what they were doing to sort it out.

The British management thinker John Seddon coined a term for this: failure demand. It’s the contact you only get because something wasn’t done, or wasn’t done right, the first time.

What gets me is the contrast. Companies spend serious money on surveys and research panels to find out what customers think. Meanwhile, customers are explaining exactly what went wrong, in their own words, all day long, to the contact center.

Most of those conversations are recorded. Very few are heard again by anyone who could change the cause.

We Built Tools to Make Complaints Go Away

I’ll own my industry’s part in this, my own company included.

Most contact centers live by average handle time and cost per contact. If they use bots, there’s also containment: the share of conversations that ended without reaching a person. Every one of those numbers tells you how quickly and cheaply a problem got closed. None of them tells you why it happened.

So, you can run a contact center that hits every target while the same issue generates calls for a year. The agent who takes the twentieth “where’s my refund?” call of the morning gets praised for keeping it under four minutes. Nobody asks why there were twenty.

Bots can make it worse. A chatbot that sends a frustrated customer around in circles until they give up counts as a contained conversation. On the report, that looks like a win. The customer still has the problem, and now they’re angrier. Because the bot “handled” it, nobody hears about it.

We sold a lot of software on those numbers. I’m not sure we always helped our customers learn much from them.

The Answers Are Already in Your Recordings

The odd part is that most companies already capture what they’d need to find the cause. They just capture it badly.

At the end of each call, the agent picks a reason from a dropdown list. That list was usually built years ago, by someone who has since moved on, for products that have since changed. The agent has a few seconds before the next call, so they pick whatever’s closest.

“General inquiry” and “Other” end up near the top. (If you run a contact center, you may have just winced.) That’s the official record of why customers get in touch, and it’s often the least reliable data in the building.

Quality teams, meanwhile, listen to a small sample of calls, mostly to score and coach agents. That’s useful work. It just isn’t built to find out why people called in the first place.

What has changed recently is that you don’t have to sample anymore. Transcription and AI analysis are now good enough to read every conversation, on every channel, and sort them by what the customer was trying to get done. So the tools are there. What’s usually missing is someone outside the contact center looking at what they turn up.

Your Agents Knew First

Agents hear the same complaint a dozen times a shift, long before

it turns up in a monthly report. They know which promo code fails at checkout and which line on the first bill confuses everyone. They know the exact words that calm someone whose refund is late, because they had to work those words out themselves.

Look at an agent’s desk, or the notes they keep open on a second screen, and you’ll find a map of what’s broken, with a workaround for each. Most companies never ask for that map.

Agents are scored on getting through calls, and when they do raise something, it usually stops at their supervisor. After the second or third suggestion disappears without a reply, most agents stop making them. I can’t say I blame them.

Get It Out of the Contact Center

None of this helps while it stays inside the support team. The people who can fix the cause usually sit in other teams, and many of them have never heard one of these calls. Some of what I’ve seen work is surprisingly low-tech.

The cheapest is to send people to listen. Ask your marketers and product managers to spend an hour a month on live calls, hearing whatever comes in rather than a few recordings someone chose for them. It’s hard to stay attached to a clever headline after you’ve heard ten people misread it.

Then swap the dropdown for the transcripts. A short weekly note with the top reasons people got in touch, quoted the way customers said them, will travel further in your company than any dashboard. Put a few of those quotes at the top. People who skim past a chart will stop and read a customer’s sentence.

Give agents a direct line, too. Something as simple as a “this keeps happening” flag, reviewed by someone who can act on it, tells agents their notes are worth something.

Tell the contact center what’s coming. When marketing plans a campaign or a pricing change, agents should hear about it before customers do. And when the calls start, you’ll know within a day whether people understood the offer.

The last one is probably the hardest: change what you celebrate. Handle time shows how efficiently you answered a question. The bigger win is a call type that disappears because someone fixed whatever was causing it.

What I Tell People Now

When contact center leaders ask me how to bring call volume down, I still show them the efficiency features. They work, and they matter. A good bot is a gift to someone who just wants to change a delivery address. But I also ask them a question: what are your customers calling about, and who outside your team has heard it?

If you’re looking at contact center platforms right now, ours included, put that question to every vendor. Any platform can tell you how fast you answered. Ask what it can tell you about why customers called.

Volume comes down when you fix the reasons people call. That’s the kind of “fewer calls” I’d rather be selling.