I started writing about customer experience in 2007.
I was managing a helpdesk in London, working longer hours than I should have, and trying to make sense of what I was learning on the job. I wrote to think. I published because I figured if I was confused about something, someone else probably was too.
That was 19 years ago. I’ve led support organizations ranging from 25 people to hundreds around the world. I’ve managed budgets up to $12M. I’ve built teams from scratch on three continents, integrated the wreckage of acquisitions, implemented ITIL frameworks in organizations that had never heard of them, and watched AI go from a contact centre punchline to a genuine operating reality.
This post isn’t a greatest-hits listicle. It’s an honest account of what I actually believe now versus what I believed when I started — and what changed my mind.
I thought metrics were the answer. They’re a starting point.
When I was a new manager, I was obsessed with measurement. If you could count it, you could improve it. Response times, resolution rates, CSAT scores, first call resolution — I tracked everything and reported upward with confidence.
It took me a few years to notice something uncomfortable: teams with excellent numbers sometimes delivered terrible experiences. And occasionally, teams with middling metrics were doing genuinely meaningful work for customers.
Metrics tell you what happened. They almost never tell you why. And without the why, you’re optimizing a number, not the thing the number was supposed to represent.
The shift for me was learning to use data as the start of a conversation, not the conclusion of one. When a CSAT score drops, that’s a question — not an answer. The answer is in the call recordings, the ticket notes, the agent who keeps flagging the same product issue that nobody upstream wants to hear about.
I still measure everything. I just don’t confuse the dashboard for the truth.
I thought great process was the solution to most problems. People are.
Early in my career I believed that if you built the right process, performance would follow. Get the workflow right, document the escalation path, define the SLA, train to the playbook. That was the job.
The organizations I ran with the best processes weren’t the ones that performed best. The ones that performed best had managers who actually knew their people — what motivated them, where they were struggling, when they were checking out before they handed in their notice.
You can’t process-engineer your way out of a people problem. And most performance problems are people problems, which are usually management problems in disguise.
The best investment I ever made in a team wasn’t a new ticketing system or a restructured routing queue. It was sitting beside an agent for two hours and watching them work — then having a real conversation afterward about what was getting in their way.
I thought AI was going to change everything. I was right, but not in the way I expected.
I wrote about AI in the contact centre back in 2018. I was cautiously optimistic, which in retrospect was still too early — most of what was being marketed as AI then was glorified keyword routing dressed up in a press release.
What I didn’t anticipate was how quickly the reality would arrive — and how different the actual impact would be from what everyone predicted. The automation wave was supposed to eliminate frontline support jobs. What it’s actually done is shift what those jobs require.
The best support agents I hire today are not the ones who are fastest at finding answers. The information retrieval part of the job is increasingly handled for them. The ones who thrive are the ones who can do the things AI still cannot: read a frustrated customer’s emotional state accurately, make a judgment call in an edge case that isn’t in any knowledge base, hold a conversation that leaves someone feeling heard even when the answer is no.
That has changed how I recruit, how I onboard, and what I measure. I’ll write more about this in future posts. But the short version: if your hiring criteria haven’t changed in the last three years, they need to.
I thought the most important relationship was with my team. It’s with the people above me.
This one took me the longest to learn, and I resisted it because it felt political.
Support leaders are chronically underfunded and under-heard in most organizations. We run critical operations that touch every customer, and we’re often the last function consulted when decisions are made about the product, the pricing, or the policies that directly affect what our teams have to deal with every day.
The reason for this, I eventually realized, wasn’t malice. It was translation failure.
Executives don’t speak in CSAT scores and average handle times. They speak in revenue, risk, and retention. The moment I learned to translate what my team did into those terms — to show what a one-point drop in CSAT was worth in terms of churn probability, to quantify what a 2-hour response time improvement meant for contract renewals — the dynamic changed.
The leaders who run the best support organizations aren’t necessarily the most technically skilled. They’re the ones who can tell the story of their team’s value in a language the business understands.
I thought I needed to have the answers. I needed to ask better questions.
Nobody tells you this when you get promoted to manager, but the job fundamentally changes. You are no longer paid for your individual expertise. You are paid for the judgment, development, and output of a team of people.
The instinct to demonstrate competence — to know the answer, to solve the problem, to have the plan — is actually counterproductive at a certain scale. I burned time and goodwill in my early leadership years by jumping to solutions before I’d properly understood problems. I hired people for their expertise and then didn’t use it.
The most useful thing I can do in most situations now is ask a question I don’t already know the answer to. “What’s in the way?” “If you were making this decision, what would you do?” “What am I missing?”
It sounds simple. It isn’t. Sitting with uncertainty long enough to actually understand something before you try to fix it runs against every instinct that gets people promoted into management in the first place.
What hasn’t changed
After all of that, here’s what I believe as much now as I did in 2007:
Service is a competitive advantage. Not a cost centre. Not a necessary evil. A genuine differentiator — when leadership treats it that way.
The people doing the work know what’s broken. They almost always do. Your job is to create conditions where they’ll tell you.
Consistency beats excellence. A team that reliably delivers good experiences is more valuable than one that occasionally delivers great ones but can’t sustain it.
This work matters. At the end of every support interaction is a person who needed help. Getting that right — reliably, at scale, with a team that’s proud of what they do — is genuinely hard. It’s worth doing well.
I built CX Master to share what I’ve learned along the way. Not as a theorist — as someone who’s been doing this job for nearly two decades and is still figuring it out.
If that’s useful to you, stick around.
I am an ITIL Expert and extremely passionate about customer service, customer experience, best practices and process improvement. I have led support, service, help desk and IT teams as well as quality and call center teams in Canada and the UK. I know how to motivate my teams to ensure that they are putting the customer first.
