My first week at Ceridian, customers were waiting as long as seven hours on hold.
Not seven minutes. Seven hours.
Tens of thousands of calls came in every day across a support environment serving corporate payroll for some of the largest companies in the country. The system was an AS400 mainframe, technically deep, aging, and prone to outages. Users (payroll managers, CFOs, senior finance and HR executives) had to be trained and certified just to use it. Many were not. So they called us. And we did their payroll for them.
The net promoter score was -52%. (not a typo)
In my first week, the Houston Texans called. I was days into the job and they ripped into me for things I had no way of knowing yet. That was my orientation. That was the environment I had walked into.
Here is what I did not understand at the time: the people responsible for this environment believed they were doing it right.
Before you can redesign the operation, you have to understand the currency.
A few days in, I made a small mistake that taught me something large.
An escalation came through and I sent a note to one of my staff asking them to handle it. Professional. Direct. Polite. What I did not include was a thank-you.
The reaction told me everything.
I started paying closer attention to my inbox. It was overflowing. Every time someone helped someone else internally, there was a thank-you. Every time a customer called in and an agent saved their payroll run, a customer thank-you email circulated through the organization on a reply-all chain. People printed those emails and taped them to their cubicle walls.
This was not a culture of politeness. It was a signal system.
In a company where average tenure was measured in decades, where I had a manager with 58 years of service, where multiple people had worked together for 30 or 40 years without ever being in the same room, organizational rank was not determined by title. It was determined by the size of your thank-you pile. By how many customers had personally reached out to say you saved them. By how many favors you had banked with colleagues who needed your expertise.
I asked someone why they were reluctant to train a new person who had just come in.
They said: "Why should I tell them what I know?"
That sentence was the diagnosis.
The currency of a culture is not what is on the mission statement. It is what people actually trade.
Once I understood that, every behavior in the environment made sense. The veterans were not being difficult. They were being rational. Their status, their value, their safety within the organization was built on the depth and exclusivity of what they knew. Sharing that knowledge freely was not generosity. In their calculus, it was self-destructive.
The result was a closed system. New employees could not be trained. Veterans who needed help from each other brokered their expertise as favors. Customers waited hours on hold because the only people who could help them were protecting the very knowledge that could have reduced those wait times.
The -52% NPS was not a customer service problem. It was the logical output of a rational, self-reinforcing cultural system.
My first instinct was the obvious one. We needed a real training program. We needed to identify which parts of the work a new person could handle quickly, route that volume to new staff, and free the veterans for complex issues. Reasonable. Sensible. Standard.
I was met with complete derision.
Not just from my own team. From sales, who had all come up through support. From project managers and quality assurance, who had all come up through support. Every person who understood this environment told me I was absolutely nuts.
I tried to get existing managers to help build the training. Nobody shared their expertise. When year-end hit, hundreds of temporary workers arrived and I could not pull a single veteran resource to build anything. The plan collapsed.
I had tried to use the existing cultural infrastructure to do something the culture was specifically designed to prevent.
The redesign had to be built around the currency, not against it.
I needed a proof of concept that did not require veterans to give anything up. I found people who used to work in support, who had the expertise, but who had moved on to other functions. They were far enough removed that sharing their knowledge was not a threat to their current standing. I worked with them to build a training curriculum.
At the same time, I built skills-based routing: a system that broke incoming work into segments, matched complexity to capability, and created a clear pathway where new hires handled a specific, manageable category of calls with confidence, while complex issues went to veterans who owned that depth.
For the new hires, I implemented a quality program built around two frameworks I rely on heavily: Cialdini's principles of influence and Prochaska's stages of change. The structure was direct. At the start of every call, confirm what the customer's issue is. This signals that the customer is being heard, which drops the emotional temperature immediately, and it anchors the scope so the interaction cannot expand without everyone acknowledging it.
At the close, summarize what was done and ask one question: "Did that solve your problem?"
The customer says yes. When the survey arrives two days later with the same question, they answer consistently with what they already said out loud. You are not manipulating anything. You are giving customers the chance to confirm resolution in real time, so their survey response reflects the actual interaction instead of two days of accumulated frustration.
New hires had no institutional knowledge to protect. They learned the method cleanly.
About 90 days in, I called an all-hands meeting in Atlanta. That is where the veterans were. The 30-year, 40-year, 58-year people. The ones with the most printed thank-you notes on their cubicle walls.
I opened a PowerPoint. I said: "I want to show you something I am genuinely confused about, and I want your feedback. I normally do not compare locations this way, but I cannot explain this one."
I put up the net promoter scores and customer satisfaction numbers side by side.
The new hires, the ones who had been with the company for 90 days, had better scores than people who had been there for decades.
The room erupted.
They went after the data first. I had expected that. I walked through every number. The data held. Then they said the comparison was unfair. Then there were complaints to HR that I was making people feel bad. All of it exactly as expected.
A month later, I showed the same slide again. The numbers were better.
That is when the question changed.
"Why did they get special training that we didn't?"
That is not an objection. That is a request. The veterans who had spent months telling me I was wrong about everything were now asking to be included in a program they had previously called impossible.
I did not win that argument in a meeting. The cultural currency did it for me. The new hires had accumulated better customer satisfaction scores in 90 days than veterans had built over 30 years. In a culture where those scores were status, that was not just a data point. It was a direct challenge to rank, and an open invitation to compete.
One more thing about sequencing that most turnaround frameworks skip.
Before I could implement real performance management, I had to solve one problem first.
In any turnaround, roughly 10 to 15 percent of the workforce will not make the change. That is not a character assessment. It is a distribution. If you start managing those individuals out before you have a functioning pipeline for recruiting, training, and onboarding new talent, attrition outpaces your capacity to staff. You are no longer running a turnaround. You are running a collapse.
I built the training engine first. I solved the ability to bring in and certify new people. Only then did I implement performance management with the confidence that natural attrition would not exceed my capacity to replace it.
The sequence is not a preference. It is a constraint. In a distressed environment with an entrenched culture, you cannot manage the performance problem before you have solved the staffing problem.
The culture at Ceridian was not broken. It was rational, self-consistent, and functioning exactly as it had been designed, across 30 and 40 years of design.
The problem was that it had been designed, slowly and without anyone intending it, around the wrong object of value.
Once I understood the currency, everything else followed.
Next in the series · 90 for 90: How I Know a Turnaround Is Actually Complete