Customer service is not about your company. It is not about your products. It is not about your policies or your processes or your service catalogue.
Customer service is measured by your customers — and the only person who can tell you whether you’re doing it well is the person on the other end of the interaction.
That sounds obvious. Most organizations don’t build their operations that way. They build inward-facing operations — designed around what’s easy to staff, easy to manage, and easy to report on — and then wonder why their CSAT scores plateau and their customers stop coming back.
I’ve been building and running contact centre and technical support operations for nearly two decades. The organizations that get this right consistently do four things well. The ones that struggle are usually failing on at least two of them.
The Foundation: What You’re Actually Building
Before we get into the four pillars, it’s worth being clear about what a well-designed customer service operation actually achieves.
The goal is not to resolve tickets. It’s to create an environment that is reliable, repeatable, and predictable — where customers know what to expect and consistently get it. Customers don’t require perfection. They require consistency. When you deliver a consistently good experience, customers come back. When the experience is unpredictable — sometimes excellent, sometimes terrible, often somewhere in between — you erode trust regardless of the average quality.
Perseverance matters more than peaks. A team that reliably delivers a 7/10 experience will outperform one that occasionally hits 10/10 and regularly lands at 4/10.
The four pillars that make consistency possible are: Staff, Training, Tools, and Measurement.
Pillar 1: Staff — The Right Person at the Right Time
Staffing in customer service is not simply about having enough people. It’s about having the right people available at the right time with the right skills to handle the work that’s actually coming in.
This distinction matters more than most leaders appreciate. A senior engineer who can solve any technical problem in your portfolio is not an effective Tier 1 agent. They’ll resolve the call, but they’ll do it more slowly than a well-trained Tier 1 agent who handles that issue type daily — and you’re paying a premium for a resource that’s being used well below its capability. The opposite failure is equally common: routing complex technical escalations to agents who don’t have the skills to handle them, resulting in incorrect resolutions, longer handle times, and frustrated customers who have to call back.
The principle is skill-based routing in its most fundamental form: matching the work to the person best equipped to do it. Before you can do this well, you need to understand your contact volume — its composition, its timing, its complexity distribution. What percentage of your contacts are Tier 1 resolvable? What requires Tier 2? What genuinely needs your senior technical staff?
Answering these questions well requires measurement — which is why Measurement and Reporting is the fourth pillar and feeds back into every other decision you make. The staffing model you build today is only as good as the data you have about what your team actually handles.
On the staffing mathematics: the tool for determining how many agents you need, at which hours, against which volume targets, is Erlang C modelling. It’s not optional — it’s the only rigorous way to connect service level targets to headcount. Most organisations that are chronically understaffed at peak hours and overstaffed at quiet periods are simply not using the right staffing model.
Pillar 2: Training — Invest in the People You’re Going to Lose
Here is an uncomfortable truth about customer service and technical support: most of your best agents will leave the team. Not the company — the team. Your top performers in support are, by definition, people who know your products deeply, understand your customers, and communicate well. Those skills are valuable across every function in the business. Sales, product, marketing, customer success — they all want people who understand the customer at that level of detail.
My advice: encourage it. Actively.
When a support agent moves into product, they carry with them an intimate understanding of where customers struggle. When they move into sales, they understand the objections and the edge cases better than any other hire. Your support team is a training ground for the rest of the organization — and the organizations that recognize this build better companies than the ones that try to hoard their support talent.
What this means practically is that your training investment should be dual-purpose: it keeps your best agents longer (because growth opportunities extend tenure), and it makes them more valuable to the organization when they do move on.
Training in a support context has three layers:
Product and process training — the foundation. Every agent needs to understand your product well enough to handle the majority of contact types without escalation. This is your onboarding curriculum and it needs to be maintained as the product evolves.
Skills training — communication, conflict resolution, active listening, note-taking, and the ability to translate technical complexity into language that non-technical customers understand. These are teachable skills that most organizations under-invest in.
Tier advancement training — the progression that allows your best Tier 1 agents to develop toward Tier 2 capability. This is what enables skill-based routing to work over time: as your team develops, you can route more complex work internally rather than escalating.
The alternative to investing in training is high attrition, inconsistent quality, and a team that’s perpetually at Tier 1 capability regardless of how long agents have been with you. Training is not a cost. It’s the primary lever you have for building a team that improves over time rather than one that simply turns over.
Pillar 3: Tools — You Cannot Serve Customers Without the Right Infrastructure
The tools your team uses directly determine what’s possible in your operation. This isn’t a technology argument — it’s an operational one.
Ticketing and case management
Without a proper ticketing system, your team has no institutional memory. Every customer interaction starts from scratch. Agents have no visibility into what the customer has experienced before, what solutions have been tried, or what commitments have been made on the company’s behalf.
There are free ticketing systems that work adequately for small operations just getting started. The caution I’d offer is that migrating away from any ticketing system — especially one that’s been in use for years — is significantly harder than choosing the right one at the outset. The data that accumulates in your ticketing system becomes your operational history, your trend data, and your basis for measurement and reporting. Choosing a system with poor data export capabilities will cost you later.
CRM and sales integration
Your support team should have full visibility into what each customer has purchased, what their account status is, and what conversations they’ve had with sales and customer success. Without this, support agents are guessing — and guessing incorrectly costs you upsell opportunities and, more importantly, creates the situation where a customer has to explain their account history on every call.
The best support teams I’ve built have had this integration in place. It changes the nature of the conversation: instead of “can you give me your account number,” you’re opening with “I can see you’ve been with us for three years and you’re on our Enterprise plan — let me look at what’s happening.” That’s a fundamentally different customer experience, and it’s achievable with proper tooling.
Monitoring and diagnostic tools
Particularly for technical support operations, the ability to proactively identify issues — and to diagnose problems in a customer’s environment quickly — is the difference between a team that leads the conversation and one that reacts to it. How often does a customer contact you with a problem that turns out to be their ISP, their hardware, or their network rather than your product? With the right monitoring tools, your team can often identify and communicate this before the customer has finished explaining the symptom.
Workforce management
This is the category that ties staffing to reality. Workforce management tools — combined with Erlang C modelling — allow you to forecast demand, build shift patterns that match capacity to volume, and track adherence in real time. Without them, you’re scheduling based on instinct and adjusting after the fact. With them, you’re making staffing decisions based on data.
The tooling investment is not about buying the most expensive platform. It’s about ensuring that your team has what they need to do the job without working around gaps in their infrastructure.
Pillar 4: Measurement and Reporting — Plan, Do, Check, Act
The Plan-Do-Check-Act cycle is the operational heartbeat of a well-run support organization. Measurement is the “Check” — the mechanism that tells you whether what you’re doing is working, and where the gaps are.
The fundamental principle of measurement in customer service is one I’ve come back to consistently throughout my career: the only true judge of whether you’re doing a good job is the customer. Your internal metrics tell you about process adherence and operational efficiency. Customer satisfaction data tells you whether the output of all that process is actually serving the people it’s designed for.
KPIs in a support context need to operate at two levels:
Team-level metrics tell you whether the operation is performing against its commitments: SLA attainment, first contact resolution, response time, resolution time, contact volume by type and channel. These are the metrics that tell leadership whether the operation is healthy and flag when systemic problems are developing.
Individual-level metrics tell you where to focus coaching and development: output, quality scores, CSAT, escalation rate, reopened ticket rate. These enable you to identify your high performers (and understand what they’re doing differently), your developing agents (and what specific gaps to address), and the individuals who — after appropriate support — may not be the right fit.
The measurement function also tells you things about your product and your processes that aren’t visible anywhere else. A spike in contacts about a specific feature after a product update tells you something the product team needs to know. A growing category of “how do I” contacts tells you that something in your onboarding or documentation isn’t working. A cluster of escalations around a specific issue type tells you that Tier 1 needs additional training or that the issue is more complex than it should be.
This is where the Voice of the Customer function starts. The data flowing through your support operation is one of the most complete pictures of the customer experience that exists in your organization — if you’re capturing it, analysing it, and sharing it with the right people.
Beyond the Phone Call
The four pillars above apply regardless of channel — phone, email, chat, portal, or whatever combination your customers use to reach you. The underlying principles don’t change: you need the right people, trained appropriately, equipped with the right tools, and operating in an environment where measurement tells you what’s working and what isn’t.
What does change is the complexity of orchestrating those pillars across multiple channels simultaneously. Omnichannel support adds coordination requirements: hand-off discipline between channels, a unified view of the customer across interactions, and service level targets calibrated to each channel’s expectations. These are solvable problems — but they require the same rigorous approach to staffing, training, tools, and measurement that single-channel operations do.
The organizations that build this foundation properly are the ones that can scale. The ones that skip it — that assume growth will sort itself out — are the ones that discover, usually at the worst possible moment, that their operation wasn’t built for the volume or complexity it’s now handling.
Build the foundation. The rest follows from it.
Related reading:
- KPIs and the Importance of Measurements
- Erlang C and Contact Centre Scheduling
- Navigating Skill-Based Routing and Scheduling
- The CX Leader’s KPI Playbook — Free Download
Hutch Morzaria is a Director-level CX and Support Leadership professional with 19 years of experience building global support organizations across SaaS, Fintech, and enterprise technology. He has hired dozens of support and CX leaders across his career and holds ITIL Expert certification across V3 and V4.
