Customer experience is an operations problem before it is a marketing one. The moments that make or break a customer, the wait for a reply, the promise that gets missed, the refund that takes three emails, all happen inside processes the COO owns.

Marketing sets the promise. Operations keeps it. That is why customer experience (CX) belongs on the COO's desk, not just the CMO's.

The practical job is to convert a vague goal ("delight customers") into things you can measure, staff, and fix on a Monday morning. This guide walks through how to do that, section by section, with the trade-offs stated plainly and a concrete example in each.

Own CX as an operating system, not a campaign

What it means. A campaign has a start and an end. An operating system runs every day and gets tuned over time. Treating CX as an operating system means you define the customer journey, assign an owner to each stage, attach a metric to each owner, and review the numbers on a fixed cadence, forever. Strong vs weak. A weak COO delegates CX to the support team and checks in when complaints spike. A strong COO maps the whole journey, from first contact to renewal, and knows which team owns each handoff. When something breaks, they can point to the exact stage, the owner, and the metric that flagged it. How to do it. Start with customer journey mapping done as an operational exercise, not a workshop poster. For each stage, write down: who owns it, what the customer is trying to do, what usually goes wrong, and one number that tells you it is healthy. That table becomes your CX operating system. Example. A subscription software company mapped its journey and found the ugliest stage was not onboarding, as everyone assumed, but the first renewal. No single team owned it, so nobody was watching. Assigning an owner and a renewal-readiness metric turned an invisible leak into a tracked, improving number.

Measure the few metrics that drive decisions

What it means. You cannot improve what you do not watch, but watching everything is the same as watching nothing. A COO picks a small set of CX metrics that connect a customer feeling to an operational lever, then holds teams to them. Strong vs weak. A weak dashboard shows twenty vanity numbers and drives no action. A strong dashboard shows a handful, each with a clear owner and a "what we do if this moves" rule. The test for any metric is simple: if it drops, does someone know what to change?

The core CX metrics and what each is actually good for:

How to do it. Pick one relationship metric (NPS or churn), one transactional metric (CSAT or CES on a key interaction), and one operational metric (first contact resolution or response time). Three is often enough to start. See operations metrics that matter for how to keep a metric set honest and avoid dashboard bloat. Example. A retailer obsessed over NPS but could not explain why it moved. Adding CES on the returns flow revealed the culprit: returns took too many steps. NPS was the symptom; effort on returns was the cause. The fix showed up in the metric they could actually act on.

Remove friction from the journey, one step at a time

What it means. Most bad experiences are not caused by rude staff. They are caused by clunky processes: too many handoffs, unclear ownership, forms that ask for the same thing twice. Improving CX is mostly process optimization applied to customer-facing work. Strong vs weak. A weak team apologises for slow service and adds more people to absorb the load. A strong team maps the process, finds the step that causes the delay, and removes it, so the same headcount handles more with less pain for the customer. How to do it. Take one high-volume journey, a support ticket, an onboarding, a refund, and walk it end to end as if you were the customer. Count the steps, the handoffs, and the waits. Then attack the biggest wait first. One fixed bottleneck beats ten small tweaks. Example. A services firm found that new customers waited days for account setup because three teams touched the file in sequence, each with its own queue. Collapsing it to a single owner with a checklist cut the wait dramatically, and complaints about "slow start" faded on their own.

Make CX a cross-functional habit, not a department

What it means. The customer does not see your org chart. They see one company. But the experience is delivered by sales, support, product, billing, and IT, each optimising for its own goals. The COO's job is to align those teams so the seams do not show. Strong vs weak. In a weak setup, teams throw the customer over the wall: sales promises, support inherits the mess, product never hears about it. In a strong setup, the same voice-of-customer data reaches every team, and a shared CX goal shows up in each team's own targets. Building real cross-functional teams is how you stop the throw-over-the-wall pattern. How to do it. Create one regular forum where sales, support, and product review the same CX data and own joint fixes. Give them a shared metric so no team can win while the customer loses. Good stakeholder engagement across departments is what keeps this forum from becoming a blame session. Example. A company noticed the same complaint appearing in support tickets month after month, but support could only apologise because the root cause lived in the product. Once support and product shared a monthly review and a joint metric, the recurring issue got fixed at the source instead of being re-explained a hundred times.

Use technology to serve the process, not replace it

What it means. CRM systems, help-desk tools, and analytics platforms are enablers. They make a good process faster and a bad process fail faster. The COO chooses tools that fit the journey you have designed, not the other way around. Strong vs weak. A weak move is buying a shiny platform and hoping it fixes service. A strong move is fixing the process first, then choosing tools that support it, with clean data flowing between them so no team is guessing. Grounding decisions in data-driven operations keeps tool choices tied to evidence rather than sales demos. How to do it. Before any purchase, write down the specific journey step the tool must improve and the metric that will prove it worked. If you cannot name both, you are not ready to buy. Prioritise integration: a tool that does not share data with the rest of the stack creates new seams for the customer to fall through. Example. A firm rolled out automated chat to cut response time. Response time improved, but CSAT fell, because the bot could not handle the questions customers actually asked and made them repeat everything to a human afterward. The lesson was not "avoid automation" but "automate the step you understood, measured, and could hand off cleanly."

Equip and motivate the people who deliver it

What it means. Frontline staff are the experience for most customers. No process or tool matters if the person on the phone is undertrained, unclear on policy, or not empowered to solve the problem. Investing in people is employee engagement as a CX lever, not a soft benefit. Strong vs weak. A weak organisation scripts every interaction and punishes deviation, so staff protect themselves instead of helping the customer. A strong one trains people well, tells them the goal, and gives them room to fix things, then trusts the CSAT numbers to catch problems. How to do it. Train on the two things that move CX most: clear communication and problem-solving within known guardrails. Give frontline staff a sensible authority to resolve common issues without escalating. Feed real customer feedback back to them so they see the impact of their work. Example. A support team was measured only on call volume, so agents rushed customers off the line to hit their number, and satisfaction suffered. Shifting the incentive to first contact resolution changed the behaviour overnight: agents took the time to solve the problem once, and repeat contacts fell.

Run a review cadence that keeps CX improving

What it means. CX is never "done." Customer expectations rise, channels change, and processes drift. The COO installs a regular rhythm of review and refinement so the experience stays healthy without a crisis to force it. Strong vs weak. A weak organisation reacts to CX only when a complaint reaches the CEO. A strong one runs a quarterly review of the journey map, the metrics, and the open fixes, and treats CX as a standing agenda item, not a fire drill. Weaving CX targets into your operational excellence routine is what makes the cadence stick. How to do it. Set a fixed quarterly review with three questions: which metrics moved and why, which fixes shipped, and which stage of the journey is now the weakest link. Assign owners and dates to the next set of fixes before the meeting ends. Example. A company that reviewed CX only annually kept getting blindsided. Moving to a quarterly rhythm meant the weakest stage was always known and being worked on, and the "surprise" complaints that used to reach the CEO mostly disappeared.

Key takeaways

• Customer experience is an operations problem: marketing sets the promise, operations keeps it, and the COO owns the processes where promises are kept or broken.
• Treat CX as an operating system with an owner and a metric for every journey stage, not a campaign that starts and stops.
• Pick a small metric set, one relationship metric, one transactional, one operational, and make each one drive a decision.
• Most bad experiences come from clunky processes, not bad people; fix the biggest bottleneck before adding headcount or tools.
• Align sales, support, and product around shared CX data and a shared goal so the seams between teams never show to the customer.
• Choose technology to serve a process you have already designed, and run a quarterly review so CX keeps improving instead of drifting.

Frequently asked questions

Why does customer experience strategy belong to the COO and not just marketing? Marketing shapes the brand promise, but that promise is delivered through operational processes, support workflows, fulfilment, billing, and onboarding, that the COO owns. When an experience breaks, the cause is almost always a process, a handoff, or a resourcing gap on the operations side. The COO is the person who can actually change those things. Which CX metrics should a COO track first? Start with three: one relationship metric such as NPS or churn, one transactional metric such as CSAT or CES on a key interaction, and one operational metric such as first contact resolution or response time. Three well-chosen numbers that each drive a decision beat a dashboard of twenty that nobody acts on. How do I stop different departments from delivering an inconsistent experience? Give sales, support, and product one shared forum, the same voice-of-customer data, and a shared CX metric so no team can hit its own target while the customer loses. Consistency across channels also needs standard operating procedures and clear service-level agreements so every touchpoint sets the same expectation. Should I invest in CX technology or fix processes first? Fix the process first. Tools make a good process faster and a bad one fail faster, so buying a platform to paper over broken workflows usually creates new problems. Before any purchase, name the exact journey step the tool must improve and the metric that will prove it worked. How do I balance cost efficiency with a better customer experience? The two are less opposed than they look. Removing friction, cutting unnecessary handoffs and repeat contacts, often lowers cost and improves the experience at the same time. Prioritise the fixes that do both, use automation selectively on steps you fully understand, and measure the effect on an operational metric like first contact resolution. How often should CX strategy be reviewed? Run a standing quarterly review covering which metrics moved and why, which fixes shipped, and which journey stage is now the weakest link. Annual reviews leave you blindsided by drift; a quarterly rhythm keeps the weakest point always known and always being worked on.