Metrics & Measurement
CX metrics worth tracking, and the ones that lie
Metrics are governance: they direct budgets and behavior. Lead with task success and retention; treat scores as thermometers, not steering wheels.
CX metrics are governance mechanisms. They direct attention, budgets, and tradeoffs whether you intend it or not. The metrics that matter are task-level and leading; a second tier of familiar scores needs careful handling; and any number tied to individual pay will be gamed.
Every metric you publish is an instruction to your teams, whether you meant it as one or not. That makes metric selection a governance decision, not a reporting chore. The metrics that matter most are task-level and leading: task completion and success rates, time on task, error rate, real time to resolution, customer lifetime value, and retention. They answer the two questions that count. Did people accomplish what they came to do, and did they stay?
Why it matters to the business
Familiar scores mislead in specific, documented ways. CEB research found the Customer Effort Score was 1.8 times more predictive of loyalty than CSAT and twice as predictive as NPS, and that 96% of customers who had high-effort experiences became more disloyal, against 9% after low-effort ones. Satisfaction can even mask defection: Bain's Reichheld found 60-80% of lost customers said they were satisfied or very satisfied shortly before leaving.
Misused metrics also decay. Reichheld himself warned that linking NPS to frontline bonuses breeds pleading and gaming, and proposed audited Earned Growth as the accounting-based check. Meanwhile the raw material is thinning: Qualtrics XM Institute data shows survey response rates fell 7-8 percentage points since 2021, and Gartner finds only 16% of customers strongly believe their feedback drives change.
How to use it
- Make task completion, task success, and customer-reported resolution your primary leading indicators.
- Measure real time to resolution, from the customer noticing the problem to a deployed fix, not internal ticket-close time.
- Use NPS only as a directional relationship signal, paired with qualitative research, and never average it across journey stages.
- Never tie handle time, features shipped, or survey scores to individual pay; validate loyalty claims against retention and repeat revenue instead.
- Before any number becomes a KPI, write down what behavior it will incentivize and how you would detect gaming.
Where teams get it wrong
Most organizations sit in the first two stages of measurement maturity, tracking business metrics only, or business metrics plus a few CX scores with complaints rationalized away, and mistake that for real measurement. Mature teams report honestly, chase root causes, and adjust KPIs when the KPIs themselves start producing bad experiences.
Ask your team
- Which of our CX metrics is anyone's bonus tied to, and how would we detect if it were being gamed?
- Can we show that customers who score us highly actually stay and spend more?
- What did we change last quarter because of customer feedback, and did we tell customers?
If a metric can be gamed, it will be.
Apply this
Reading about cx metrics worth tracking, and the ones that lie is one thing. Seeing where it applies in your journey is the useful part.