Customer support is one of the most important parts of the customer experience, and one of the hardest to manage without data. Are customers waiting too long? Are problems actually solved? Is the team overloaded? Is the new AI assistant helping or annoying people?
Here are the 12 customer support KPIs that matter most, with formulas and practical tips.
Speed
1. First response time (FRT)
What: time between a customer’s message and the first human or meaningful response.
FRT = average(time of first response - time ticket created)
Use the median as well as the average; a few very late tickets distort averages. Measure per channel (email, chat, phone), since expectations differ.
2. Average resolution time
What: time from ticket creation to resolution.
Long resolution times often point to dependencies (waiting for warehouse, finance or suppliers) rather than slow agents.
3. Time to resolution by category
Break resolution time down by topic (returns, delivery, technical). This shows where processes, not people, need fixing.
Quality
4. Customer satisfaction score (CSAT)
What: share of customers rating their support experience positively.
CSAT = positive ratings / total ratings x 100%
Ask one simple question right after resolution. Read the comments; they are more valuable than the score.
5. First contact resolution (FCR)
What: share of issues solved in the first interaction, without follow-ups.
FCR = tickets resolved in first contact / total resolved tickets x 100%
High FCR means less effort for customers and less work for the team.
6. Reopen rate
What: share of tickets reopened after being marked resolved. A high rate means tickets are closed too early or answers do not really solve the problem.
7. Customer effort score (CES)
What: how easy it was to get the issue solved. Useful when customers have to repeat themselves or switch channels.
Volume and efficiency
8. Ticket volume (and contact rate)
Track total tickets, and also tickets per order (contact rate) for e-commerce:
Contact rate = tickets / orders x 100%
A rising contact rate signals problems in products, delivery or the website, even if the team keeps up.
9. Tickets by reason
The most actionable report: what are customers contacting you about? If “Where is my order?” is the top reason, better tracking emails and proactive updates may remove many tickets entirely.
10. Backlog
Open tickets older than your target response time. A growing backlog is an early warning.
AI and self-service
11. AI resolution rate
What: share of conversations fully resolved by an AI assistant without human handover.
AI resolution rate = AI-resolved conversations / total AI conversations x 100%
Always measure it together with CSAT on AI conversations. An AI that “resolves” conversations by making customers give up looks good on this metric and bad on satisfaction.
12. Handover rate and reasons
What: share of AI conversations handed to humans, and why (customer request, low confidence, out of scope, sensitive topic).
Handover reasons tell you what content or integrations to add next. See AI customer support for e-commerce.
KPI overview
| KPI | Measures | Watch out for |
|---|---|---|
| First response time | Speed | Auto-replies counted as responses |
| Resolution time | Speed | Waiting on other departments |
| CSAT | Satisfaction | Low response rates |
| FCR | Effectiveness | Definition consistency |
| Reopen rate | Quality | Tickets closed too early |
| CES | Effort | Survey fatigue |
| Contact rate | Underlying problems | Seasonal peaks |
| Tickets by reason | Root causes | Inconsistent tagging |
| Backlog | Capacity | Old tickets hidden in queues |
| AI resolution rate | Automation | Must be paired with CSAT |
| Handover reasons | AI improvement | Missing categories |
How AI changes support metrics
When you introduce AI (assistants, reply drafts, automatic tagging):
- First response time drops sharply for questions AI can answer
- Ticket volume for humans drops, but the remaining tickets are harder, so average human handling time may go up. That is expected
- Tagging becomes consistent when AI classifies tickets, making “tickets by reason” much more reliable
- New metrics appear: AI resolution rate, handover rate, AI CSAT
Set expectations with management before launch, so a higher average handling time is not misread as a drop in performance.
Building a simple support dashboard
Start with five numbers, reviewed weekly:
- Ticket volume and contact rate
- Median first response time
- CSAT
- Top 5 contact reasons
- Backlog
Add AI metrics when you introduce AI. Calculate the value of improvements with the ROI of AI automation guide.
Frequently asked questions
Which support KPI is the most important?
There is no single one. A combination of speed (first response time), quality (CSAT) and root causes (tickets by reason) gives the most complete picture.
Should agents be measured individually on these KPIs?
Carefully. Individual speed targets can push agents to close tickets too fast. Use KPIs mainly to improve processes and coaching.
How often should we review support KPIs?
Weekly for operational metrics, monthly for trends and root causes.
Want better support with less effort?
I help e-commerce teams set up support metrics, automatic ticket classification and AI assistants that answer with real order data. I am also building Conviro, an AI customer support platform for online stores. See my AI integration service or request a consultation.