Most helpdesks track 20+ metrics. That's noise. These 6 are the ones that correlate directly with customer retention, repeat purchases, and support efficiency. Track these weekly, ignore the rest.
First Response Time is the single strongest predictor of customer satisfaction in e-commerce support. It measures the time between when a customer submits a ticket and when they receive the first human (or AI) response.
Why it matters: Customers who receive a response within 1 hour are 2.5× more likely to make a repeat purchase than those who wait 24+ hours. For live chat, the expectation is under 60 seconds.
How to improve it:
- Use AI to auto-respond to common ticket types instantly (WISMO, order status)
- Set up auto-acknowledgment: "We received your message and will respond within [X] hours"
- Prioritize channels by urgency — chat first, then social, then email
- Stagger agent shifts to cover peak hours (usually 9am–12pm and 5pm–8pm)
CSAT is a post-interaction survey, typically a simple "How would you rate your experience?" on a 1–5 scale. It's the most direct measure of whether your support is working.
Why it matters: CSAT below 85% usually indicates a systemic problem — not just individual bad interactions. Common causes: slow response times, incorrect information, impersonal responses, or unresolved issues.
How to improve it:
- Always resolve the issue in one interaction when possible (FCR)
- Use the customer's name and reference their specific order/product
- Follow up after complex issues to confirm resolution
- Train agents (or AI) on empathy and tone, not just accuracy
FCR measures the percentage of tickets resolved in a single interaction — no follow-ups, no escalations, no re-opens. It's a measure of both agent capability and process quality.
Why it matters: Every follow-up interaction doubles your cost per ticket and halves the customer's satisfaction. A customer who has to email three times about the same issue is likely a lost customer.
How to improve it:
- Give agents access to order data, tracking, and customer history in one view
- Empower agents to make decisions (refunds under $X, replacements, credits) without manager approval
- Write comprehensive SOPs that cover edge cases, not just the happy path
- Use AI that can pull real-time data to answer data-driven questions instantly
Average Resolution Time measures the total elapsed time from when a ticket is created to when it's marked resolved. Unlike FRT, this includes all back-and-forth, escalations, and waiting time.
Why it matters: Long resolution times indicate process bottlenecks — maybe agents lack authority to resolve, or information is scattered across tools, or escalation paths are unclear.
How to improve it:
- Map your escalation path — who handles what, and how fast should each level respond?
- Automate information gathering (AI can collect order details before handing to a human)
- Set SLAs for each ticket priority level and track adherence
- Identify your slowest ticket types and create dedicated SOPs for them
Cost Per Ticket is your total monthly support spend (agent salaries + tools + overhead) divided by the number of tickets resolved. It's your bottom-line efficiency metric.
Why it matters: If your cost per ticket is rising, you're becoming less efficient as you scale. If it's falling, your automation and processes are working. This metric tells you whether growth is sustainable.
How to improve it:
- Automate high-volume, low-complexity tickets (WISMO, order status, policy questions)
- Use self-service tools (knowledge base, FAQ) to deflect tickets before they're created
- Consolidate tools — every SaaS subscription in your support stack adds to cost per ticket
- Consider flat-fee support models that don't scale linearly with volume
CES measures how much effort a customer had to put in to get their issue resolved. It's typically a 1–7 scale question: "How easy was it to get your issue resolved?" Higher is better.
Why it matters: Research consistently shows that CES is a better predictor of customer loyalty than CSAT or NPS. Customers don't need to be delighted — they need to not be frustrated. Reducing effort is more impactful than exceeding expectations.
How to improve it:
- Don't make customers repeat information — pull up their order history before they ask
- Offer multiple channels but don't force channel switching ("email us to start a return" on a chat widget)
- Resolve issues in one interaction whenever possible
- Make self-service actually work — test your FAQ, chatbot, and tracking page as a customer would
Benchmarks by Store Size
What "good" looks like depends on your stage. Here are realistic benchmarks based on monthly order volume:
| Metric | Small (<500 orders/mo) | Mid (500–2,000) | Large (2,000+) |
|---|---|---|---|
| First Response Time | <4 hours | <1 hour | <15 minutes |
| CSAT | 85%+ | 90%+ | 92%+ |
| FCR | 70%+ | 80%+ | 85%+ |
| Avg Resolution Time | <24 hours | <8 hours | <4 hours |
| Cost Per Ticket | <$15 | <$8 | <$5 |
| CES | 5+ | 5.5+ | 6+ |
Daily fluctuations are noise. Weekly trends are signal. Set up a simple dashboard that shows each metric's weekly average, and review it every Monday. If something drops two weeks in a row, investigate.