First contact resolution (FCR) measures how often a customer's issue is resolved during the first support interaction. To use it well, define what counts as an interaction and how you will detect a repeat contact about the same issue.
A closed ticket is not necessarily a resolved problem. An order-status reply, for example, may answer the immediate question but still leave the customer unsure what to do about a missing package.
Define the measurement before setting a target
Use a documented formula:
FCR = issues resolved on the first interaction ÷ eligible issues measured × 100
Choose an observation window for repeat contacts and keep it consistent. Define whether an email exchange, chat session or phone call counts as one interaction. Record exclusions, such as duplicate tickets, and explain how conversations across channels are linked.
For a hypothetical cohort of 100 eligible issues, suppose 72 were resolved on the first interaction and did not reopen within the chosen window. FCR for that cohort is 72%. If some cases have not yet completed the observation window, keep them separate until they can be assessed fairly.
The result depends on those definitions and on the work in the queue. Compare similar issue types and channels before treating a change as an improvement. There is no single benchmark in this guide that every ecommerce team should adopt.
Find out why customers need to contact you again
Review a sample of repeat contacts and classify the reason. Useful categories include missing information, an action that was not completed, unclear instructions, a policy exception and a dependency on another team or carrier.
For each category, trace the customer's journey. Did the first response answer the question? Did the agent have the necessary order details and permission? Was the next step clear? This review produces more useful work than asking agents to close tickets faster.
Give agents information and authority
Keep product information, delivery policies and escalation routes in a shared knowledge base with a clear owner. Make it easy to find the relevant order or previous conversation without asking the customer to repeat everything.
Define which actions an agent may take, which require approval and how to obtain that approval. Include examples in new-agent onboarding, then practise ambiguous cases as well as straightforward ones.
Training should include a failed order lookup, an exception request and a case requiring another team. A useful handoff preserves the customer's context and makes ownership explicit.
Use self-service where it completes the task
Start with a recurring customer need, such as finding a policy or checking shipment progress. Test whether the customer can reach a useful answer and whether they have a clear route to a person when that answer is insufficient.
If you evaluate Chad or another support tool, use representative store workflows and confirm the outcome in the connected system. Do not count a displayed answer or attempted action as a completed resolution without checking what happened.
Report self-service completion separately from agent FCR unless your measurement definition explicitly combines them. Otherwise, moving easy contacts out of the agent queue can change its FCR even when the overall customer experience improves.
Ask for feedback and inspect reopened cases
Use a short follow-up question to understand whether the customer considers the issue resolved. Compare that response with ticket status and repeat-contact data. Review mismatches rather than assuming one signal is always correct.
Combine FCR with response time, resolution time and customer effort. A process that increases FCR by making customers wait excessively during one interaction still needs attention.
Improve one repeat-contact cause at a time
Choose one frequent cause, make a specific change and review a comparable set of cases after rollout. Record the observation window, volume and issue mix alongside the result. Avoid attributing every movement to the change when promotions, staffing or delivery disruptions also affected the queue.
Ask agents what still prevents them from completing work. Repeated friction may point to tooling, permissions or training gaps, and belongs in a broader review of agent workload and retention.
