What is first contact resolution?
First contact resolution (FCR) is the percentage of support tickets fully resolved in a single interaction, with no follow-up, reopen, or repeat contact from the customer. An FCR of 80% means eight of every ten customers get their issue completely handled the first time they reach out. A good first contact resolution rate for ecommerce sits at 75-85%; the cross-industry average is around 70%, and top performers reach 80-85%.
FCR is not the same as first response time, which only measures how fast the first reply goes out, and it is not CSAT, which measures how the customer felt. FCR measures completeness: did the person actually get what they came for, or did they have to come back? You can have a one-minute first response time and a terrible FCR if that fast reply is 'let me look into this and get back to you.'
Operators fixate on FCR because it sits upstream of almost everything else that matters. A resolved-first-time ticket costs less, scores higher on satisfaction, and never reappears in your queue as a repeat contact. A ticket that needs three rounds to close costs roughly three times as much to handle and leaves the customer more frustrated with each round.
First contact resolution (FCR) = tickets resolved on the first interaction divided by total tickets, over a period. 'First interaction' means one channel session: a single chat, a single email reply, or a single call. If the customer has to follow up or reopen the ticket, it does not count as resolved on first contact.
Why FCR is worth tracking
FCR is the closest thing support has to a leading indicator. Most metrics tell you what already happened; FCR tells you whether a ticket is about to generate more tickets. Industry research from SQM Group has tracked the same relationship for years: for roughly every one-point gain in FCR, CSAT moves up about a point, and call-center studies put customer satisfaction at 86% when an issue is resolved on first contact versus 42% when a repeat contact is required.
The cost story is just as direct. A repeat contact is a second ticket on the same problem, which means a second human touch, a second context-load, and usually a more annoyed customer who now leads with the history. If your blended cost per ticket is, say, six dollars, every avoidable repeat contact is six dollars you spent producing nothing but friction.
There is also a retention angle. Customers who have to chase a resolution churn at higher rates than customers who got a clean answer the first time, even when the eventual outcome is identical. The effort of the second contact is what sticks. That is why low-effort, one-and-done resolution shows up in nearly every modern CX framework, from CES to 'one-and-done' rate.
One caution before you anchor on the number: FCR is a means, not the goal. The goal is a resolved customer who does not have to think about the issue again. Treat FCR as the diagnostic that tells you whether your process is producing that outcome, and you will use it well. Treat it as a target to be hit at any cost and you will start gaming it, which the last section covers.
| What FCR drives | Resolved first contact | Needs a repeat contact |
|---|---|---|
| Customer satisfaction | ~86% satisfied | ~42% satisfied |
| Cost to serve | 1 ticket of handling | 2-3x the handling cost |
| Effort / friction | Low (one-and-done) | High (customer chases it) |
| Repeat-contact load | None | Adds to next day's queue |
| Churn risk | Lower | Elevated |
SQM Group's long-running call-center data places the cross-industry FCR average near 70%, with world-class teams in the 80-85% range. Ecommerce sits at the higher end of that spread because so much of the volume is structured, data-grounded questions an agent can answer outright.
FCR benchmarks by ticket type
A single blended FCR number hides the thing you actually need to know: where you are losing resolutions. Ecommerce FCR ranges from roughly 65% to 85% blended, but that average is the weighted result of categories that behave very differently. Order-status questions resolve on first contact almost every time; complex disputes and damaged-goods claims rarely do, because they need carrier investigation, photos, or a judgment call.
Use the table below as a diagnostic, not a scoreboard. Compare your FCR per category against the typical and strong columns. The gap tells you where to invest. If your WISMO FCR is 78%, something is broken in your data access, because order status should resolve north of 90%. If your dispute FCR is 60%, that is roughly normal and not worth chasing to 90%.
Shipping issues and disputes have structurally lower ceilings. Improving them usually means better intake, getting the order number, the item, and a photo in the first message, or faster carrier data access, not better-worded replies.
Your own ceiling is also a function of mix. A store that sells perishable food or fragile glassware will carry more damaged-goods and dispute volume, which pulls the blended number down no matter how good the team is. A store selling apparel staples with a simple returns policy will run higher on the same effort. Benchmark yourself against your category mix, not against a competitor whose ticket profile looks nothing like yours.
| Ticket type | Typical FCR | Strong FCR |
|---|---|---|
| Order status (WISMO) | 85-95% | 95%+ |
| Return eligibility questions | 75-88% | 88-93% |
| Product / pre-sale questions | 78-88% | 88-94% |
| Account / login issues | 72-85% | 86-92% |
| Billing and refund requests | 65-78% | 80-87% |
| Shipping issues / lost packages | 60-75% | 78-84% |
| Complex disputes / damaged goods | 50-68% | 70-80% |
| Overall blended FCR | 65-80% | 82-88% |
How to measure first contact resolution
FCR is simple to define and surprisingly easy to measure wrong. The number you report depends entirely on how you define 'resolved' and over what window you watch for the customer coming back. Pick one method, document it, and keep it stable so the trend means something.
The most practical approach for ecommerce is the reopen window: count a ticket as first-contact-resolved if the customer does not reopen it or start a new ticket on the same issue within 48-72 hours. Most modern help desks can report this automatically once you set the window. Three methods, from cleanest to most operational:
- 1Post-resolution survey. Send a one-question survey after closing: 'Was your issue fully resolved?' This is the cleanest signal because it asks the customer directly, but response rates are low, so it samples rather than measures every ticket.
- 2Reopen / repeat-contact tracking. Mark a ticket resolved on first contact if no follow-up or reopen lands within 48-72 hours. This captures every ticket automatically and is the default most teams run on. Watch out for new tickets about the same issue under a different thread.
- 3Agent-logged resolution. The agent tags whether they resolved the issue in that interaction. Cheap to implement but the least reliable, since agents over-report their own first-contact wins.
A 24-hour reopen window flatters your FCR; a lot of customers come back on day two or three. A 7-day window is stricter but starts catching unrelated new orders. For ecommerce, 48-72 hours is the standard compromise. Whatever you choose, segment FCR by ticket type, because the blended number alone never tells you where to act.
What drives FCR up
High FCR is mostly a function of having the right information and the right authority at the exact moment of first response. When an agent can see the order and is allowed to act on it, resolution happens in one touch. When either is missing, you get a deferral, and a deferral is an FCR failure even when it is polite.
Every 'let me check and circle back' reply converts one ticket into two and resets the customer's clock. Teams that cut deferrals, by giving agents data and authority, often see FCR jump several points without changing headcount or response time at all.
Information at the moment of response
- Live order data inline: the actual order, fulfillment status, tracking, and customer history visible in the ticket, so the agent answers from fact instead of going to look it up.
- A knowledge base your agents and AI can actually answer from, kept current, so nobody has to check with a manager before replying.
- Good intake: a ticket that arrives with order number, item, and a photo can close immediately; one that says only 'my order is wrong' guarantees a second message.
Authority to finish the job
- Agents empowered to issue refunds up to a set cap, start returns, and offer credits without an approval cycle resolve on first contact instead of opening an escalation chain.
- Clear, unambiguous return and refund policies, so the answer is a rule the agent applies, not a judgment they have to escalate.
- A culture that treats 'I'll look into this and get back to you' as a last resort, not a default. Either resolve it now or escalate it cleanly with full context.
What tanks FCR
Low FCR almost always traces back to a short list of systemic gaps rather than bad individual agents. The good news is that systemic problems are fixable once you name them. Here are the usual culprits, in rough order of how often they show up in ecommerce queues:
- Missing data at first response: agents lack order, carrier, or policy access, so they either stall or send a second message to get it.
- Vague policies: 'we evaluate each case individually' cannot be applied consistently, so agents escalate for approvals that should not need to exist.
- Authority gaps: agents who cannot take any action without a manager turn instant transactions into back-and-forth.
- Partial answers: replying 'tracking says delivered' without addressing 'but I never got it' creates a guaranteed follow-up.
- Time pressure: rushed agents on an understaffed queue give incomplete answers. Volume without coverage is one of the most reliable predictors of low FCR.
- Fragmented channels: a customer who starts on chat and follows up by email looks like two tickets, and the second agent lacks the first agent's context.
If FCR is low, resist the urge to coach agents harder first. Pull FCR by ticket type and by channel. A category-specific dip almost always means a data or policy gap, not a skill gap. You fix those with integrations and clearer rules, not more training.
How AI affects FCR
AI typically raises FCR on the categories it is built to handle, and the mechanism is specific: an AI agent connected to your store either answers from live data immediately or escalates with full context. There is no 'acknowledged but not resolved' middle state, and that middle state is what generates most repeat contacts in human queues. Studies of AI agent-assist consistently find meaningful FCR gains on the inquiry types AI is built to handle, because the model surfaces the right answer, or the right escalation, in the moment rather than after a delay.
For order status, AI FCR approaches 95%+ when it can read live order and tracking data, because the answer is always available in the moment. For return eligibility, AI applies a documented policy consistently without needing an approval, so its FCR holds up well. The trick is that AI is only as good as the data and rules you connect it to: a model with no order access is just a faster way to send a vague reply.
There are categories where AI FCR should be lower than human FCR, and that is the system working as designed. Complex disputes, damaged-goods negotiations, and issues spanning systems the AI is not connected to call for judgment. A well-scoped agent escalates those to a human with the full thread attached, which keeps AI FCR honest and lifts human FCR on the escalations, because the human starts with everything they need. For more on where that line sits, see our guide on [confidence thresholds for autonomous resolution](/blog/confidence-thresholds-autonomous-resolution).
| Category | Human FCR | AI FCR (well-configured) | Notes |
|---|---|---|---|
| Order status / WISMO | 88-94% | 94-98% | AI reads live data, more reliable |
| Return eligibility | 75-85% | 82-90% | AI applies policy consistently |
| Product questions | 78-87% | 80-88% | Depends on catalog data quality |
| Billing disputes | 65-78% | 60-72% | AI is conservative, escalates edge cases |
| Complex complaints | 55-68% | Escalates to human | Right call, do not automate judgment |
How to improve FCR in 30 days
You do not improve FCR by exhorting people to resolve more. You improve it by removing the specific reasons resolutions slip to a second contact. Run this as a focused sprint and you can move the blended number several points in a month, because most of the wins are structural rather than behavioral.
- 1Segment FCR by ticket type and channel. You cannot fix an average. Find the categories sitting well below their strong benchmark, that is your target list.
- 2Audit deferrals. Pull a sample of 'I'll get back to you' replies and ask, for each, what information or authority the agent was missing. Patterns appear fast.
- 3Close the data gaps. Connect order, tracking, and customer history into the ticket view so agents answer from fact. This single move usually fixes WISMO and order-status FCR.
- 4Rewrite the vague policies. Turn 'case by case' into explicit rules with caps (refund up to X, return within Y days). Rules are resolvable on first contact; judgments are not.
- 5Raise agent authority. Give agents a spending cap for refunds and credits so routine transactions close without an approval chain.
- 6Improve intake. Add the right fields or questions to your widget and forms so the first message arrives with order number, item, and a photo when relevant.
- 7Automate the data-grounded categories. Put an AI agent on order status, returns, and product questions so those resolve instantly and consistently, and free humans for the judgment cases.
- 8Re-measure and hold the window steady. Recheck FCR by category after two to three weeks against the same reopen window, and keep iterating on the laggards.
WISMO and order-status tickets are usually the largest single slice of ecommerce volume and the easiest FCR to fix, because the answer is pure data. Connecting live order data and automating that category is the highest-leverage move most stores can make in their first month.
Mistakes that game the metric instead of fixing it
FCR is easy to inflate without resolving anything, which is exactly why you should pair it with CSAT and repeat-contact rate. When teams are pushed on FCR in isolation, the metric goes up and the customer experience quietly gets worse. Watch for these.
The most common one is closing tickets prematurely so a reopen does not count against you, then letting the customer start a fresh thread. Your FCR looks great; your repeat-contact rate balloons. The two metrics only make sense read together, which is why most mature teams track them side by side along with CSAT.
- Premature closing: marking a ticket resolved before the customer confirms, so follow-ups land as new tickets that do not dent FCR.
- Channel-hopping blind spots: a chat resolution that the customer re-raises by email counts as first contact on both if your systems do not link them.
- Over-narrow definitions: defining 'first contact' so tightly (or the reopen window so short) that the number is technically true and practically meaningless.
- Optimizing FCR against CSAT: rushing to close to protect the metric, which trades a small FCR gain for a real satisfaction loss.
- Ignoring category mix: a quarter with more WISMO and fewer disputes will show higher FCR for free. Read the trend per category, not just the blend.
FCR is only trustworthy next to repeat-contact rate and CSAT. Rising FCR with rising repeat contacts means you are closing tickets, not solving problems. Rising FCR with steady or rising CSAT is the real thing.
How Bookbag raises FCR
Bookbag is built around the two things that drive first contact resolution: information and authority at the moment of response. It is an AI support agent for ecommerce that connects natively to Shopify, WooCommerce, and BigCommerce, so it reads live order, fulfillment, and tracking data and answers WISMO and order-status questions outright, the category where most stores leak FCR. It also tracks orders, processes returns and exchanges, and issues refunds within the caps you set, which is the authority part: it finishes the job instead of deferring it.
Crucially, Bookbag is an agent, not a script that deflects. It works from your knowledge plus live store data and resolves the ticket completely, or it escalates to a human with the full conversation and order context attached. That clean handoff is what keeps AI FCR honest on the judgment cases and lifts human FCR on the escalations, because your team never starts from zero. It runs across the website widget, email, WhatsApp, Instagram DM, and Messenger from one place, so a customer who switches channels does not become a second, context-less ticket.
Pricing is flat monthly plans with message-credit allowances and a spend cap you control, not per-resolution fees, so improving your resolution rate never produces a surprise bill. Most stores connect their store, import help docs, and drop in the widget in well under a day. If you are weighing options, our comparison pages lay out the tradeoffs honestly.
- Live order data inline, so order-status and WISMO tickets resolve on first contact, often above 95%.
- Takes real actions, returns, exchanges, refunds within your rules, so it does not defer resolvable tickets.
- Clean escalation with full context, which protects AI FCR and improves human FCR on the cases that need judgment.
- One agent across web, email, WhatsApp, Instagram, and Messenger, so channel-hopping does not fragment resolution.
Key takeaways
- A good ecommerce FCR is 75-85%; the cross-industry average is near 70% and top teams hit 80-85%.
- FCR predicts CSAT and cost: resolved-first-contact customers report ~86% satisfaction versus ~42% when they have to come back.
- FCR varies sharply by ticket type. WISMO should clear 90%+, while complex disputes are structurally lower, so measure per category.
- The usual FCR killers are missing order data, vague policies, and agents without authority to act, all structural, all fixable.
- Well-configured AI lifts FCR on data-grounded categories by answering completely or escalating cleanly, never deferring.
- Never read FCR alone. Pair it with repeat-contact rate and CSAT, or you will reward closing tickets instead of solving them.