Your CSAT survey says 91%. Your Amazon seller feedback says something else entirely. Both numbers are real. Ecommerce sellers are the only support teams juggling two different satisfaction scores at once, one you control, one the marketplace controls for you. This guide covers the formula, what counts as a good score, and five specific changes that move the number, not just five vague tips about “being more responsive.”
TL;DR
- CSAT = (positive responses ÷ total responses) x 100, tracked per interaction, not as a single lifetime score.
- 82% is the ecommerce median, 92%+ is top-quartile. Treat 85%+ as a solid working target.
- Ecommerce sellers deal with two CSAT layers: your own survey, and marketplace feedback scores you don’t control the wording of.
- Survey immediately after resolution, response rates drop from 25-35% at zero delay to under 10% after a day.
- The fastest CSAT gains come from order context, category-level fixes, and getting ahead of problems before the customer has to ask.
What is CSAT and how do you calculate it?
Customer Satisfaction Score (CSAT) measures how satisfied a customer was with one specific interaction, not your brand overall. You ask a simple question right after a ticket resolves, usually a 1-5 rating or a thumbs up/down, then calculate the percentage who rated it positively.
CSAT = (Positive responses / Total responses) x 100
A positive response is typically a 4 or 5 on a 5-point scale, or a thumbs up on binary scales. Run 500 resolved tickets through a CSAT survey, get 410 positive ratings out of 460 responses, and your CSAT is 89%, not 82%. The denominator is responses received, not tickets sent.
What is a good CSAT score for ecommerce?
82% is the ecommerce median, according to Gorgias benchmark data, with top-quartile merchants clearing 92% or higher. Treat 85%+ as a realistic working target, above median, genuinely competitive, and achievable without needing a top-quartile support operation to hit it.
A score sitting below 75% usually isn’t a training problem. It’s a structural one, slow response times, agents without order context, or a specific ticket category dragging the whole average down.
The two CSAT layers only ecommerce sellers deal with
A SaaS company has one satisfaction number. Ecommerce sellers have two, and they measure different things.
Your own CSAT survey measures the support interaction itself: did the agent solve the problem, how fast, how it felt. Amazon seller feedback and eBay feedback scores measure something broader, the entire order experience, product quality, shipping speed, packaging, everything, not just the ticket. A buyer can leave 1-star Amazon feedback over a late carrier delivery that had nothing to do with your support team, and that score still counts against your account.
Treat marketplace feedback as an external, noisier CSAT proxy, not a replacement for your own survey. It tells you what the market thinks of the full order experience. Your own CSAT tells you specifically whether support helped or hurt.
How often should you survey customers?
Immediately after the ticket closes, not an hour later, not the next day. Response rates run 25-35% at zero delay and fall under 10% by the 24-hour mark, the experience is fresh right after resolution and stale by the time a delayed survey lands.
Cap it at one survey per resolved ticket, and keep it to a single question. Response rates below 15% mean your sample size is too small to trust, if that’s happening, timing is almost always the first thing to fix before touching the question itself.
5 ways to improve your CSAT score
1. Cut the wait before the customer has to ask again
Response speed correlates directly with CSAT, a customer waiting three hours for a simple order-status answer rates that experience worse than one who waited three minutes, even if both got the same correct answer eventually. Route routine tickets like WISMO and tracking through AI Agent so the wait disappears entirely for the tickets that don’t need a human’s judgment.
2. Answer with order context, not a generic template
A response that opens with “can you provide your order number” reads as the company not knowing who the customer is. eDesk’s Customer View surfaces order history, tracking, and prior tickets the moment a conversation opens, so the first reply can answer the actual question instead of asking for information the system already has.
3. Fix the category dragging the average down, not the average itself
A blended CSAT score hides where the real damage is happening. Break scores out by ticket category in Insights & Reporting, and you’ll usually find one category, returns, damaged items, wrong-size exchanges, sitting well below the rest. Fix the templates and process for that specific category instead of trying to lift the whole average at once.
4. Get ahead of problems before the customer has to ask
A proactive shipping-delay message sent before the customer notices scores better than the fastest reactive reply to their complaint. Set up automated alerts for known delay triggers, carrier exceptions, stock-outs, so the seller reaches out first instead of waiting for the ticket.
5. Route unhappy customers to a human before they hit a public review
A 1 or 2-star CSAT rating should trigger an immediate internal follow-up, not just log to a spreadsheet. eDesk’s Reviews & Feedback feature sends selective review requests to customers who already rated their experience well, and flags low scorers for a human to reach out to directly before they leave public feedback instead of a private CSAT response.
How does Amazon seller feedback differ from CSAT?
Amazon feedback is a 1-5 star rating on the entire order experience, left voluntarily by the buyer, and it feeds directly into your Order Defect Rate if it’s 1 or 2 stars. CSAT is a survey you control, sent after a specific support interaction, measuring that interaction alone. eBay works similarly: feedback covers the whole transaction, and sellers can only leave positive feedback for buyers in return, so it’s a one-directional signal you can influence but not fully shape.
The practical difference: you can rewrite your CSAT survey question tomorrow. You can’t rewrite what a buyer decides to rate you on Amazon. Improve support quality and both numbers move, but they’ll never move in exact lockstep.
Key Takeaways
- CSAT = (positive responses / total responses) x 100, calculated per interaction, not as a blanket brand score.
- 82% ecommerce median, 92%+ top-quartile, 85%+ is a solid working target.
- Marketplace feedback (Amazon, eBay) is a noisier, external CSAT proxy covering the whole order, not a replacement for your own survey.
- Survey immediately after resolution, delay kills response rate faster than almost anything else.
- Fix the worst-performing category specifically, cutting response time, and catching unhappy customers before public feedback move the number fastest.
Your action plan:
- Calculate your current CSAT and check your survey timing, immediate or delayed.
- Break your CSAT out by ticket category to find which one is dragging the average down.
- Set up AI Agent on routine tickets so response time stops being the bottleneck.
- Build a proactive alert for your most common delay trigger (carrier exceptions, stock-outs).
- Set a rule to flag every 1-2 star CSAT response for a same-day human follow-up.
FAQ
What is a good CSAT score for ecommerce?
82% is the ecommerce median, according to Gorgias benchmark data, with top-quartile merchants reaching 92% or higher. 85%+ is a realistic, competitive target for most sellers.
How do I calculate CSAT?
Divide the number of positive responses (typically a 4 or 5 on a 5-point scale) by the total number of responses received, then multiply by 100. Use responses received as the denominator, not tickets sent.
How often should I survey customers?
Immediately after each resolved ticket. Response rates run 25-35% at zero delay and drop under 10% within 24 hours, so timing matters more than the survey question itself.
How is Amazon seller feedback different from CSAT?
Amazon feedback rates the entire order experience, product, shipping, packaging, not just support, and it factors into your Order Defect Rate. CSAT measures satisfaction with one specific support interaction that you control the timing and wording of.
Want CSAT tracked by category with order context built into every ticket automatically? Book a Free Demo and we’ll show you how it works.
