How it works
You ask a short question, such as "How satisfied are you with this answer?", with a rating from 1 to 5.
The formula: CSAT = responses rated 4 or 5 ÷ total responses × 100.
Out of 800 responses in a month, 680 ratings of 4 or 5 give a CSAT of 85%.
Two things to keep in mind when reading it:
- Some companies show an average rating (for example 4.3 out of 5) instead of a percentage. Both work, as long as you stick to the same convention over time.
- When few customers respond, the ones who do tend to be very happy or very unhappy. Read the response rate alongside the score.
Why it matters in e-commerce
In e-commerce, CSAT is measured at specific moments: after a conversation with customer service or an assistant, after delivery, after a return. It quickly flags what's going wrong (an off-topic answer, a delay, a badly described product). Broken down by question topic, it shows where in the shopper journey customers are let down, and so where to fix things first.
Real-life example
A home goods retailer sends a one-question survey at the end of each conversation. The overall score looks fine, but responses about delivery dates score far lower than responses about products. The team finds that delivery times on product pages are vague, rewrites them, and watches that topic's score over the next month.
Common confusion
CSAT and NPS (Net Promoter Score) often get mixed up. NPS asks, on a 0 to 10 scale, whether the customer would recommend the brand, and reflects the relationship as a whole. CSAT covers one specific interaction. A customer can be happy with a conversation and still unlikely to recommend the brand, or the other way around.