CSAT (Customer Satisfaction Score) is a post-interaction survey metric: it asks a customer to rate a specific experience — most often a support contact, an order delivery, or a return — on a defined scale, then reports the share of respondents who picked the top of the scale. The 5-point scale (1 = very dissatisfied, 5 = very satisfied) is the most common DTC implementation; 1–7 and 1–3 variants exist but show up less often.
The standard reporting form is top-box share (the percentage of respondents who picked the top box) or top-2-box share (the percentage who picked the top two), not the average of the 1–5 scores. Most survey tools default to displaying a 1–5 mean (“4.6 CSAT”), but Likert scales are ordinal — the gap from 4 to 5 isn’t guaranteed to equal the gap from 3 to 4 — so averaging them assumes spacing the scale doesn’t promise. “84% top-2-box CSAT” is the reporting form to standardize on.
CSAT and NPS measure different things. CSAT is transactional: how was this specific interaction. NPS is relational: how do you feel about the brand overall. A brand with high CSAT and low NPS has a recognizable operational pattern — support reps are competent and friendly, but the underlying product or category isn’t generating loyalty. Reporting whichever number is higher to the board obscures which one of those signals the business is actually getting.
CSAT is most useful as a per-touchpoint diagnostic — sliced by support agent, fulfillment lane, return reason, channel, or SKU — rather than as a single brand-level health number. Brands that also want a friction signal alongside satisfaction add CES (Customer Effort Score), which asks “how easy was it” rather than “how satisfied are you.”