Customer Satisfaction (CSAT) Research: A Practical Setup Guide¶

CSAT is one of the simplest survey formats to build and one of the easiest to run badly - inconsistent timing, wording that drifts over time, and results nobody looks at after the first report. Here's how to set it up as a real program rather than a one-off.

Keep the Core Question Fixed¶

A CSAT program's value comes from tracking the same measure consistently over time. Decide on your core satisfaction question's exact wording and scale before you launch, and don't tweak it later - see Comparing Results Across Survey Waves: A Methodology Guide for why wording consistency matters this much.

Pair the Score With a "Why"¶

A satisfaction score alone tells you the temperature, not the cause. Follow your core scale question with an open-ended "what's driving that score" question, and run it through Text Analytics (see How Text Analytics Works: From Open-Ends to Categories) to surface the actual reasons behind a low or high score.

Deciding When to Trigger the Survey¶

CSAT is often triggered by a specific event - after a purchase, after a support interaction, after onboarding - using a Hidden Field to carry context about which trigger point the respondent came from (see Passing Data into Your Survey with Hidden Fields), so you can segment results by trigger type later.

Interpreting the Score Responsibly¶

A single average score can hide real variation - see Interpreting NPS and Scale Question Results for reading the distribution behind the average, not just the topline number.

Segmenting Meaningfully¶

Cross-tab your CSAT results against whatever context you have - product line, customer tenure, support channel - to find where satisfaction actually varies rather than treating your customer base as one undifferentiated group. See Reading Cross-Tab Results: What to Look for Beyond the Percentages.

Building a Recurring Report¶

Set up your CSAT report against a Combined Survey view from the start if you're running this repeatedly (see Building a Report for a Recurring Tracking Study), so new response periods flow into the same tracking view without a rebuild each time.

FAQ¶

Should CSAT be a 5-point or 10-point scale?
Either works - the more important thing is picking one and keeping it consistent across every wave you run.

How soon after an interaction should I send a CSAT survey?
Sooner is generally better for recall accuracy - a survey sent days after an interaction tends to get less accurate reflection than one sent within a day or two.

Can I run CSAT continuously rather than in discrete waves?
Yes - a continuously running survey triggered by ongoing events works the same way as a periodic one; just be consistent about how you segment and report on time periods when comparing.

Should low scores trigger any follow-up action?
That's a business process decision outside the survey tool itself - some teams use the open-ended "why" response to flag specific accounts for direct follow-up, which is a manual process based on what the response contains.

For the tracking-study mechanics behind a recurring CSAT program, see Building a Report for a Recurring Tracking Study.