Description
A short, fixed satisfaction survey (one labelled rating plus open 'why?' questions) shown continuously to random samples of users and tracked over time within each population.
Continuous satisfaction tracking listens to customers over time rather than proving the effect of a release. Google's Happiness Tracking Survey samples users, not page views, into weekly random buckets and shows a small non-blocking invitation; anyone invited is excluded for at least twelve weeks. Page one asks overall satisfaction on a fully labelled 7-point bipolar scale, then open questions on frustrations, wished-for capabilities and what people like best. Chapman and Rodden generalise this into a CSat programme: a defined population, one ordinal rating (5 or 7 points, labelled, with midpoint), one open 'why?', minimal demographics, a baseline period after which wording and sources are frozen, and reporting of proportions with confidence intervals and top-2-box. Microsoft's NSAT is a scoring variant: a 4-point scale scored as % Very Satisfied minus % dissatisfied, plus 100. Rigour comes from random sampling, frozen instruments and within-population comparison only.
Attributes
- Discipline
- Survey & Psychometrics
- Loss type
- Active
- Stage
- Post-launch
- Data
- Mixed
- Effort
- Months
Steps
- Define the active-user population, the channel (in-product, email to known customers, or panel with validated screener) and the target number of completes per period.
- Randomly assign users (not page views) to weekly buckets, show one bucket a non-blocking invitation, and exclude invitees from re-sampling for at least twelve weeks.
- Ask the overall satisfaction question first on a fully labelled 5- or 7-point scale, then one or two open questions ('why did you give that rating?', frustrations, best-liked aspects).
- Pre-test wording, run a baseline period to settle sources and scale, then freeze everything; if using NSAT, score % Very Satisfied minus (% Somewhat + % Very Dissatisfied) plus 100 (range 0-200).
- Analyse rating proportions and top-2-box with confidence intervals over time within each population, marking launches on the time series.
- Read or code all open answers split by positive and negative raters and frequency-sort themes to prioritise.
- Report sample characteristics, trends with CIs, cautious group differences, comment themes and recommendations; keep closed-loop customer follow-up separate.
Sources
- Measuring the User Experience (Albert & Tullis, 2022)
- Understanding Your Users (Baxter et al., 2015)
- Quantitative User Experience Research (Chapman & Rodden, 2023)
- Mail and Internet Surveys (Dillman, 2007)
- The Ultimate Question 2.0 (Reichheld & Markey, 2011)
- Data Empowerment (Williams, 2024)
- Customer Analytics for Dummies (Sauro, 2015)
Also known as
- Happiness Tracking Survey (HaTS)
- HaTS
- Customer satisfaction (CSat) tracking programme
- CSat survey
- Satisfaction tracking survey
- In-product satisfaction tracking survey
- Microsoft Net Satisfaction (NSAT)
- NSAT
- Change-aversion tracking
- Relationship NPS
- Transactional NPS
Categories
- Analytics & Behaviour
- Survey & Psychometrics