Definition
A soft launch sends a survey or test to a small share of its eventual audience, in exactly the way the full launch will, and then pauses until the results have been read. The tutorial on pretesting a questionnaire makes it the last stage of a pretest: after think-aloud sessions and route checks, a small random part of the sample receives the real invitation, and the questionnaire is fixed and frozen only once that data has been read.
Two features separate it from a preview. The people are the real audience, chosen at random rather than picked because they are willing, and nobody is watching them answer. That is what lets it find problems the observed sessions cannot: how people behave on their own, with the real invitation, on their own devices.
What a soft launch reads
For a survey, the pretesting tutorial names three things to look for. Drop-off, read question by question, shows where people give up; a sharp fall at one question points at that question. Speeding and straight-lining show respondents who finished implausibly fast or answered every grid in the same column. And repeated "other, please specify" answers show a list that lacks an option people need. The tutorial also checks that the data arrives in the shape the analysis plan expects.
For an experiment, the reading is simpler. The A/B testing guide suggests a short soft launch to a small share of traffic before a test starts counting, to confirm that events arrive from real visitors.
In practice: surveys, commissioned fieldwork and tests
The survey launch checklist keeps a soft launch in the questionnaire group, next to the think-aloud sessions and route tests, and reminds the team to leave time for it before the first full invitation. The guide to survey and questionnaire design argues for testing with the audience before launch, because they find confusing questions and broken routing faster than internal review.
When an agency runs fieldwork, the soft launch becomes something the client asks for. The guide to commissioning a research agency suggests requesting one, with an early look at completion by question, the respondent mix against quotas and a sample of open answers, and agreeing in advance what happens if a check finds a problem.
Scale matters. For a single open question on a page, the tutorial on on-page polls and surveys asks for the question to be read aloud to someone outside the team; the fuller pretest, soft launch included, is for longer questionnaires.
Two decisions belong in the plan before the slice goes out: whether its answers will be kept if nothing changes, and how the slice is drawn, ideally at random from the same draw as the rest of the sample so fieldwork can carry on from it.
A veterinary group rewrites its client survey after a small first send
Hendory Veterinary Group planned a client satisfaction survey for 6,000 pet owners across 9 practices. Its pretest had used 5 think-aloud sessions, and the questionnaire looked ready. The soft launch went to 300 owners drawn at random from the list. Of 92 who started, 61 finished, and 22 of the 31 who left did so at one grid asking them to rate 12 aspects of their last visit. In the "other" box for the question on reasons for choosing a practice, 14 people wrote some version of "the vet who treated my previous pet". The grid was cut to 5 aspects, a new option for continuity of care was added, and both changes went into the change log. The full send went out a week later with the questionnaire frozen.
Related terms
A ScreenerThe short questionnaire that confirms each research candidate fits a study, asking about recent behaviour without revealing which answers qualify. decides who is eligible for research; a soft launch checks whether the instrument works once eligible people meet it. A soft launch can expose a Leading QuestionA survey or interview question whose wording contains or suggests its expected answer, producing measurement bias rather than evidence. only indirectly, for example when answers cluster suspiciously, which is why wording is checked in the think-aloud sessions first. An A/B TestControlled comparison of two variants. uses a soft launch as a technical check before it starts counting, not as a first result. A Usability TestA research session in which a person attempts a realistic task on a site, app or prototype while an observer notes where they struggle. watches a person attempt a task and hears why they struggle; a soft launch shows where many unobserved people struggle, without the why.
Common confusions
- A soft launch is a pretest. It is one part of a pretest, the part with the real audience and no observer. It cannot explain why people left a question; that needs the sessions.
- Colleagues can be the soft-launch group. They know what every question means. The value comes from a random slice of the real audience.
- Soft-launch results are an early read of the findings. For a test, they confirm that tracking works; for a survey, they check the questionnaire. Neither is a result to report.
- Any small product release is a soft launch. On this site the term means a survey or test released to a small share first. A quiet product or app release is a different practice with a different purpose.
