People do not abandon surveys because they hate helping you.
They abandon surveys because, somewhere around the middle, it stops feeling like a favor and starts feeling like unpaid work. Too long. Too vague. Too many required fields. Too many questions that clearly do not apply.
The good news is that completion is mostly mechanical. Vendor data is consistent about where surveys break: drop-off rises roughly 17% once you cross ~12 questions or ~5 minutes, and short surveys can out-respond long ones by double digits (one trial saw 64% vs 51%). So most of these twelve tips are really one idea (reduce the cost of finishing) applied in different places.
1. Tell people why you are asking
A one-line reason changes the whole tone:
We are improving onboarding, and this should take under three minutes.
That beats dropping someone into question one with no context. Be accurate about the time, too, see tip 8.
2. Make the first question easy
The first question should create momentum, not friction. Do not open with a long text box, a salary field, or "tell us everything that happened." Start with something simple and relevant so the respondent commits before the harder questions arrive.
3. Cut the survey in half
Most surveys are too long because teams add every question they are curious about. Curiosity is not a good enough reason. If a question will not change a decision, remove it. Given the 12-question cliff, this is the single highest-leverage edit you can make.
4. Use one open-text question well
Open text is valuable but expensive for the respondent. One good open-ended question beats five lazy ones. My favorite:
What is the main reason for your score?
It turns a rating into something you can act on, without adding five more fields.
5. Ask at the right moment
Timing often matters more than copy. Ask about onboarding inside onboarding and the answer is fresh; ask a week later and you are asking someone to reconstruct a memory. A trigger that fires within the hour consistently outperforms a weekly batch email sent days after the fact. This is one of the reasons we built the Formaly SDK.
6. Use logic to skip irrelevant questions
Nothing makes a survey feel careless faster than irrelevant follow-ups. If someone has never used a feature, do not ask them to rate it. Branching is not a power-user luxury, it is how you keep a longer survey from feeling long for any individual respondent.
7. Use honest progress cues
People continue more readily when they know what is left, but be careful: research shows a slow-moving progress bar can actually lower completion by signaling length. Textual cues ("just a few questions left") and named stages often work better. Whatever you use, keep it honest.
8. Avoid fake urgency
Do not promise "only takes 30 seconds" if the survey takes five minutes. Respondents are doing you a favor; the moment they catch the lie, trust drops and so does effort.
9. Design for mobile
Assume the respondent is on a phone, distracted, and one bad interaction from leaving. Mobile-optimized surveys see about 10% higher completion than desktop-first ones, so avoid giant grids, tiny tap targets, and required text boxes.
10. Use reminders carefully
If you know who the respondent is, one well-timed reminder helps. More than that feels like pressure. Survey Invites make this manageable: send once, see who opened and completed, and follow up only with the people who actually need a nudge.
11. End cleanly
A good ending matters. Thank the respondent, tell them what happens next if it is relevant, and do not surprise them with another step after they think they are done.
12. Read the abandoned questions
If people consistently drop off at the same question, that question is the problem, too personal, badly worded, or brutal on mobile. Completion funnels turn that into something you can actually debug, instead of guessing.
The practical lesson
Completion rate is earned. A survey that is short, clear, well-timed, and respectful will usually beat a longer survey with better incentives.
The respondent experience is not a layer on top of the data, it is the thing that determines whether you get usable answers at all.