Building Surveys That Get Completed

Why completion rate matters more than response count

Two thousand invitations sent and two hundred responses is one kind of story. Two thousand sent and a thousand started but only two hundred finished is a very different story — you designed something that lost 80% of the people who were already willing to help. Started-but-abandoned responses are signal that the survey itself is the problem, not the audience.

The good news: most drop-off has a small number of common causes, and they're all under your control.

1. Right-size the survey

The single biggest predictor of completion rate is how long the survey feels. Not how long it actually takes — how long the respondent thinks it'll take when they're on the third question and deciding whether to keep going.

• Target five minutes of real time. For most transportation surveys, that is about 8 to 12 questions if a few of them are text or map questions, more if they're radio or single-select.
• Cut every question that doesn't earn its place. For each one, ask: "what decision will this answer change?" If the answer is "none in particular, but it'd be nice to know," cut it.
• Beware of the "while we're at it" question. Every stakeholder who hears about your survey will want to add their pet question. Most of them shouldn't go in.

2. Question order: build commitment before you ask anything hard

There's a psychology principle behind survey order: people who have already invested a small effort are much more likely to keep going. That means the first two or three questions should be easy, low-stakes, and ideally a little bit interesting.

• Start with engaging, easy questions. "Which of these best describes how you usually travel to work?" is a great opener — one tap, no typing, and the respondent is now invested.
• Put the map question in the middle. Map questions are the highest-value thing TIPS lets you ask, but they take effort. Place them after the Traveler has answered a few easy questions and is committed.
• Save the demographic and "anything else?" questions for last. Demographics feel intrusive at the start and harmless at the end. Free-form "anything else?" boxes produce great qualitative data but are usually skipped early.

3. The profile-step trade-off

TIPS lets you require a Traveler Profile before the survey starts — first name, last name, email, ZIP. Each one of those is a knob, and each one is also a drop-off opportunity.

• Turn the profile step off for fully public, anonymous data collection (open community input, QR-on-a-poster surveys).
• Turn it on when you need to tie responses to identities or send verification emails — for example, a member survey of an MPO advisory committee.
• If you only need an email address (so the response counts as Verified in reports), turn off the name and ZIP fields. Every required field that doesn't matter loses you respondents.

For a deeper dive on what Verified means and why you might want it, see Three Ways to Invoke a Survey.

4. Design for mobile, because that's where the responses come from

For most transportation surveys, more than half of responses come from phones. For QR-code surveys posted in public space, it's closer to all of them. Design accordingly.

• Test on a phone before you send. TIPS has a Preview Email button on the Send Invitations page — send the invitation to yourself, open it on your phone, take the survey. Whatever annoys you will annoy your respondents tenfold.
• Keep question text short. Long question prose wraps to a wall of grey text on a phone screen. Aim for one or two sentences max per question.
• Use radio buttons and dropdowns instead of free text wherever possible. Typing on a phone is slow and respondents will abandon a long text question that they would have answered on a laptop.
• Give map questions their own page. A map question that shares a page with three other questions on a phone is unusable. On its own page, it's a great experience.

5. Page breaks: free completion-rate insurance

Page breaks do two important things in TIPS:

1. They reset the perceived length. A 12-question survey feels long on one page and reasonable in three pages of four questions each.
2. They save progress. TIPS saves the response at every page submit. A Traveler whose phone dies on page 3 can open the invitation link again the next day and pick up where they left off.

A good rule of thumb: a page break every 3 to 5 questions, and always before a map question.

6. Required fields: make every star earn its place

Every required field that the respondent doesn't want to fill in is a chance for them to close the tab. Be selective. Use these rules:

• If you can analyze the data without an answer to this question, don't require it.
• If the question is sensitive (income, household composition, disability status, immigration status), never require it. Optional sensitive questions get more honest answers than required ones.
• Required map questions should be Route Selectors or Rankers, not Route Creators — "draw your route" is too much to require, but "pick one of these three" is reasonable.

7. Wording: be neutral, be specific, be plain

• Avoid leading words. "How can we improve Route 47?" already tells the respondent that the agency thinks Route 47 needs improvement. Try "How can we change Route 47 to better serve you?"
• Avoid double-barreled questions. "Is the service frequent and reliable?" cannot be answered honestly if it's frequent but unreliable. Split it into two questions.
• Avoid jargon. "ADA paratransit eligibility" means something to you. To a respondent, it's a wall. "Door- to-door rides for people with disabilities" is better.
• Always offer a neutral middle. A 4-point scale forces respondents to fake a position. A 5- or 7-point scale with a real middle option produces more honest data.
• Offer "I don't know" / "Prefer not to say" on demographic and opinion questions. Forced answers are often guesses.

8. Show progress — and label the finish line

Survey takers want to know how much longer they have. TIPS shows a progress bar by default; leave it on. Two other small details that matter:

• The button on the last page should say Submit Survey, not "Continue". The Traveler should know they're about to finish.
• The Thank-You page is your only chance to say thank you. Customize the title and message in the Survey Builder. Generic "Thanks!" pages feel like the agency didn't really care.

9. Test the survey before you send it to anyone

Two cheap, high-value tests every survey should pass before launch:

• The Preview Email test. Use the Send Invitations page's Preview Email feature to send yourself the real invitation email. Open it on your phone. Click through. Take the survey. Whatever you found awkward, your respondents will find worse.
• The pilot list test. Build a small Mailing List of 10–25 friendly testers (coworkers, board members, a small subset of your real audience). Send to them first. Look at the completion rate, time to complete, and the text responses. Fix what's broken. Then send to the big list.

10. Timing: when to send, when to remind

• Best initial send window: Tuesday through Thursday, 9:30am to 11:00am respondent local time, for consumer audiences. Avoid Monday mornings (inbox cleanup) and Friday afternoons.
• Reminder timing: Two to four days after the original send. A well-timed reminder typically lifts overall completion by 10 to 25 percentage points.
• Open invitation window: Keep the survey open for one to two weeks. Past that, you mostly get Travelers who hate-completing it to make the reminders stop.

TIPS keeps the reminder workflow safe: on the Send Invitations page, the recipient list is broken into "new", "previously invited (no response)", and "already completed" buckets. The middle bucket — reminders — is checked by default; the "already completed" bucket is unchecked. You can't accidentally re-send to a Traveler whose response is already in your reports.

Common questions about survey completion rates

How long should a survey be to get a high completion rate?

Aim for under five minutes of real respondent time. In practice, that is roughly 8 to 12 well-designed questions for a typical transportation survey, depending on how many require typing or map drawing. Completion rates fall sharply past the 5-minute mark and tend to collapse past 10 minutes.

Should I require an email address before the Traveler can start the survey?

Only if you actually need to tie the response to a known person — for example, to send a verification email, prevent duplicate responses, or follow up later. Every required field on the pre-survey profile form reduces the completion rate. If you do not need an email, do not ask for one. TIPS lets you turn the profile step off, on, or on but with email optional.

What is the best order to put questions in?

Start with one or two easy, low-stakes questions to build commitment. Put map questions and longer text questions in the middle, once the Traveler is invested. Save demographic questions, sensitive questions, and free-form "anything else?" questions for the end. This pattern minimizes drop-off because respondents who reach the middle have already invested enough effort that they are likely to finish.

How do I design a survey that works well on a phone?

Assume the majority of responses will come from mobile. Keep questions short enough to fit on one screen without scrolling. Avoid wide tables and grid questions. Use the TIPS preview-email feature to send yourself a real invitation and take the survey on a phone before launching. The map questions in TIPS are touch-optimized but still benefit from being on their own page break.

Should I use page breaks?

Yes — for two reasons. First, page breaks give the Traveler a psychological reset and make a long survey feel shorter. Second, TIPS saves progress at every page break, so a Traveler whose phone runs out of battery can resume where they left off when they open the link again. A good rule of thumb is one page break every 3 to 5 questions, and always before a map question.

How do I avoid leading questions?

Read each question out loud and ask yourself: "could a respondent who disagrees with me answer this honestly?" Replace charged words ("improve", "fix", "fair") with neutral ones ("change", "alter", "evenly distributed"). Avoid double-barreled questions — anything with "and" in the middle that asks about two things at once. Always offer a neutral middle option on rating scales.

What is the best time to send a survey invitation?

For consumer audiences, mid-morning Tuesday through Thursday produces the highest open and completion rates. For agency and stakeholder lists, slightly later in the morning often works better. Avoid Monday mornings (inbox cleanup) and Friday afternoons (weekend mode). For reminders, two to four days after the original send is the sweet spot.

Should I send a reminder email?

Yes, almost always. A well-timed reminder can lift overall completion by 10 to 25 percent. TIPS makes this easy: when you send to a list you have already invited, the "Send reminder emails to N Travelers who have not responded" option is checked by default, and Travelers who have already completed are excluded.