How to get survey respondents for market research
How to get survey respondents is only half the question. The other half is how to get respondents whose answers you can trust. This guide covers sample size basics, screening, attention checks and completion codes, how to spot low-quality answers fairly, and what a paid survey sample costs when you recruit through micro tasks.
Where survey respondents come from
| Source | Good for | Watch out for |
|---|---|---|
| Your customers and email list | Satisfaction, churn reasons, feature priorities | Only reaches people who already chose you |
| Social media and communities | Niche audiences, early validation | Self-selected enthusiasts; slow to reach numbers |
| Research panels | Demographic quotas, representative samples | Higher cost per complete; minimum order sizes on some |
| Micro-task platforms (e.g. Tasklify) | Fast, affordable general-population and device-owner samples, concept and message tests | Not a probability sample; needs good quality controls |
Micro-task respondents are best for questions where you need many real opinions quickly: which headline is clearest, what people expect a product to cost, how they describe a problem in their own words. For market sizing or published statistics, use a panel with proper quota sampling. Tasklify’s survey respondents page shows the setup.
Sample size basics
For a simple percentage (“what share of people prefer A?”), precision improves with the square root of the sample size, so quadrupling the sample halves the margin of error. For a simple random sample at 95% confidence, the worst-case margin is roughly:
| Completes | Margin of error (±) |
|---|---|
| 50 | 13.9 percentage points |
| 100 | 9.8 percentage points |
| 200 | 6.9 percentage points |
| 400 | 4.9 percentage points |
| 1,000 | 3.1 percentage points |
Three practical rules follow:
- Size for your smallest subgroup. If you want to compare iPhone and Android users, each group needs enough completes on its own.
- Treat convenience samples as directional. Crowdsourced respondents are not randomly drawn from the population, so the table is a best case, not a guarantee.
- Qualitative questions need fewer people. For open-ended “why” answers, 30–50 thoughtful responses often show the main themes.
Plan for 5–15% of responses to be removed in cleaning and add that to your target.
Screening respondents
Hard criteria: use targeting
Country, device and experience are set in the task itself: pick specific countries or Tier 1 only; Android, iOS or Desktop; a minimum approval rate (80–95%); a minimum number of completed tasks; or Plus-only. Respondents outside those filters never see the task.
Soft criteria: screen without giving the answer away
If you write “only for dog owners” in the title, some people will become dog owners. Instead, keep the task title neutral and ask the qualifying question inside the survey among neutral options (“Which of these pets, if any, live in your home?”).
Two-stage screener pattern
- Task 1, screener (2 min): 4–6 neutral questions. Everyone who completes it is paid. Qualified respondents see a unique invitation code on the last page.
- Task 2, main survey: the task states that a valid invitation code from the screener is required. Respondents paste it as a text answer; you check it against the codes you issued.
Everyone is paid fairly for the time they spent, nobody is rejected for not qualifying, and unique codes cannot be shared.
Attention checks and completion codes
Completion codes
Put a code on the last page of the survey and ask for it as a text answer proof. Most survey tools can show a unique value per response (for example the response ID), which is far better than one shared code. Also ask for the respondent’s Tasklify username inside the survey so you can match survey rows to submissions.
Attention checks
- Instructed item: “To show you are reading, select ‘Somewhat disagree’ for this row.” Use one or two, not five.
- Consistency pair: ask the same fact twice in different forms (age early, birth year late).
- Open-ended question: one question that needs a real sentence. Copy-pasted, irrelevant or identical answers stand out.
- Timing: record completion time; a common rule of thumb is to review anyone faster than about a third of the median.
Task template for a paid survey
Title: 10-minute survey about how you shop for groceries
Instructions: We are researching how people plan their weekly shopping. There are no right or wrong answers; we want your honest experience. The survey works on phone and desktop.
Steps: 1) Open the survey link. 2) Enter your Tasklify username on the first page. 3) Answer every question honestly. 4) Copy the code shown on the final page.
Proof note: Paste the completion code as a text answer. The survey contains attention checks; submissions that fail them, or that use a wrong or reused code, will be rejected. One response per person.
Avoiding low-quality answers
Most bad survey data is caused by the survey, not the respondents. Before blaming the sample, check these:
| Cause | Symptom | Fix |
|---|---|---|
| Survey too long for the reward | Speeding, straight-lining near the end | Cut questions or raise the reward; price by minute |
| Grids on mobile | Same column chosen for every row | Use single questions; test on a phone |
| Leading questions | Suspiciously positive answers | Neutral wording, balanced scales |
| Required open-ended questions everywhere | “good”, “ok”, “nothing” | One or two open questions, clearly optional elsewhere |
| Visible screening criteria | Everyone qualifies | Neutral title, hidden screener |
On the platform side, Tasklify flags duplicate text answers and duplicate files across submissions, and you can require a minimum approval rate. Always soft-launch: open 10–20 slots, look at the data, then add the rest. For the review workflow in detail, see crowdsourcing quality control.
If your survey collects personal data, keep it to what you need, explain how you use it, and make sure you have a lawful basis. The Acceptable Use Policy prohibits processing third parties’ personal data without one.
Survey respondent costs
You pay reward × respondents + a 12% service fee. Rewards are held until approval. With current suggested rewards:
| Study | Setup | Rewards | Fee | Total |
|---|---|---|---|---|
| Quick concept test | 100 × $0.35 (5 min) | $35 | $4.20 | $39.20 |
| Standard survey | 200 × $0.70 (10 min) | $140 | $16.80 | $156.80 |
| Screener | 500 × $0.15 (2 min) | $75 | $9 | $84 |
| Main survey, qualified only | 150 × $1 (15 min) | $150 | $18 | $168 |
The minimum top-up is $10; the first top-up includes a 10% bonus up to $5 for tasks. Details on the pricing page.
Frequently asked questions
How many survey respondents do I need?
It depends on how precise the results must be and which subgroups you want to compare. As a rough guide for a simple percentage, 100 responses give about ±10 percentage points, 400 about ±5 and 1,000 about ±3, assuming a random sample. Plan enough completes for your smallest subgroup.
What is a survey completion code?
A code shown on the last page of the survey that respondents paste into the task as proof. Unique codes per response let you match every submission to an actual completed survey and stop people from sharing one code.
Can I reject respondents who fail attention checks?
Yes, if the task says the survey contains attention checks and that failing them leads to rejection. Rejections must be tied to published rules, never to the opinions a respondent gave.
How do I screen survey respondents without them gaming the screener?
Use platform targeting for hard criteria like country and device. For behaviour or ownership, hide the qualifying question among neutral ones, do not reveal the criteria in the task title, and consider a short paid screener that issues unique codes to qualified respondents.
How much do survey respondents cost?
On Tasklify you pay the reward per respondent times the number of respondents plus a 12% service fee. A 10-minute survey at the suggested reward with 200 respondents costs a little over the rewards total because of the fee.