Collect feedback without annoying your users
Survey Frequency & Sampling Controls
Balance data collection with user experience. Control exactly how often visitors see your feedback surveys with four frequency modes and adjustable sampling rates. Avoid survey fatigue while still collecting statistically meaningful data.
Why Frequency Control Matters
Too many surveys = annoyed users who leave your site. Too few = missed insights. Getting the balance right protects both UX and data quality.
Survey Fatigue Is Real
Users who see too many surveys stop responding—or worse, leave your site entirely. The cost of over-surveying is higher than under-surveying.
Data Quality Suffers
Fatigued users give rushed, lower-quality responses. You get quantity but not quality—and the data skews negative from annoyance.
Trust & Perception
Respectful feedback collection builds trust. Aggressive surveying damages your brand perception—exactly what you're trying to measure.
Four Frequency Modes
Choose how often any single user can see your feedback survey.
Show once ever per visitor
After a user sees or responds to the survey, they never see it again. Uses browser storage to remember.
Best for:
- • One-time feedback (e.g., onboarding experience)
- • High-value pages where you only need to ask once
- • Sites with frequent return visitors
Show once per browsing session
Each time a user returns (new session), they may see the survey again—but only once per visit.
Best for:
- • General website feedback collection
- • Sites where user experience varies by visit
- • Combining with sampling for high-traffic pages
Wait X days between surveys
After responding or dismissing, the user won't see the survey again for your specified cooldown period.
7d
Light
14d
Recommended
30d
Conservative
90d
Minimal
Best for:
- • Tracking changes over time with repeat visitors
- • Periodic pulse checks without over-surveying
- • Balancing data freshness with respect for users
No frequency limits
Survey can appear on every qualifying page view. Only use with low sampling rates.
Warning:
“Always” without sampling creates survey fatigue. Combine with 5-10% sampling on high-traffic pages, or use only for single-page apps where users rarely revisit the same page.
Sampling Rate (0-100%)
Show your survey to only a percentage of eligible visitors. Essential for high-traffic sites.
How Sampling Works
Even when all targeting and frequency rules pass, sampling determines whether the survey actually appears.
- 100% — Show to all eligible visitors (no sampling)
- 25% — Show to 1 in 4 eligible visitors
- 10% — Show to 1 in 10 eligible visitors
- 5% — Show to 1 in 20 eligible visitors
Example: 10% sampling
1 of 10 eligible visitors sees survey
When to use lower sampling (5-25%)
- •High-traffic pages (10,000+ daily visitors)
- •Pages where conversion is critical (checkout, signup)
- •When you already have enough responses for statistical significance
- •Testing a new survey before full rollout
When to use higher sampling (50-100%)
- •Low-traffic pages needing more data
- •Time-sensitive feedback collection
- •Pages with short average session duration
- •When response rate is already low
Recommended Settings by Traffic Level
Use these as starting points. Adjust based on your response rates and user feedback.
Frequency
Once per session
Sampling
100%
Expected responses
10-50/day
Rationale: You need every response you can get. Full sampling is fine because users won't see it often anyway. Focus on maximizing data collection.
Frequency
Cooldown (14 days)
Sampling
50-100%
Expected responses
50-200/day
Rationale: Balance is key. Cooldowns ensure return visitors aren't over-surveyed. Start at 100% sampling and reduce if you're getting more responses than you need.
Frequency
Cooldown (30 days)
Sampling
10-25%
Expected responses
100-500/day
Rationale: You have plenty of traffic. Low sampling protects UX while still generating statistically significant data. Longer cooldowns prevent repeat surveying.
Frequency
Once or Cooldown (90 days)
Sampling
5-10%
Expected responses
500-1000/day
Rationale: Be aggressive about protecting UX. Even 5% sampling gives you thousands of responses. Use “once” for one-time surveys or long cooldowns for ongoing collection.
How to Avoid Over-Surveying
Signs you're surveying too much, and how to fix it.
Response rate dropping over time
What it means: Users are tuning out your surveys
Fix: Increase cooldown period or reduce sampling rate
Angry feedback about the survey itself
What it means: Users are explicitly telling you it's too much
Fix: Reduce frequency immediately. Add longer cooldown.
Scores trending negative without site changes
What it means: Annoyance is contaminating your feedback data
Fix: Lower sampling. The data you're getting isn't accurate.
Multiple surveys showing per session
What it means: Different surveys competing for attention
Fix: Coordinate across surveys. One survey per session max.
Bounce rate increasing on surveyed pages
What it means: Survey is driving users away
Fix: Use less intrusive triggers (scroll, exit intent) and formats (slide-in vs modal).
Low completion on follow-up questions
What it means: Users are rushing through or abandoning
Fix: Make follow-ups optional. One question is enough.
Frequency & Sampling FAQ
How does frequency control work for anonymous visitors?
Valerie uses browser local storage to track whether a visitor has seen or responded to a survey. This persists across sessions but is device-specific—the same person on a different device is treated as a new visitor.
What happens if a user clears their browser storage?
They'll be treated as a new visitor and may see the survey again. This is rare enough that it doesn't significantly impact your data. If you have logged-in users, we can tie frequency to user ID instead.
Can I have different frequency settings for different pages?
Yes. Create separate survey configurations with different frequency and sampling settings. Your checkout page might use "once per session" with 100% sampling, while your blog uses "cooldown 7 days" with 25% sampling.
Does sampling affect data quality?
No. Sampling is random, so a 10% sample is statistically representative of the full population. You get fewer total responses but the data isn't biased. This is standard survey methodology.
How many responses do I need for reliable data?
For page-level scores, aim for at least 20 responses for directional insights, 100+ for statistical reliability. Valerie shows confidence levels so you know when to trust a score.
What's the right balance between frequency and sampling?
Start conservative (lower frequency, lower sampling) and increase if you're not getting enough responses. It's easier to increase data collection than to recover from survey fatigue.
Collect Feedback Respectfully
Configure frequency and sampling controls that respect your users while gathering the insights you need.