Better experiments.Remarkably accurate results.
We test your creative decisions with real people
Case studies
BoranSample: a clearer offer improved customer preference by 18 points.
South Street CafeSample: customers selected a stronger hero photo for the listing.
Patriot Garage DoorSample: testing identified the message most likely to earn a call.
Liora's CateringSample: real customers compared positioning before launch.
How it works
We use simulated environments to generate more reliable predictions.
01 We build your experiment
Upload photos, names, ads, descriptions, or offers. We place them into a realistic customer experience and prepare balanced comparisons.
02 Real customers evaluate your options
Verified participants compare your ideas head-to-head and tell us what they genuinely prefer.
03 We identify the winner
We analyze every comparison, preference strength, and confidence level to determine what performs best.
04 You receive actionable insights
You get a clear result and practical recommendation so you know exactly what to do next.
Pricing
Choose the study size and support you need. We’ll match you to the right plan automatically.
| Tier | Price | Participants | Recommended # of Options iThe lower the number of options, the higher the confidence. | Confidence | Best For | Additional Features |
|---|---|---|---|---|---|---|
| Starter | Free | 15 | 2 | Low | Trying Sigan | Randomized experiment design, preference winner, strength of preference, and standard recommendation |
| Green | $24 | 30 | 2 | Medium | Lower-risk decisions | Everything included above, basic audience targeting and multi-element testing |
| Blue | $39 | 50 | 2–3 | High confidence for testing 1 element iAn element is one thing you are testing, such as an item name, photo, price, or rating. | Important items, stronger ability to determine close results | Everything included above, price sensitivity analysis |
| Purple | $59 | 80 | 2–4 | High confidence for testing 2+ elements iAn element is one thing you are testing, such as an item name, photo, price, or rating. | Important items, testing multiple options with high confidence | Everything included above, advanced audience targeting and deeper analysis across multiple elements |
| Black | $109 | 125 | 2–5 | Very high confidence for testing 2+ elements iAn element is one thing you are testing, such as an item name, photo, price, or rating. | Most important content, DoorDash header photos, highest-priority items, Google listing information, multi-element testing | Everything included above, priority processing |
- Participants
- 15
- Recommended # of Options iThe lower the number of options, the higher the confidence.
- 2
- Confidence iAn element is one thing you are testing, such as an item name, photo, price, or rating.
- Low
- Best For
- Trying Sigan
- Additional features
- Randomized experiment design, standard recommendation
Partner program
Become a
Business Partner
The Business Partner program is built for businesses looking to continuously optimize their online presence through a high volume of experiments at a considerably discounted price.
We offer this program because we’re actively validating our recommendations against real-world business outcomes and improving our models.
In return for the discount, Business Partners agree to share agreed-upon performance data after implementing recommendations.
Apply to become a partnerYou receive
- 50% off10 experiments
- Dedicated account advisorto help you plan experiments
- Personalized recommendationsbased on your results
We receive
- Real-world result sharingwhen decisions are implemented
- Permission to use resultsfor research and case studies
Our privacy policy
We use anonymized details to ensure your privacy. Everything we show to testers except explicitly what you’re studying contains randomized information.
Who are our testers?
As primarily a group of statisticians, our goal is to ensure the absolute highest data quality.
We maintain a network of 1,000+ verified consumer testers and 250+ enterprise testers.
Every tester is paid top of the market rates and has an associated performance rating based on how accurately their preferences predict real-world outcomes.
Testers whose choices consistently outperform the average receive higher ratings and are selected more often for future studies.
Additionally, if your study requires specialized testers, we not only recruit participants but also validate their data quality before the study begins.
Become a tester