DE
BRData-backed decisions with sequential statistics
VWO’s enhanced SmartStats - Bayesian powered sequential testing engine empowers your experimentation with unparalleled flexibility. Set statistical configurations to match your business needs and get swift, reliable results you can trust.
Say hello to VWO’s enhanced SmartStats
Our enhanced SmartStats - Bayesian powered sequential testing engine gives you full control of the statistical parameters of your experiments to ensure precise, statistically robust results for confident decision-making.
Configure metrics your way
Create metrics using any event, customize their statistical settings, and save them for reuse in your campaigns.
Get error-free, reliable reports with sequential statistics
Get error-free real-time reports with our stats engine that auto-adjusts for peeking and multiple comparisons.
Accelerate experiment results
Reduce visitor loss by quickly disabling underperforming variations, enabling faster evaluation of others.
Be sure of your experiments’ integrity
VWO monitors experiment health, performs background checks, and notifies you of issues, ensuring you can relax.
A flexible sequential statistics engine that helps you scale
Our stats engine enables you to tailor metrics and statistical parameters, estimate test duration, and optimize experiments for speed or power.
Optimize experiments on low-traffic pages
Minimum Detectable Effect (MDE) is the smallest conversion change you want to detect. Raising MDE speeds up experiments on low-traffic pages, leading to quicker conclusions, though it might miss minor improvements.Prioritize significant improvements
The Region Of Practical Equivalence (ROPE) is a range where conversion changes are considered equivalent to the baseline. This helps our stats engine prioritize significant improvements over minor ones.Balance efficiency and accuracy
False positive rate (FPR) gauges the chance of mistakenly identifying significant effects. Adjusting FPR balances efficiency and accuracy: higher FPR boosts efficiency in non-critical campaigns, while lower FPR ensures accuracy in critical ones.Control the power of your experiments
Statistical power is the probability of detecting a true effect. Increase power for critical metrics to catch smaller uplifts and decrease it for less critical metrics to save resources while ensuring reliability.Save and reuse metrics with statistical defaults
Set the statistical parameters and desired improvement direction (upward for metrics like conversions and revenue, downward for metrics like bounce rate and page load time) for each metric based on your requirements. Save these metrics with their values as defaults to reuse in other campaigns.Set guardrails to protect critical business KPIs
Configure guardrail metrics to automatically pause tests if configured metric conversion threshold values are breached. Guardrail metrics minimize losses and protect critical KPIs when experiments don’t go as expected.Get crystal-clear experiment results for deeper analysis
Our enhanced, easy-to-interpret reports distill experiment information that enables you to make data-driven business decisions.
Get clear comparisons against the baseline
Identify if variations outperform, match, or fall below the baseline with our advanced stats engine. Results are declared at a 95% probability threshold. Detailed graphs offer clear insights for effective decision-making.
Track recommendation history
Track campaign progress with a visual progress bar showing the estimated time to completion, decision recommendations for variations, any breached experiment vitals, and corresponding timestamps.
Segment-specific analysis made easy
Quickly filter reports by date, visitor segments, or dimensions for precise segment-specific insights. Access probability density graphs directly within your reports for detailed variation analysis.
Get comprehensive oversight at a glance
Toggle between a simple and an in-depth view to see conversions, expected conversion rates, expected improvement, probability of improvement, and statistical configurations in a unified view.
Get error-free results that you can bank on
Get error-free results for all your experiments with built-in handling for common errors such as peeking and multiple comparisons.
Built-in sequential testing
Sequential testing engines monitor probability values to recommend deploying or disabling variations at statistical significance. Our stats engine ensures accurate reports by adjusting probabilities based on maximum sample size.Built-in Bonferroni correction
Testing multiple variations raises the risk of false winners. Traditional Bonferroni correction shifts decision boundaries, but our advanced method integrates uncertainty into probability values. The correction provides nuanced analysis and controls error rates, ensuring more reliable results.Run fixed horizon tests
Fixed horizon test calculates results only after collecting the entire sample size. Conclusions are made after the campaign's pre-set duration ends, making it ideal for enterprises ready to allocate significant visitor traffic to their experiments.Get the best combination of test modes, sample size, and duration
Choose the best-suited test modes, estimate sample size, and set campaign durations for optimal results.
Reduce conversion loss with timely recommendations
Our advanced stats engine helps you decide when to disable or roll out variations, ensuring statistical accuracy. It prevents conversion loss by recommending you disable variations with less than a 5% chance of success, speeding up evaluations and focusing on successful changes.
Play around with our sample size calculator
*coming soon
Get an estimate of how many visitors your campaign needs to conclude even before you begin your campaign.
Relaxed comparison conditions with non-inferiority tests
Achieve faster results in experiments where both improvement and equivalence against the baseline are acceptable.
Ensure clean experimentation
Our stats engine runs background health checks to ensure reliable results. If issues are detected, you can ignore them or take appropriate action based on their severity.
24/7 Comprehensive vitals check
Our stats engine runs critical checks on data tracking, experiment conduct, and potential issues like Simpson’s paradox, ensuring comprehensive and trustworthy experiment results.
Avoid data collection issues *coming soon
Sample Ratio Mismatch (SRM) is a sanity check to identify data collection and test setup discrepancies. If an SRM check fails, our stats engine will notify you of unreliable campaign results.
Ensure no outliers skew the results *coming soon
Ensure data integrity with built-in Outlier Detection running in the background. Switch on the optional Outlier Replacement to automatically replace deviations with the mean and receive notifications for the same.
We strive to continuously enhance our reports
Our latest updates offer a perfect mix of power, flexibility, and speed, enabling quick, reliable, and easy-to-understand experiment results. Stay tuned for more report enhancements.
Repository of stats blogs
Learn how our customers use behavioral data insights to improve user experiences, run data-backed experiments, and impact conversion metrics.
