Detect subtle, significant traffic changes
We’ve spent years refining a neural network model to improve the sensitivity and accuracy of our SEO tests.
This proprietary model is trained on huge amounts of data over thousands of websites.
It means our platform can detect smaller uplifts in traffic based on lower visitor numbers than other approaches, with higher confidence.
The detail behind our testing methodology
Our neural network model accounts for external influences like seasonality, competitor activity, and search engine algorithm updates.
Any time you set up a test, the platform will:
- Create statistically-similar buckets of control and variant pages.
- Create a forecast of expected traffic to the group of pages you’re testing on.
- Measure the organic traffic to the two buckets.
- Calculate whether any difference in traffic compared to the forecast is statistically significant.
Set up for significance success
We use a proprietary bucketing algorithm to produce the most reliable split tests.
We ensure both buckets have similar levels of traffic and are statistically similar.
This means they trend up and down at similar times, helping to eliminate pre-existing bias in the buckets.
Greater sensitivity, faster results
Winning tests are great, but the sensitivity of our algorithm also ensures you don’t roll out losers.
When a change causes only a small drop in traffic, it’s often hard to see.
Our algorithm spots even small changes quickly and with confidence. So, you can take two steps forward with no steps back.
Easy analytics integration
The SearchPilot platform connects to your analytics platforms.
We offer integration with Google Analytics (including GA360) and Adobe Analytics, as well as custom reporting to help you best communicate results from our A/B tests to your organization.
Webinar replay: Drive Performance With These 4 Disciplines for SEO Teams
How to make sense of SEO A/B test results
