Foundation · 2 min
What you need to know
How should ecommerce conversion rate be measured?
Choose a defined outcome and divide completed outcomes by the eligible sessions, users or attempts. Label the numerator, denominator, exclusions, date basis and counting method beside the rate.
Should conversion rate include all visitors?
Not automatically. Exclude verified bots, employees, test traffic and ineligible experiences, but do not remove real low-intent visitors merely to improve the rate. Preserve an all-traffic view for reconciliation.
Why can overall conversion rate be misleading?
An overall rate can hide device, source, country, product and customer-mix shifts. It can change even when every segment’s own conversion stays stable.
How should conversion be compared across devices?
Calculate the same conversion definition separately by device category, then compare funnel steps, traffic mix and sample size. Do not interpret device as the cause without investigating the experience.
How should conversion be compared across countries?
Use the same eligible outcome and denominator within each country, then account for availability, price, currency, delivery and legal differences. Small-country rates need uncertainty context.
How should conversion be compared across customer types?
Separate clearly identified first-time and returning customers using consistent identity rules. Report unmatched guest activity rather than silently assigning it to either group.
How should conversion be compared across products?
Compare product-view-to-purchase or another product-level rate using stable item IDs and availability rules. Avoid using store sessions as the denominator for products most visitors never saw.
Why can conversion rate rise while revenue falls?
Conversion can rise while revenue falls if traffic, average order value, price or high-value product mix declines. Decompose revenue instead of judging the rate alone.
A quick store check
Write the conversion numerator, denominator, exclusions and time zone beside the report before comparing segments.
Foundation · 2 min
Frequently asked questions
Should orders or customers be used for conversion rate?
Either can be valid. Label the numerator and denominator and avoid comparing rates built from different units.
Should bot traffic remain in the conversion denominator?
No when it can be reliably identified, because non-human visits make customer conversion performance misleading.