Product sales per variant

This widget displays the top 10 products contributing the most to each variant's performance. It helps you understand which products drive engagement, revenue, or conversions across different test versions.

By comparing product metrics between variants, you can uncover how each variation influences user purchase behavior, product interest, and overall experiment outcomes.

Metrics

The chart includes:

1. Variant filter
   A filter to select orders of a specific variant.
2. Currency
   A filter to select orders in a specific currency.
3. Product
   The name or identifier of the product.
4. Quantity
   The total number of units sold.
5. Quantity share
   The percentage of all units sold that this product represents.
6. Revenue
   The total revenue generated by the product.
7. Revenue share
   The percentage of total revenue that comes from this product.

Interpretation

This widget helps you analyze how product performance differs across variants in your experiment. Imagine Variant A generated sales of 200 products in total, 30 of which were white T-shirts. This means 30 white T-shirts were sold in Variant A, making up 15% of all units sold for that version.

We calculate this metric using the formula:

$$Quantity \space share = \frac{units \space of \space product \space sold}{total \space units \space sold}$$

Now consider revenue. If those 200 products together brought in $6,000, and each white T-shirt was priced at $25, then the white T-shirts accounted for $750 in sales. That corresponds to 12.5% of the total revenue for Variant A.

We calculate this metric using the formula:

$$Revenue \space share = \frac{revenue \space of \space a \space product}{total \space revenue}$$

What the number tells you

Here are some common insights you can get from this widget:

• Differences in product rankings across variants reveal which version better highlights or motivates users to buy a given product.
• If a promoted or featured product appears among the top performers in one variant, it may indicate that this variant's changes positively influenced its sales.
• If expected products don't perform well across variants, it may suggest weak user interest or a need to adjust targeting, pricing, or messaging.