Suggested reading:
Why did impressions for my campaign with the Manual Merchandising algorithm decrease suddenly?
For Manual Merchandising, you explicitly choose which products to recommend. If some of those selected products go out of stock, your widget may show fewer products or even no products at all, resulting in a decrease (or zero) in impressions.
Example: Check your Product Catalog Management to ensure that all product IDs in your manual selection are in stock.
Why is there no impression for my campaign?
There could be several reasons if your Smart Recommender campaign has no impressions:
Your filters or strategy might be too restrictive, so there are no products matching the recommendation criteria.
All products meeting the strategy criteria could be out of stock.
The widget’s container location on your website may have changed or been removed (for example, due to a website update).
Your segment and rules might be too narrow, resulting in very few users qualifying to see the widget.
The widget placement might be too far down the page, causing users to leave before scrolling to it.
Example: If your campaign is set only to recommend “Men’s Formal Shoes” and several key SKUs are sold out, no products would be available for recommendation, and thus no impressions would be logged.
Why is the impression count less than the click count of Smart Recommender Analytics?
Since an impression is logged only when at least 50% of a product card is visible in the viewport, users may click on a product before scrolling far enough for it to log an impression.
Another possibility is an integration issue (especially with JavaScript SDK-based campaigns).
Example: A user may quickly tap on a product even if only 30% of the product is visible, causing a click but not meeting the criteria for an impression.
Why does the sum of category analytics revenue not match revenue in overall analytics?
Each product can belong to multiple category layers (e.g., Men > Clothes > Shirts), so the same revenue is attributed to multiple categories. This means that summing revenue across categories will result in a figure higher than the actual overall revenue.
Example: If a shirt generates 100 USD in revenue, it might be counted in “Men,” “Clothes,” and “Shirts” categories, leading to a summed revenue of 300 USD across the categories, even though the actual revenue is 100 USD.
Why can product clicks be less than product add-to-cart counts?
Users may first click a recommended product to view its details and then decide to add it to the cart later. Each add-to-cart action is logged separately, even if it’s for the same product.
Example: If a user clicks on Product A and later adds it to the cart twice (maybe due to increasing quantity), you’ll see two add-to-cart logs despite only one product click.
Why do I see different algorithms for the same campaign when filtering different date ranges?
While your campaign runs, you may adjust strategies or experiment with different algorithms over time. When you filter by date, you’re essentially viewing the performance of each algorithm as it was applied during that period, showing you exactly how changes in strategy affected the results.
Example: In January, you might have used a “User-Based” algorithm. In February, you switched to “Most Popular.” Filtering by these dates lets you see the performance differences between the two approaches.
Why is Assisted Revenue showing high values but low Direct Revenue?
Assisted Revenue is recorded when a shopper clicks on a recommended product and then, in the same session, purchases a different product. If your campaigns are effectively driving product discovery but not resulting in the purchase of the originally clicked item, you may see higher assisted revenue.
Example: A shopper clicks on Product E but ends up buying Product F. The revenue for Product F is recorded as Assisted Revenue. If this pattern is typical, it indicates that recommendations are sparking interest in broader product discovery, even if not directly converting the clicked product.
Why does the same product show different metrics across campaigns or placements?
A product can appear in multiple campaigns (for example, on the Home Page and within Category Pages) or in different widget placements. Each instance is tracked independently for impressions, clicks, and conversions. Evaluate performance within the context of where and how the product is displayed.
Example: Product X might have 500 impressions in a Home Page campaign but only 200 in a Category Page campaign. The difference may reflect the varying visibility and attractiveness of the placements.
Why do I see high impressions but low CTR for some campaigns?
This can happen when the widget’s placement is suboptimal (e.g., below the fold or in a less engaging part of the page), or when the design isn’t attractive (lacking compelling images or calls-to-action). It’s also possible that product assortments do not match user intent.
Example: A campaign placed at the very bottom of the page may log many impressions because users pass through that area, but if the products aren’t compelling, clicks remain low.
Why are conversions dropping after I changed the strategy in a variant?
When you change the campaign strategy (for instance, switching from “User-Based” to “Manual Merchandising”), there may be a period of adjustment as user behavior shifts. It can take time for a new approach to align with customer expectations. Also, ensure that product stock and presentation remain consistent.
Example: After switching to a new algorithm, you might initially see a 20% dip in conversions until the algorithm is fine-tuned to your audience’s behavior.
Why does my product appear in Top 100 Analytics but not in individual campaign results?
If a product is recommended in multiple campaigns, its cumulative performance is captured in the Top 100 Product Analytics. However, if an individual campaign’s threshold for visibility or engagement isn’t met, the product might not stand out in that specific campaign’s results.
Example: A best-selling product might perform well overall when aggregated but appear low in a single campaign if that campaign had limited impressions or targeted a niche segment.
Why does a product have impressions but zero clicks or conversions?
Possible reasons include:
The product is consistently in the first slide (hence, viewed frequently), but it isn’t engaging enough for users to click.
The product image, price, or description may be unappealing.
There may be issues with the widget design that hide key information. Example: A product might be well-stocked and displayed but have an unappealing thumbnail, resulting in lots of impressions but few clicks or purchases.
Why are duplicated products shown in the Top Product list with different IDs?
If your catalog uses multiple SKUs or variant IDs for the same product (for example, different colors or sizes), each variant is tracked separately in the analytics. This causes the same product to appear more than once in the Top Product list.
Example: A t-shirt in red and blue may appear as separate entries if each has a different SKU.
Why do the same campaigns perform differently across platforms (mobile vs. desktop)?
User behavior, screen size, and layout differences greatly affect engagement. A carousel that works well on desktop—where users have a larger screen—might be less effective on mobile, where space is limited and scrolling behavior differs.
Example: On a desktop, a well-designed carousel might have high engagement, whereas on mobile, the same carousel appears too small and results in lower CTR and conversions.
Why do I see click logs but no revenue, even after many sessions?
This situation may arise if:
The product is not competitively priced or attractive enough.
It frequently goes out of stock.
Users are merely curious and don’t follow through to purchase.
Example: If Product Y gets a high number of clicks due to appealing imagery but is priced too high compared to alternatives, it may generate many click logs without resulting in revenue.