--- title: "Step 4: Analyze Recommendation Performance" slug: "step-4-analyze-recommendation-performance" description: "Evaluate your recommendation widgets' performance in Mapp Intelligence with insights on user interactions, conversions, and effective algorithms." tags: ["Recommendation Tracking", "Performance Analysis", "Ecommerce Metrics"] updated: 2026-10-07T10:27:08Z published: 2026-10-07T10:27:08Z canonical: "docs.mapp.com/step-4-analyze-recommendation-performance" --- > ## Documentation Index > Fetch the complete documentation index at: https://docs.mapp.com/llms.txt > Use this file to discover all available pages before exploring further. # Step 4: Analyze Recommendation Performance Once Recommendation Tracking is collecting data, you can evaluate results directly in Mapp Intelligence. Analyses help you understand which recommendation widgets perform well and which products contributed to conversions. --- ## Available dimensions Recommendation Tracking generates **two types of dimensions**: 1. **Widget context dimensions** based on the predefined E-Commerce parameters (placement, position, algorithm) 2. **Product-level dimensions** based on the loaded, viewable and clicked recommendation events (product + product category) | **Dimension** | **Description** | | --- | --- | | Algorithm (in reco widget)* | Recommendation algorithm used (for example, similar items). Based on E-Commerce parameter ID 767. | | Placement (in reco widget)* | Where the widget appeared (for example, PDP, homepage). Based on E-Commerce parameter ID 765. | | *Yourcategoryname* loaded (in reco widget) | Category of loaded products. One dimension per category. | | *Yourcategoryname* viewable (in reco widget) | Category of viewable products. One dimension per category. | | *Yourcategoryname* clicked (in reco widget) | Category of clicked products. One dimension per category. | | Products loaded (in reco widget) | Lists products loaded into widgets | | Products viewable (in reco widget) | Lists products that became viewable in widgets | | Products clicked (in reco widget) | Lists products clicked in widgets | * These names are customer-defined. The system uses the label you configure when creating the parameter. > [!NOTE] > Note > > The extension does not send product categories. A product receives its category once, through your regular product tracking. With the Unified Product Catalog, categories are assigned only through the catalog import. If a product is measured for the very first time in a recommendation widget without a category, it does not receive one, and it does not appear in the category dimensions. --- ## Available metrics | **Metric** | **Description** | | --- | --- | | Qty Products loaded (in reco widget) | How often a product was loaded into a widget | | Qty Products viewable (in reco widget) | How often a product became viewable in a widget | | Qty Products clicked (in reco widget) | How often a product was clicked | | Product position (in reco widget)* | Product's position inside the widget (1, 2, 3 …). Based on E-Commerce parameter ID 766. Useful for filtering and breaking down performance. | | Value Products loaded (in reco widget) | Total value of products loaded into widgets. The value is taken from the product price submitted with the event. | | Value Products viewable (in reco widget) | Total value of products that became viewable in widgets. The value is taken from the product price submitted with the event. | | Value Products clicked (in reco widget) | Total value of products clicked in widgets. The value is taken from the product price submitted with the event. | * These names are customer-defined. The system uses the label you configure when creating the parameter. Loaded, viewable and clicked always narrow down in that order. A product has to be loaded before it can become viewable, and viewable before it can be clicked. See [Recommendation Tracking](/v1/docs/recommendation-tracking-plugin-intelligence) for what triggers each status, and [Step 2: Configuring and Activating the Plugin](https://docs.mapp.com/docs/recommendation-tracking-step-2-configuring-and-activating-the-plugin) for the limits that can cap the loaded and viewable counts. > [!NOTE] > Note: Automatic Qty metrics > > For each widget-context dimension, an additional Qty-metric is created automatically (e.g., Algorithm of reco widget → Qty Algorithm of reco widget). These Qty-metrics simply count how often the parameter was passed. They are mainly useful for diagnostics, not for business reporting. --- ## Example Analyses These examples demonstrate how to evaluate the performance of your recommendations in Mapp Intelligence. All analyses use data collected via the Recommendation Tracking extension. To analyze orders or revenue, use standard e-commerce metrics such as Qty Purchased Products or Revenue together with filters on recommendation parameters, as shown in [example 2](/v1/docs/step-4-analyze-recommendation-performance#2-which-recommended-products-lead-to-purchases). In addition to the examples below, many other recommendation-related dimensions and metrics are available. You can freely combine them to explore algorithms, product categories, or audience segments in more depth. ### 1) Which recommended products receive the most user interaction Shows which recommended products receive the most attention and clicks. Use this analysis to identify items with strong engagement potential. **How to build** - **Dimension:** Products - **Metrics:** Qty Products loaded (in reco widget), Qty Products viewable (in reco widget), Qty Products clicked (in reco widget) - **Optional:** Create a custom metric to calculate a click rate. Clicks divided by viewable products tells you how well a product converts once it has actually been on screen; clicks divided by loaded products mixes that with how much of the widget was ever seen. You can use this as a starting point to decide which products should be promoted more prominently. --- ### 2) Which recommended products lead to purchases Highlights products that are not only clicked but also purchased within a seven-day period. This helps identify items with real commercial impact and shows which recommendations effectively drive conversions. **How to build** - **Dimension:** Products - **Metrics:** Standard e-commerce KPIs (e.g., Qty Purchased Products, Value Purchased Products) ![](https://cdn.document360.io/554b9b98-6720-4d8b-9919-c7b203d72648/Images/Documentation/CleanShot 2025-11-12 at 09.47.22.png) - **Filter:** Product position (in reco widget). filter for any value![](https://cdn.document360.io/554b9b98-6720-4d8b-9919-c7b203d72648/Images/Documentation/CleanShot 2025-11-12 at 09.44.57.png) This filter is required because the Products dimension and eCommerce metrics are not specific to recommendations. Filtering by "product position (in recommendation widget)" ensures that only purchased products that were previously tracked as part of a recommendation are included in the analysis. While other eCommerce parameters can also be used to identify recommendation-based purchases, the product position parameter is the most reliable choice because it is attached automatically whenever a recommended product is clicked and later purchased. > [!NOTE] > Note > > A click-to-purchase connection is recognized if the purchase happens **within seven days** after the recommendation click. --- ### 3) Which widget placements and positions perform best Shows where recommendation widgets generate the highest engagement and how product position affects results. **How to build** - **Dimension:** Placement (in reco widget) - **Metrics:** Qty Products loaded (in reco widget), Qty Products viewable (in reco widget), Qty Products clicked (in reco widget) - **Optional:** Filter or break down metrics by Product position (in reco widget) to compare positions 1, 2, 3 … Homepage and PDP placements often perform differently. This view helps refine your widget layout strategy. Comparing loaded against viewable products per placement also shows how much of a widget ever reaches the screen. A placement that loads many products but makes few of them viewable sits too far down the page. **Advanced option:** Add Products in a pivot table to explore which specific products perform best within each widget and position. --- ### 4) Which algorithms generate the most effective recommendations Shows how different recommendation algorithms perform and which products get the most clicks within each algorithm type. **How to build** - **Dimension:** Algorithm (in reco widget) - **Metrics:** Qty Products clicked (in reco widget) **Advanced option:** Add *Products clicked (in reco widget)* in a pivot table to compare product performance across algorithms. ## Related - [Qty Products viewable (in reco widget)](/qty-products-viewable-in-reco-widget.md) - [Value Products clicked (in reco widget)](/value-products-clicked-in-reco-widget.md) - [Value Products viewable (in reco widget)](/value-products-viewable-in-reco-widget.md)