Ranking plays a key role in guiding shoppers through search and navigation landing pages. When rankings do not match shopper expectations, users may struggle to find relevant products, leading to lower engagement and missed revenue opportunities.

The Insights dashboard helps you identify rankings where shoppers are not interacting with products as expected. By reviewing these rankings and addressing the underlying issues, you can improve product discovery and overall performance.

This article outlines a how to review ranking performance and resolving common issues using the Insights ranking performance report.

Learn more:  For a more general introduction to Insights visit Inside Insights: A User Guide.

Establish a benchmark

Once you have identified the ranking you want to review, define what “good” performance looks like. To determine whether it is under-performing, establish a benchmark ranking that serves as a point of comparison.The benchmark ranking should be similar in context to the ranking to be reviewed, such as:

• Similar categories
• Similar product types
• Similar shopper intent

Avoid comparing rankings for unrelated categories as differences might be expected, e.g. dresses vs. socks, or rankings for categories that will show natural differences, e.g. men’s wear vs. women’s wear.

Example

As a fashion retailer, if you want to compare the rankings used for the categories “dresses” and “skirts”, ask the following questions:

• Do these categories use the same ranking? If so, does the strategy reflected by the ranking make sense for both of these categories?
• Does the ranking used for either of these categories heavily use result modifications?

Warning: Do not use heavily merchandised pages as either target or benchmark ranking. Results of heavily merchandised pages are not primarily driven by rankings. 
For example, if you use multiple result modifications for the "skirts" category, it is not an appropriate benchmark for the "dresses" category.

Identify under-performing rankings

Once you have established a benchmark, review and compare the high-level KPIs of the benchmark and the ranking in question:

• View conversion (CTR)
• Basket conversion
• Purchase conversion
• Revenue per unique impression

(1) Use the Insights ranking performance report and adjust the table to show the metrics above. These metrics will help you understand how well a ranking does on a given page. 

(2) Select the rankings you want to review and click “Compare”.

(3) Adjust the time range to see potential trends in the ranking performance, e.g. one month. Ensure you select the KPI of interest to be be presented in the graph.

Note: Depending on which of these metrics deviate from the benchmark, it will be an indicator for where in the journey you are losing customers. For example: A high CTR but a low basket conversion indicates that users are not finding relevant products despite engaging with the page.

(4) Other friction or performance issue indicators are:

• Facet interaction rate
• Paging interaction rate

If either of these metrics are high, this could indicate that your users are not finding what they are looking for early and are forced to heavily rely on filters or frequently go to the next page.

Note: A high facet interaction rate is not always an indicator of an under-performing ranking. For example, you can expect the facet interaction rate to be high for pages that invite mission-driven purchases, e.g. video games category often filtered by console. Similarly, a high paging interaction rate might be expected for diverse and large product catalogs as, in this context, customers would be expected to browse and look at multiple pages.

Warning: An extremely low facet interaction rate might be a sign that the facets shown on a given page are not relevant to the customer and facet performance should be further investigated.

Diagnose the cause

Once you have used the Insights ranking performance report to identify an under-performing ranking, determine why it is not performing well.

• Identify funnel drop-offs: Using the metrics mentioned above, identify at which point of the customer journey performance drops off.

• Review front-end experience: Always validate findings by reviewing the live page. Ensure that the ranking performs as expected on the live page, that all relevant products are visible and no key products are missing. For example, if you review the “new arrivals” ranking on the front-end but older products are shown at the top, this might indicate an implementation issue. Review the ranking details to ensure your configuration matches your intent and is set up correctly.

Note: Ensure that the results you are reviewing are driven by the ranking itself and are not heavily overwritten by result modifications.

• Distinguish between implementation and strategy issue: If the ranking does not behave as intended when reviewing the front-end, you are likely dealing with an implementation issue. However, if the ranking works as configured but is not effective, your chosen strategy is likely to be the issue.
• Review your strategy: Review the goal of your strategy and the relevance of the strategy for the page of interest. Verify that the strategy currently in place reflects your ideal strategy, typically a balance between (1) user-focused factors, such as popularity and trends, (2) availability factors, such as stock levels and size availability, and (3) business factors, such as margin and own-brand prioritization.

Warning: If business factors, such as margin, outweigh user-focused factors, this can negatively impact user engagement and conversion.

Apply optimizations

Once you have identified the cause, apply targeted improvements:

• If there is an implementation issue: Review ranking configuration, validate attribute data quality, and ensure that the ranking logic matches the intended strategy.

• If engagement is low: Revisit your strategy, improve relevance signals, and adjust ranking logic toward user intent. Ensure you use attributes that accurately reflect and predict user intent.

Note: For more in-depth information on which attributes to use to best predict user intent in your specific use case, reach out to your Experience Consultant.

• If users browse too much (high pagination with low facet interaction): Improve top-of-page relevance, ensure the most relevant products appear first, and review whether facets accurately reflect the category’s diversity to better support filtering.
• If users rely heavily on filters (high facet interaction): Ensure the top results already surface the most relevant products, especially for mid-sized categories, so users are not forced to depend on filters to find suitable items.
• If users click but don’t convert: Review pricing competitiveness, improve product selection focusing on products that drive conversions, and ensure that the products shown are available.
• If revenue is low but engagement is strong: Adjust the product mix and promote higher-value or higher margin products where relevant.

Track and validate changes

Validate your changes by reviewing changes in the front-end. Moreover, use A/B testing to thoroughly validate your optimizations. You can view A/B testing results in the Insights A/B testing report. This will help you confirm whether changes lead to measurable improvements in engagement, conversion, and revenue.