Basic Filters are essentially inclusion and exclusion rules that narrow down the scope of the recommendation output. Some filters are useful for enhancing the contextual accuracy of recommendations, while others serve as effective merchandising tools.

For instance, basic category filters can limit recommended products to the same category that a user is viewing, transforming generic cross-selling recommendations into suggestions for alternative products. When combined with price filtering and relative conditions, the same recommendation can be tailored to mainly show slightly more expensive alternatives, creating a dynamic upselling effect.

Advanced dynamic filters support the creation of precisely paired recommendations based on product attributes and serve as a merchandising tool. These filters operate on the principle of "if this, then that" conditions, requiring more manual input but providing a higher level of control.

The scope of each filter is limited to the recommendations where the rules are applied. Global exclusion rules allow the removal of products from all Nosto features.

The majority of filters operate based on comparing the page's context with the inclusion or exclusion criteria defined by the filters. For example, brand, category, custom field, and tag-based filtering can be configured to match the current page by including products that meet the criteria or, conversely, excluding products that have the same criteria. Handpicking with custom attributes is also supported, which is commonly useful for exclusions like removing free gifts or gift cards. The logic behind it is relatively simple: include products with the same attribute or include products without the same attribute or simply select the attribute.

Filters with numeric attributes, such as price, have specific options explained separately within each filter type.

Note that filters are dependent on the integration and data mapped to Nosto. For example filters such as margin, ratings&reviews make use of the corresponding data, which is optional for Nosto to operate, yet required for respective filters to work.

Brand is manufacturer, supplier or as the name suggests, brand itself

Category is product’s hierarchical taxonomy or product range such as /womens/clothing/jackets. Note: On Shopify, collections map as a Nosto category, but Shopify’s product type as a category is available as Custom Fields > Product Type

Color Matching makes use of Nosto’s artificial intelligence and color detection.

Custom fields are custom attributes mapped to Nosto. These commonly feature useful attributes like product type, gender target audience, marketing campaign details and similar. As the exact attributes and values differ from site to site, use Catalog Explorer to review the integration on your site.

Inventory level looks into product inventory level and supports inclusions and exclusions based on the given input. Full availability option excludes all products that are not available in all variants e.g. in all sizes.

Margin is the gross margin of the product comparing supplier cost to the selling price. Margin filters support including products only based on the given criteria.

Show new products looks into product creation date and also optionally shopper’s visit history. In short, Nosto can recommend products that have been added to the inventory since the given criteria in days or since shopper’s last visit to the website. The latter operates with the logic: New products added since your last visit.

Price filters are intended for multitude of use cases. Basic exact minimum and maximum value delimit the recommended products within the given criteria and are useful when the intention is to remove very low-priced items such as add-ons from recommendations. Relative price filtering ties the relation either to the product viewed or shopping cart content. Former supports dynamic upselling filtering such recommending only slightly cheaper options when the minimum input has value such as -10%. Latter has different modes depending on the intended use case, where relative prices are tied to average, most expensive or cheapest item in cart.

Ratings and reviews are product’s ratings and review-count supporting inclusion rules, which effectively can exclude products that don’t meet the given criteria.

Tags are custom attributes as mapped to Nosto. Strict matching option expects the fields to fully match with each other, whereas when left disabled the logic is contains the same tags.

Discount filter includes only products that have been discounted by comparing list price to the current price. When price < list price, product is considered as discounted e.g. on sale.

As the name suggests, exclusions are the opposite to inclusions and add support to exceptions to the inclusion rules. For example, include only new products, but exclude a certain category.

ProductIDs is support removing handpicked products from the recommendation. The opposite, hand-picking products what to recommend is supported by a recommendation type Cherry-picked recommendations and by more advanced Dynamic Bundles.

Filters may be applied strictly by selecting Show only filtered results, but this setting might leave product slots in the recommendation empty as unfilled if filters are too restrictive. The amount of products is defined in the recommendation creation with the maximum number of products -option.

Fill with filtered bestsellers will apply given filter settings strictly, but in case there are empty product slots, best selling items respecting the filter criteria will take place of the remaining slots.

Fill with unfiltered option will fallback to products across all ranges based on popularity, but by keeping filtered products prioritised in the recommendation.

Advanced Dynamic filtering supports creating hand-picked and exact recommendation pairing or merchandising based on following product attributes:

The feature uses behaviorial data just like regular filtering by showing the best and most relevant options, but by narrowing down recommendations to hand-selected product subsets. The easiest way to comprehend the functionality is to perceive dynamic filtering as a set of conditional clauses: “If something is true, then do the following.”

Advanced Dynamic Filters are applicable for Product Based Recommendations, Cart Based Recommendations and for Best Sellers. The context is based on viewed product on product based recommendations, all products held in cart on cart based recommendations and on viewed category for Best Sellers.

Advanced dynamic filtering supports filling the available item slots based on the given criteria, similarly to basic filters. Start by adding a case, which will reveal filter options and proceed by adding a condition clicking add case.

Arguably a very common option is to do recommendation pairing based on product categories. As an elaborative and practical example, if a customer is viewing product in tees, dynamic filtering allows narrowing down cross-sellers to most relevant options in a specific and pre-defined product category or categories such as pants, whereas regular filtering might recommend ties but also other relevant supplementary products across all product ranges. Adding conditions will effectively override what recommendation engine considers as the most relevant product recommendations by applying advanced filter settings explicitly.

In the example below, when a product viewed belongs to category Tees the recommendation will feature products only from categories pants, shoes and caps.

As an optional else-rule, a secondary rule using criteria products from other categories as a secondary rule will generate a regular recommendation outcome.

Certain brands might not allow cross-selling of other brands on their product detail pages due to brand guidelines or other business rules. Dynamic filtering can narrow cross-sellers only to the same brand, but by cross-selling other brands when cross-selling is not limited.

In the example if a product brand is Apple and category is tech, the condition below would only allow other Apple branded tech products in the recommendation, while with other brands the cross-selling of other brands is allowed.

Rule regression and order is an imperative importance when multiple conditions as cases are applied, which also allows sophisticated use-cases of dynamic filters. Cases are read from top to bottom or from first to last, meaning the last one should always be the fallback, whereas the first one should always be the most explicit rule you want to apply.

If a dynamic filter setting doesn’t yield an expected result, it’s likely that another rule takes precedence and re-ordering the cases solves the issue. Drag and drop cases to get the preferred output and end-result

Multiple conditions enable support for use cases when particularly specific recommendations are needed. Conditions are inclusive meaning all conditions need to be matched in order the recommendation to yield preferred outcome. In the example a tee by Patagonia from the subset of outdoor would always be featured with Patagonia caps within price range 10-50.