Skip to content
Go to raptorservices.com
  • There are no suggestions because the search field is empty.

Module adjustment and tuning

What is module tuning?

Each recommendation module in the Raptor Control Panel can be tuned using a set of weighting parameters that influence which products are shown and in what order. Tuning lets you shift the balance of recommendations to match your business context, for example by favouring products from a shopper's purchase history, boosting high-margin items, or increasing variety across the site. You do not need a developer to adjust these settings.

How to open a module for tuning

  1. In the Raptor Control Panel, navigate to Recommendations in the sidebar and click Website.
  2. Find the module you want to adjust and click it to open the module page.

    web-1

  3. In the Mandatory Input Parameters box, enter a test value to preview the module output. You can find valid values (such as a product ID, category ID, or brand ID) by clicking the magnifying glass icon and selecting a parameter from the live data stream.
  4. Click Test to preview the current output.

Module Adjustment 1

Scroll down to the Optional Input Parameters box. This is where all tuning parameters are found. 

🔍 Note: To make recommendations personalised for each visitor, add the visitor's cookie ID or user ID at the top of the Optional Input Parameters list. Without this, the module returns non-personalised output. 

Tuning parameters

The parameters below control how the recommendation algorithm weights different signals. Each parameter accepts a numeric value. Higher values give that signal more influence relative to the others.

VisitHistoryWeight
Boosts products that the visitor has previously viewed. Use a higher value when shoppers tend to visit products multiple times before buying - common in fashion, sportswear, and travel.

BuyHistoryWeight
Boosts products the visitor has previously purchased. Increase this weight for stores where repeat purchases are likely, such as grocery, DIY, or B2B supply.

SimilarU2IBoost (look-alike boost)
Boosts products that are similar to what the visitor has already viewed, using twin analysis. Works well alongside RelatedU2IWeight.

RelatedU2IWeight (cross-sell boost)
Boosts products related to what the visitor has already purchased. Often used together with SimilarU2IBoost across different recommendation modules.

💡 Example: An outdoor retailer might combine a moderate SimilarU2IBoost with a higher RelatedU2IWeight to show similar jackets alongside complementary accessories like gloves or base layers.

MerchandisingBoosts
Lets you boost specific products based on business criteria of your choice. Common use cases include promoting high-margin products, items with high stock levels, or private label products.

👀 Use case: A retailer preparing for a seasonal clearance event can use MerchandisingBoosts to surface products that need to move, without overriding the personalisation logic entirely.