Website personalisation: Product recommendation elements – Ometria Help Centre

Note: This guide is for retailers using Ometria's website personalisation offering - if you don't have this feature, you can still use Website product recommendations.

Create product recommendation elements for website personalisation.

Page elements for recommendations

You'll need a product page element set up and configured to display on the right page on your website - this might require development work.

We also recommend setting up a second page element with a broader strategy and filters to use as your Plan B (fallback).

💡 It's common for retailers to use the same product recommendation page element in multiple places throughout the site.

Select the Page Type > Product page:

Set up the default strategy for your page element as usual and select the recommendation type you want to use.

Recommendation types

There are two primary types of recommendations:

Type	Recommendation	Description
Crowd based recommendations	You may also like	Shows products based on other visitors' behaviour, e.g. people who viewed/bought this also viewed/bought that.
	Frequently bought together
Product based recommendations	Similar products	Shows products similar to each other, based on filters you set up.
	Complete the Look

Crowd based recommendations

Crowd based recommendations look at behavioural data on your website.

Note: If these recommendations aren't available to you, you might require further development.

Select configure to mix and match the behaviours you'd like to reflect:

When setting up crowd based recommendations it's important to think about how much data might be available for your products.

For example, newer products might not have as much purchase data available, and retailers with broad and diverse product ranges might end up with diverse recommendations.

Configuring your recommendations

For product based recommendations, we recommend using filters to control the products displayed, but you can also use any of the following to customise the look:

Merchandising blends - Define the order in which products display
Merchandising templates - Define the layout in which products display and lock certain products in place
AdSpots - Use adspots for inline ads and other kinds of content
Product groupings - Group sets of products together (e.g. clothing ensembles)

Manual merchandising

You can use manual merchandising to pin products and make sure they are always promoted in your recommendations.

To apply manual merchandising to your product recommendation, make sure you select the right page element in the Element field:

Set up filters for your recommendations

Use filters to control which products display in your recommendation elements.

For example, you can create a filter to display only products that are the same colour as the product the visitor is currently viewing.

Go to the Settings tab in your page element or select Configure if you're using a product based recommendation type:

Select +Add under Filters and select an Attribute to filter by, e.g. season, colour, product type, size, etc.

Complete the fields in the Filter Editor:

Field

Description

Condition

For most attributes, choose between:

Equals - displays products with the same attributes as the current item (e.g. the same colour, season etc.)
Does Not Equal - displays products with different attributes from the current item (e.g. a different colour, season etc.)

For numerical attributes, e.g. price, choose from:

Equals
Is Greater Than
Is Greater Than Or Equal To
Is Less Than
Is Less Than Or Equal To
Is A Range

Filter by the "current" item

Toggle this to Yes.

This means that the recommendations will display based on the product the visitor is currently viewing.

Create as many filters as you need.

💡 Remember that the more granular your filters are, the fewer products will be available to display.

Filter logic

Filter conditions use AND/OR logic, so you can choose for all filters to apply (AND), or any of the filters (OR).

You also have further options for AND logic:

Logic	Description
AND - match all	All of the filters you add will apply.
AND - fallback filtering	Set a threshold number of products, e.g. four. If Ometria can't find four products that match, it will ignore the last applied filter and try to find more. This process continues until four matching products are found.
AND - progressive filtering	Ometria tries to find all products that match all the criteria you chose, and then falls back to OR filtering after that. E.g. If you choose Red AND dresses, Ometria shows red dresses first, then dresses, then other red products, giving your customers more to choose from.
OR - match any	Ometria looks for products that match any the criteria you chose.

Set up a plan B

Plan B is a fallback you can set up in case no recommendations are found, or if the number of recommendations doesn't meet your set threshold.

Use the Plan B tab to call a completely different page element:

Field

Description

Fallback Threshold

Enter the minimum number of recommended products you'd like to display before reverting to Plan B.

E.g. If you enter 4, the fallback element will be used when Ometria can't find more than three products that match your conditions.

Fallback Element

Select the page element to use as a fallback.

We recommend setting up a page element with a deliberately broad filter, e.g. "equals current category" to guarantee products displaying.

You can use the same fallback element for multiple recommendation elements.

Additional Merchandising Strategies

Use this tab to set up additional merchandising strategies.

See: Merchandising strategies

You could create strategies based on different customer segments, e.g. show customers in a "high-spenders" segment luxury products.

Basket recommendations

To set up recommendations based on the products in a site visitor's basket/cart, work with your development team to set up a page element which is called when the basket is open.

Select the Similar Products recommendation type.

You can configure your basket recommendations in the same way as any other, but we recommend leaving your filters very broad to maximise the chances of products showing up.

Note: Basket recommendations are based on the most expensive item in the visitor's basket. This is because customers are more likely to buy products that 'go with' their most expensive purchase.

Home page recommendations

Home page recommendations are actually small product list pages (PLPs) with a limited number of products per element.

Note: Because they're not product page type elements, home page recommendations don't display recommendations based on products. They show products based on the filters, blends and strategies you set.

First, you'll need a page element set up and configured to display on the home page on your website - this might require development work.

Select the Page Type > Home page:

Follow the steps for setting up a page element and define the merchandising blends, filters etc. as required.