A sequence is the order in which products are organised when an end-user interacts with a category page. Sequences can be constructed using a sorting rule, and multiple highlights or manually added products. Sequences can either be targeted towards your entire audience base, or towards individual segments.  

Our real-time segmentation allows you to create unique sequences that vary between segments according to different wants and needs. Sequences can be configured to target either your entire audience base or individual segments. By combining aspects of both personalization and merchandising rules allows ecommerce brands to create category page sequences that are both more engaging and built to accomplish the goals they want.

Merchandising rules is a top level term used for both sorting rules, and highlights. Merchandising rules are reusable configurations that can be created and applied to one or multiple categories. Merchandising rules are built by using different combinations of Merchandising rules fall into two different groups based on what use-case is intended. 

Sorting rules allow you sort an infinite amount of products in a specific order according to different product attributes and performance metrics of your choosing. When using sorting rules you can select one or more product attributes or performance metrics and assign specific weighting to each of them according to your goals. By adding personalization rules on top of these merchandising rules, you are able to create highly relevant category page sequences that take customer behavior into account but always keep your business goals top of mind when sorting products.

Learn how to use Sorting Rules here: Introduction to Sorting Rules, and how do i create them?

Highlights are merchandising rules that allow you to permanently place products at the top of the category page and hero them. This can be used to create end of aisle display-like experiences on category pages by ensuring certain products are highlighted in the most prominent or engaging spots. The business cases are endless; retailers can optimize toward an assortment of different business goals like increasing profitability, liquidating surplus inventory, and much more. Much like Sorting, a combination of product attributes and performance metrics can be used to curate the order in which these products are displayed.

Learn how to use Highlights here: Introduction to Highlights, and how do i create them?

Product attributes are descriptive fields for products such as Brand, Category, Price, Supplier cost etc. These fields describe the product in one way or another and using them for merchandising rules allow you to group similar kinds of products together. 

Example: Promote products with a high price, or demote products with a low price, and adjust the weighting between different attributes to match your business needs.

A full list of available Product Attributes and Performance Metrics are available in this article: Product Attributes and Performance Metrics

Product performance metrics are calculated based on behavioral and transactional data, and consists of relative data points such as conversion rate, add to cart rate, inventory throughput etc. Using these fields will allow you to showcase products which are either performing poorly or well on certain measurement points. 

Example: Promote products with a high conversion rate, or demote products with a low conversion rate, and adjust the weighting between different attributes to match your business needs.
​
A full list of available Product Attributes and Performance Metrics are available in this article: Product Attributes and Performance Metrics

You can define weights for product attributes or performance metrics on an attribute to attribute level. When configuring sorting rules and using the sliders, products are always evaluated on a scale from 0 - 100 for each individual attribute. This means that a shirt can be on scale 100 for the attribute price, and on scale 60 for the performance metric conversion rate. The weight defines which of the attributes are more important when compared against each other. 

Read more about this here: Ranking Score and Weighting Factors

When configuring sorting rules with multiple attributes and weights, a total match score will be calculated based on how well the products score on the given scales. The total match score will be represented by a percentage score next to the product in preview. For example weighting price and conversion rate against each other, one of the shirts might have a match score of 93%, whereas another product might have a match score of 54%. A very high match score is expected when using only one attribute, or a textual field such as brand or category.  A low match score does not necessarily mean that the product is irrelevant, but is a likely causality of adding multiple weighting factors which diminish the total score exponentially. 

Read more about this here: Ranking Score and Weighting Factors

Sequence testing can be used through A/B Testing and Optimization which allows you to test your category merchandising sequences against each other. This means that for example, you can test two sorting rules for one category or a combination of sorting rules, highlighting rules and pinned products for a sequence against another. Firstly, before setting up a category merchandising sequence test, you have to have at least 2 variations of a specific sequence setup - or this won't be visible to test under A/B testing at all. Variations will fall under a specific segment for that category sequence such as the "everyone" segment. One segment with however many variations will then be chosen to sequence A/B test in the set up.
For example to set up a variation to test, you need to go to the category sequence itself and then click the 3 dots as so where it will then say "create new variation".
​

You can then set up the new variation for your category sequence for example as below with a name and choose an existing sort rule/ new rule/ pin products.

An example of setting up a sequence test is below - with the category for "bags" and the "everyone" segment that has 2 variations against this to choose from.
​
​

Sequences can be tested against each other only by manually defining the traffic split between differing variations.

One other aspect to bear in mind is that currently - the preview debug toolbar will show an A/B sequence test - however, this is not currently supported. Although, you can be sure that the conditions you have set for the variation of your sequence will be the one that's showed within an A/B sequence test and once the sequence test is live, the data will start to appear. You can also start to see data on the variations when checking in the sequence for the category itself in the category merchandising module and setting the date to that day onwards e.g.
​
​

It's important to set up different sequence testing, so you can define the best sorting rule method for each segment in each category, based on your strategy of conversion rate, CVR or AVV.