Custom rules and filters
Both branching controlled by slides and branching controlled by tree can be set to recommend products based on rules. Through the use of rules, you can define which products will be recommended in a fashion similar to natural language. For example, you can say things like: if the user selects “fashion” and “flat,” then matching products will be recommended. If the user selects “fashion” and “heel” or “comfort,” but not “expensive,” then the other matching products will be recommended.
Before defining rules, you have to define the valid terms that will be used to make up the rules. For example, in the rule “flat and not slide,” the terms will be “flat” and “slide.” Those terms have to be mapped to answers so that Pickzen knows which terms are met and not met.
To map a term to an answer, edit the answer settings and go to the advanced tab. There, you can enter a Term ID for that answer. For example:
In this example, the “Flat” answers will be mapped to the “flat” term.
Activating rules mode in branching controlled by slides
To activate this mode, enter general settings and check “Customized rules” in the advanced tab. When complete, step three in the breadcrumb will become highlighted:
You can define your questions and products as usual and in step three you can create the rules. This way, once the user answers the questions, the recommended products will be selected based on the defined rules.
Activating rules mode in tree-controlled branching
To activate this mode, click “Show advanced editor” in the endpoint editor:
This way, the products recommended at the endpoint will be determined based on the defined rules.
Once all of the terms you want to use in the rules are defined, you can define the rules in a panel similar to this one:
The rules are defined in rule groups (and each rule must contain at least one rule group). In the bottom beige panel you can see the defined terms and operators. Products will be recommended based on the first rule met in each rule group. This means that Pickzen considers the rules of a group from top to bottom. When a rule is met, its defined products will be added to the recommendation list and it will continue on to the next group (if applicable).
For example, if we had this rules group:
- If the user answers “flat,” the 6 defined products will be recommended.
- Otherwise, if the user doesn’t select “flat” but “heel” and “expensive,” the 3 defined products will be recommended.
- Otherwise, if the user answers “heel” and “economic” and either slide or not expensive the 2 defined products will be recommended. In this last case, the use of parentheses is mandatory because the “and clause” has preference over the “or clause.” If the parentheses were omitted, the meaning of the rule would be: “heel” and “economic” and “slide” or “not expensive.”
- Lastly and optionally, if no rules are met in a group, we can use the special “any” rule as a default rule, so their products will be recommended.
As mentioned above, if more groups are defined, they would add their recommended products to the final recommendation list.
If you’d like to recommend variants, sizing for example (S, M, L), you’ll want to recommend products according to the size the user selects and normally, you’d need a to create lot of rules because each rule would define the products to recommend a unique size. If you also wanted to recommend by “color” (an additional variant) even more rules would be necessary as for each size you’d also have several colors.
To recommend variants most easily, you can use filters, found in the second tab. The rules select all the products to recommend (including all variants) and filters retain only those products that meet the variants.
For example, if based on rules, the products determined to be recommended are the following:
- Nike Air Max S
- Nike Air Max M
- Nike Air Max L
- Adidas Prohere M
- Adidas Prohere L
We can set the filters as:
This way, in the size answer slide, all products of that size will be included.