Link Search Menu Expand Document

Creating an Optimizer

In this page you have a grid displaying all your optimizers. You can add a new one by clicking on the Add New Optimizer button.

You will then be prompted to the optimizer create form containing the following fields :

searchoptimizers_general

Parameter Default value Description
Enable optimizer Yes If the optimizer should be enable or not.
Store View   The store view for this optimizer. An optimizer can only be assigned to one store view.
Model Constant The model to use for scoring :
- Constant : to apply a constant score
- Based on attribute value : to apply a boost that will be proportional to an attribute value
- Based on behavioral data : to apply a boost that will be proportional to the selected metric (Details bellow)
Optimizer Name   The name of the optimizer.
Active From   Start date of the optimizer. Use it for temporary events.
Active To   End date of the optimizer.
Request Type   The request to apply the optimizer :
- Catalog Product Search : the catalog search results.
- Catalog Product Autocomplete : product results in the autocomplete
- Category Product View : the catalog navigation.
- Quick Order Suggest Search : product results in the quick order suggest search.
- Related Products : products in the related product blocks.
- Upsell Products : products in the upsell product blocks.
- Cross-sell Products : products in the Cross-sell product blocks.
- Visitor Products : products in the visitor product blocks.

searchoptimizers_rule

This fieldset figures a rule editor where you are able to chose any combination of conditions you want to match.

Eg :

  • “Color is Blue” to apply a boost on products having the value “Blue” for the “Color” Attribute
  • “Only discounted products” to apply a boost on products actually having a special price applied
  • “Only in stock products” to display immediately salable products on top of your product list and so on …
Optimizer based on metric (behavioral)

boost-metric

Parameter Default value Description
Metric   Select the metric the boost will based on. Boost can be based on daily views, weekly views, sales, daily sales, weekly sales, conversion rate and daily conversion rate. For every metric, you can select the average metric or the exact number
Boost impact   Low
Medium
High
Boost value   Value for the multiplication

Calculation is as follow :
Low : log (attribute value * boost value). For example : Boost value = 5

  • Metric = 100 –> product score will be multiplied by log(5*100) = 2,70
  • Metric = 5000 –> product score will be multiplied by log(5*5000) = 4,40
  • Metric = 8000 –> product score will be multiplied by log(5*8000) = 4,60


Medium : √ (attribute value * boost value). For example : boost value = 5

  • Metric = 100 –> product score will be multiplied by √(5*100) = 22,36
  • Metric = 5000 –> product score will be multiplied by √(5*5000) = 158,11
  • Metric = 8000 –> product score will be multiplied by √(5*8000) = 200


High : attribute value * boost value. For example : boost value = 5

  • Metric = 100 –> product score will be multiplied by 5*100 = 500
  • Metric = 5000 –> product score will be multiplied by 5*5000 = 25000
  • Metric = 8000 –> product score will be multiplied by 5*8000 = 40000