Skip to main content
Price affinity predictor

Price Affinity Predictor uses Artificial Intelligence (AI) to predict the price band that a new visitor is most likely to go on to buy at.

Updated over 5 months ago

The Price Affinity Predictor is best utilised on your website homepage, when making product recommendations to visitors that you’ve never seen before. It can also be used in conjunction with other tactics, such as popular products.

The Price Affinity Predictor can be used:

  • to filter product recommendations on your homepage, so you can show products to a cold start visitor that are more likely to be at a price point that they are interested in.

  • to drive homepage content, so you can change the content displayed depending on the AI outcome of high, medium or low price point affinity.

Both of these options are aimed at reducing the bounce rate of new visitors and increasing your return on investment for new visitor acquisitions. Price is a key influencer to visitors bouncing.


Before you start

Things you need to know:

  • Price Affinity Predictor may not be available on your account straight away. Our system assesses your account for suitability with the AI, for example, that your product price data is across a sensible range and that the volume and quality of data is adequate. If the account is assessed as suitable, the Price Affinity Predictor is automatically activated. If you're unable to use the Price Affinity Predictor and would like to discuss this further, contact your account manager.

  • You should be familiar with how to create a new SmartBlock and Slot.
    Learn more in Create and edit a SmartBlock and Personalize your triggered messages and site content.


How Price Affinity Predictor works

The AI uses the large volume of data we collect across our system to create many data models that predict which price band a person may go onto purchase at. The AI regularly tests each available model against your site's data, choosing the best model to use at any particular time, so you always see the most benefit from the tool.

The AI performs sophisticated analysis on all the products on a website and automatically puts each product into a price band of High priced, Medium priced and Low priced. It also reviews this at intervals, and may automatically shift products between bands, adapting to price changes and sales.

Settings

The only configurable part of the AI is being able to remove product categories from the data it evaluates to determine the most successful model to use. This is useful if you have a product category with very different price bands, such as price extremities or where the price is calculated in a different way, maybe by weight, and this data is likely to negatively impact the AI.

To add product categories for removal, expand the User menu and go to Settings > System Configuration > Artificial Intelligence Settings.

Use Price Affinity Predictor for product recommendations

For best returns we suggest that Price Affinity Predictor is used mainly on your website homepage and for those visitors that you have no, or little, information about.

To show products that are predicted to be in a price range of interest to the website visitor:

  1. Create a Recommendation SmartBlock with your required layouts and tactics.

  2. Edit the SmartBlock, select the Filtering tab and, under Products, add the Price Affinity Predictor filter.

  3. Add your Recommendation SmartBlock to your homepage. Slot rules or Experiences can be used to only show this SmartBlock to new visitors.

By adding the Price Affinity filter to your recommendations, the price point of items shown are more likely to be of interest to your cold start visitors.

This filter is normally applied alongside the tactic or data source being used, such as New products or Frequently purchased products, and any other filters being applied, such as Must be in stock or Have a high review rating.

Use Price Affinity Predictor to determine content shown

To show content based on the Price Band that the Price Affinity Predictor thinks a person is most likely to go onto buy at:

  1. Create your different SmartBlock content for high, medium and low price points. You need a SmartBlock for each price band that you want to show different content for.

  2. Create a Slot, and under the Slot Rules, for Person, add the Purchase Price Prediction rule to the decision tree on the right.

  3. Under the configuration of the rule, select if the content should show for High, Medium and/or Low price predictions, and then add the SmartBlocks you’d like to show for this rule.

The Price Affinity Rule can be added multiple times, to show different content for each predicted price band, and can be combined with any other rules that are available.

How do I see if the AI is working?

The best way to test the impact that the AI is having is to run a split test. This can be done using Slot rules or Experiences by dragging your control SmartBlock on top of your AI SmartBlock. You can then view data for the two through the Comparison report.

Did this answer your question?