Gone are the days of displaying the same product recommendations to everyone. Now, you can deliver relevant recommendations to visitors of your website that factor in the products they liked, viewed, or purchased in the past.

With product recommendations, you’ll discover new potential for an increase in sales and engagement. Your visitors will experience less frustration searching or manually hunting for products they need.

Why start using product recommendations

Product recommendations is a marketing strategy that lets you display products your visitors may be searching for. Once your developer configures product recommendations on your website, you’ll be able to:

• Market to individuals
  Combine current and past behavior to recommend products based on each user’s context, page views, purchases, and more.
• Customize models
  Decide what inputs, algorithms, lookback timeframes, and goals matter most to your business.
• Influence results
  Use business rules to fine-tune, filter, and modify the recommendation results before you present them to your visitors.
• Integrate your product catalog
  Your developer can use Personalization’s APIs to access the latest details on inventory, pricing, categories, and custom variables.

All this is possible thanks to algorithms, models, and business rules you can apply to define and fine-tune your recommendations.

How product recommendations work

In product recommendations, we use flexible machine-learning algorithms that combine real-time understanding of individual customer behavior with browsing history to deliver spot-on experiences centered around data and not opinions.

As you gather data about your customers’ purchase behaviors, you can:

• recommend products they can place in their shopping cart as an addition to what’s already in it
• suggest products similar to the ones they viewed in the past
• show products other customers, with similar interests, viewed or bought
• present items that can replace the contents of their cart

Customers will see your recommendations in designated and pre-defined areas of your website, called zones.

Set up product recommendations

1. It’s your developer that configures product recommendations for you. In this step, they:
   • create and upload the product catalog
   • configure Acoustic Exchange and your analytics library to work with Personalization
   • deploy the Personalization library
   • set up zones
2. Once these are complete, you’re on your way to creating your first product recommendations. To begin, think about the common behaviors of visitors on your website. You’ll need to map these behaviors to algorithms you can add to models in your strategies. Then, you need to think about business rules you want to add to fine-tune your recommendations. Check out Models, algorithms, and business rules.
3. Your next step will be configuring product recommendations strategies. You'll start with a primary strategy and then move on to a fallback strategy. Why would you need two strategies? You can use the fallback strategy when there aren’t enough results generated with your primary strategy. You’ll also connect your strategies to the rules you define.
4. Finally, go on to previewing your strategies before you publish them.