Recommendations

With the Recommendations app you can build customized lists of relevant product recommendations using what Bronto knows about your products and contact behavior. These recommendations can be used to build highly-targeted email messages and campaigns that contain products relevant to your contacts.

Important: If you do not have the Recommendations app and would like information about purchasing it, see Bronto's App Center or contact your account manager.

A recommendation is a list of items from your product catalog that meet a set of criteria that you define as you build a recommendation. When you use a recommendation in a message, what products best match its criteria are determined at the time a message is sent.

There are two versions of the Recommendations app - Standard, which can build recommendations based on what Bronto knows about your products and Premium, which can build recommendations based on what Bronto knows about your products and contacts' order and browse history.

Recommendations Standard

Recommendations Standard helps you to build recommendation lists based on what Bronto knows about your products. This includes things like a list of your top-rated products, new arrivals, or best sellers. For example, with Recommendations Standard you could:
  • Build a list of high inventory items in order to promote them when you want to move inventory. Or you could build a recommendation list using the Sale Price Effective Start Date in order to promote an upcoming sale.
  • Use recommendations to automate the inclusion of product data in daily or weekly promotions.
  • Build a recommendation that includes small quantities of high priced items and use it to build a promotion geared towards contacts on a high spender segment or list in order to try to close out your stock.
Important: Recommendations uses the product data you have already imported into Bronto. If you need information about importing product data see Products.

Recommendations Premium

Recommendations Premium helps you to build recommendation lists based on what Bronto knows about your products and your contacts' engagement data. This intersection of contacts' order and browse history combined with product data allows you to create recommendations that are highly personalized based on contact behavior. For example, with Recommendations Premium you could:
  • Build a recommendation of products that contacts frequently also buy when they have bought another product and send this recommendation to contacts who have bought the first item.
  • Use information about which products are frequently ordered together to build recommendation you can use in a "Don't forget!" message.
  • Build a recommendation of products that a contact is most interested in using the contact’s most recently browsed and most recently ordered products.

Before You Can Begin

Recommendations uses data stored in Bronto, so the quality of a recommendation can only be as good as the data it's built upon. Therefore, before you begin using Recommendations you should:
  • Import a product catalog and implement a plan for keeping your product data up-to-date. The more robust and accurate your product data is, the better your recommendation results will be. See Products for more information.
  • If you have product variations, make sure that your catalog uses Parent Product IDs to link variant products. A parent product is the main product record for a group of products that are variations of each other. The parent product should not have a value saved in Bronto for its Parent Product ID. The Product ID for the parent product should be stored as the Parent Product ID for all of its variants. Each variant will have its own unique Product ID. Using this parent product and variant model allows Bronto to see more in depth connections between products, which leads to more sophisticated recommendation results.
  • Import historical product data into Bronto in order to create a rich set of data for Recommendations to pull from. These historical records should have archived status in Bronto so they are not used in recommendations.
    Tip: Bronto automatically archives products when product data that was previously imported is not in the most recent import file. So to archive historical records, import a data feed that contains your historical product data and then immediately import the product data feed that only contains your current product data.
For Recommendations Premium you should also:
  • Configure Bronto to capture contacts' browse activity. Recommendations can use up to 90 days of a contact's browse data. See Capture Browse Data for more information.
  • Make sure you're using Bronto's order service and check that you've properly configured Bronto to capture order data.
  • Verify that the product IDs coming from your order data match the product IDs in your catalog.
  • If you recently configured order service, import historical order data into Bronto. Bronto can use up to 5 quarters of order data when building recommendations so the more data you have the better the recommendations will be.

Building Recommendations

Bronto provides recommendation templates that cover common use cases. These templates can be used as a foundation to build more complex recommendations. But you aren't limited to just these scenarios - any product field in your product catalog can be used to build general recommendation criteria.

With Recommendations Premium, you also have the option of building personalized recommendations using Bronto predictive models and/or personalized settings. The available predictive models are defined in Create A Premium Recommendation. Personalized settings are special filters that are applied using a point-of-reference product or contact. For example, you might want to limit a list of recommendations to products that are the same brand as your reference product. To do this you would set Related Products criteria and include the reference product in your message or configure settings on the Reference Products tab of the recommendation as you create it.

The results of a recommendation automatically update as changes are made to your data in Bronto. Therefore, the more up-to-date your product, order, and/or browse data is the better your recommendations will be.

Draft vs Published

Any recommendation can exist in up to two versions:
  • Draft: A draft recommendation reflects the most recent edits to a recommendation, but a draft recommendation cannot be used until it is published.
  • Published: A published recommendation is the version of the recommendation that can be used. You can edit a published recommendation, but this creates a draft version of the recommendation where your edits will be saved until the draft version is published. A published recommendation can be made either active (available for use) or inactive (not available for use).

Each recommendation can only have one draft and one published version. When you publish a draft, the previously published version of a recommendation is overwritten.

Published recommendations can be included in messages using rec loops. The product information that corresponds with your rec loop settings will be inserted into the message when it is sent. Replacing rec loop placeholders at the time a message is sent allows Bronto to provide up-to-date recommendations. See Recommendation Messages for more information.