Segmentation

Segmentation models divide your audience further into groups called "segments". Segments allow you to break up your target audience into messageable groups of customers that share a similar attribute or interest that can be sent specific content as a whole. Customers are only able to be placed in one segment allowing the single customer to be sent content customized to their segment directly. If cases where a customer is qualified for more than one segment, the customer is placed in the first segment for which they qualify.

Every segmentation model has a Leftovers segment which includes all customers who do not meet any other criteria in any other segment. This ensures that all customers in the audience are included in exactly one segment. Once segments are defined, you are able to create specific messaged directed at each segment.

Note: The Leftovers segment does not appear in when you use the segment type based on "Equal Sized Segments Randomly Selected". This will be discussed later.

Segmenting your audience allows you to send different messages to different customers sharing the same interests or attributes (e.g. resident of California, fan of baseball). Many of the reports in Connect display results of each segment, allowing you to determine the segments that have the best response and conversion (purchases). You can also send the same message to more than one segment if the same message applies to multiple or all segments.

By using segments, you can target different messages to different groups of customers. As well, you can break down campaign reports by segment, allowing you to measure and compare the behaviors of the different segments. Segmentation models can even be used if sending the same message to the entire campaign audience. In this case, the audience is one single segment.

Segmentation models use the same filter types that are used in creating audience filters. In this case, where audience models use the filters to pare back the targeted audience from the entire conversation customer pool, segmentation models use the filters to partition the audience into groups. Segments can be grouped by:

Customer Attributes: Attributes associated directly with the customer entity (e.g., city of residence, gender)
Recency Statistics: Statistics tracking the customer's recent behaviour when receiving messages (e.g., links clicked, conversions from messages)
Frequency Statistics: Statistics tracking the overall interaction with the customer (e.g. total number of messages sent to customer, percentage of links clicked)
Monetary Values: Statistics tabulating the value of business represented by the customer (e.g. total value of customer conversions, total value of conversions from customer referrals)

Segments are also able to be created based on equal or custom-sized segments that are randomly selected by Connect. That is, segments can be created automatically and randomly filled with customers from the audience. If equal-sized segments are used, the number of segments are indicated and filled with audience customers equally. If custom-sized segments are used, segments are created and determined to be a certain percentage of the audience customers. In both cases, Connect places customers from the audience into the segments randomly.

For more information on how Connect uses segmentation models, please see the following: