Segmentation

Overview

Segmentation is the ability to isolate similar groups of customers, based upon demographics or demonstrated patterns, and selectively display recommendations to these customers. A segment is a group of customers with a common demographic or demonstrated pattern to which content can be targeted based on affinities and patterns specific to that group.

For example, an office supplies retailer might define segments such as:

• Frequent ink, toner, paper purchaser

• Frequent computer purchaser

Merchants can use segments to push more personalized marketing to segmented customers. When setting up segments, ensure that they are clearly defined and avoid setting up users in many different segments as the results might not be as expected.

If the merchant has segments built, these segments can be leveraged to:

• Create merchandising rules on the Omnichannel Personalization dashboard to target segments.

• Take advantage of segment-based strategies for more granular and personalized recommendations.

Segmentation in Recommend

Recommend offers a multitude of strategies for tailoring the customer experience. Each strategy leverages merchant catalogs and customer behavior to recommend products. A traditional strategy may show recommendations based on products viewed by other customers. If segment information is provided to Omnichannel Personalization, this information can be used to understand the behavior of a subgroup of customers on merchant sites.

For example, if segments “gender” and “age” are given, Omnichannel Personalization uses segment-based strategies to generate more granular and personalized recommendations (for example, if a woman is shopping for a sweater, she will see recommendations for sweaters that other women within a certain age group have shopped). Algonomy believes that it allows for increased customer engagement and revenue.

A segment that is assigned to a customer remains in their history and viewed by the segment evaluator. This is separate from viewing the User Profile on the dashboard, where only the fifteen most recent events are viewable.

Strategies available for Segmentation

InSegment strategies require segments that are passed or created in one of the following ways:

• Segments passed in the instrumentation or through the recsForPlacement API, also known as "user segments". These take precedent over other types of segments. The InSegment strategies look at instrumented segments if available, before moving to other types of segments.

• Segments created on the dashboard through the segment builder, also known as "custom segments" or "dynamic segments."

• Segments uploaded in batch.

Recommend currently doesn’t exclude specific segments from evaluation by a micro-segment. However, segments built in Segment Builder are the last to be reviewed by strategies. So, if the customer belongs to an instrumented segment or batch-uploaded segment that has enough data behind it to fire, the Segment Builder segment will be suppressed.

Algonomy has built the following segment-based strategies to align with how retailers market differently to various segments.

Strategy Name	Strategy Description and Message
TopSellersInSegment	Returns the top-selling products among users in a given segment. Strategy Message: “Shop <Merchant_Name> Top Sellers In <This Segment>”
CategoryTopSellersInSegment	Keys off one or more categories and returns the most frequently bought products in the categories for the given segment. Strategy Message: “<Category_Name> Top Sellers In <This Segment>”
PurchaseCPInSegment	Returns an ordered list of recs based on likelihood of purchases from shoppers who bought the seed product in the same segment. Strategy Message: “People Who Bought <This Product> Also Bought Products In <This Segment>”
BrandTopSellersInSegment	Returns the top-selling brands among users in a given segment. Strategy Message: “<BRAND> Top Sellers In <This Segment>”
ClickCPInSegments	Keys off of a specific product and segment and returns products most frequently viewed given a view of the current product; informs the customer of substitutable items, enabling comparison shopping on a per segment basis. Sample Message: “People In <This Segment> Who Viewed This Product Also Viewed…”
ViewedPurchaseCPInSegments	Keys off a product and a given segment and return products likely to be purchased by viewers of the context product in the same segment. Sample Message: “People Who Viewed <This Product> Ultimately Bought In <This Segment>”
PersonalizedViewedPurchasedInSegments	Products that are frequently bought with other products that the shopper has recently viewed (based on recent history) within a given segment. Sample Message: "People In <segment> Who Viewed <this product="product"> Ultimately Purchased></this></segment>
PersonalizedClickCPInSegments	Products that are frequently viewed with products that the shopper has currently viewed in a previous or current session within a given segment. Sample Message: "People In <segment> Who Viewed <this product="product"> Also Viewed"</this></segment>
MultiItemPersonalizedClickCPInSegments	Products that are frequently viewed with multiple products that the shopper has currently viewed in a previous or current session within a given segment. Sample Message: "People In <segment> Who Viewed Products In Your Recent History Also Viewed
PersonalizedPurchaseCPInSegments	Products that are frequently bought with a product that the shopper has recently purchased within a given segment Sample Message: "People In <segment> Who Purchased Products In Your Recent History Also Purchased"
MultiItemPersonalizedPurchaseCPInSegments	Products that are frequently bought with multiple products that the shopper has recently purchased within a given segment Sample Message: "People In <segment> Who Purchased Products In Your Recent History Also Purchased"
LeastExpensiveViewedPurchasedInSegment	The lowest priced products of those that are frequently bought with the given product within a given segment. Sample Message: "People In <segment> Who Viewed <this product="product"> Ultimately Purchased></this></segment>
BestSellingNewProductsInSegment	Products that are the bestselling of new arrivals within a given segment. New Arrival and New Product strategies have a lookback setting that can be changed in site configuration, and is default set to 31 days. Sample Message: Top Selling New Arrivals In <segment>
NewArrivalTopVIewsInSegmentStrategy	Products that have arrived in the last 30 days (by default) and have the most visits for last 7 days for user segment based on viewed Sample Message: Top Viewed New Arrivals In <segment>
TopViewsInSegmentStrategy	The most visited products for last 7 days for user segment based on views Sample Message: Top Viewed In <segment>
NewToYouInSegment	The top purchased products users have not already seen according to that user's profile. Sample Message: New To You In <segment>

 

A strategy can have 1 or 2 hints. The first hint is coded in the strategy message with this variable name “$STRATEGY_HINT$”, the 2nd hint is coded with “$STRATEGY_HINT2$".

Strategy	$STRATEGY_HINT$	$STRATEGY_HINT2$
CategoryTopSellersInSegment	category	segment name
TopSellersInSegment	segment name
PurchaseCPInSegment	product shown	segment name
BrandTopSellersInSegment	brand	segment name
ClickCPInSegments	product shown	segment name
ViewedPruchaseCPInSegments	product shown	segment name

 

The strategy hint is the segment name as it is passed in the instrumentation.

R3_COMMON.addSegment('[segmentId]', '[segmentName]’);

If the segment name is not passed, then the hints display nothing.

Requirements to Enable Segmentation

For segments to be supported by segment-based or InSegment, strategies, segments can be:

• Passed through instrumentation or when calling the recsForPlacements API.

• Created through the Segment Builder on the Algonomy dashboard (Custom or Dynamic Segments)

  • In site configurations, enable/disable whether these segments are used by InSegment strategies with the "Use dynamic segments by strategy" configuration.

• Uploaded in batches through Algonomy's batch upload service.

Instrumentation changes are not necessary to use InSegment strategies. Segments do not require a user ID. An entry is created in the User Profile Service and dynamic segments leverage UPS. The only requirement is that there is a stable user identifier, it does not need specific user ID.

Instrumentation

To use a customer's segment in recommendations, additional parameters must be passed in the request for dynamic content from Omnichannel Personalization.

R3_COMMON.addSegment('[segmentId]', '[segmentName]');

Note: This should be done on every page for a segmented user.­­

API

Use the sgs parameter to send the customer's segment every time you call recsForPlacements.

Passing Multiple Segments

For both Instrumentation and API, a user can be assigned to multiple segments, and those segments can be passed and used by recommendations. The order in which they fire is right to left. To put it another way, the last segment passed goes first.

Example: User A is part of Returning Visitors segment and Loyalty Customers segment, and those are passed as: Returning Visitors, Loyalty Customers.

The result is the Loyalty Customer segment gets leveraged first, and then the Returning Visitors segment.

Using Segments in Email

InSegment strategies can be leveraged in email recommendations. Specific parameter is not available for segment ID, but if you pass sgs parameter in the mail API, it passes through and is leveraged. The caveat is, if you try to do the same call with different segment IDs, the first call works as it gets cached.