Skip to main content

ES Loyalty Boost 2.7.1 release notes

note

These release notes cover features introduced or enhanced since ES Loyalty Boost 2.7.


What's in this release


Enhancements

Audience Enhancement

note

The Audience features in ES Loyalty Boost v2.7.1 require ES Loyalty v4.1.1 or later to be installed.

Overview

Previously, Boost used its own Audience Selector — configured on a tab in the Excel Offer Pool file — to define inclusion logic at the Offer Template and Slot levels. This was used exclusively for Boost purposes.

This enhancement replaces that method with the ES Loyalty Audience feature set. Users can now access active ES Loyalty Audiences when configuring slot-level audience inclusion logic, and assign an Audience to an Offer Template directly via the Excel Offer Pool.

How targeting logic works

When selecting offers, the effective audience is the intersection of the audience defined at the offer level and the audience defined at the slot level. The following examples illustrate how this works:

Slot-level audienceOffer-level audienceResult
EveryoneEveryoneEveryone is included
Employees onlyEveryoneEmployees only
EveryoneCherry pickers excludedEveryone except cherry pickers
Everyone except employeesEveryone except cherry pickersEveryone except employees and cherry pickers
Ontario residentsEveryoneOntario residents only
Ontario residentsQuebec residentsNo members included — the two audiences cancel each other out
Employees onlyNon-employees onlyNo members included — the audiences nullify each other
Members linked to a financial partnerMembers not linked to a financial partnerNo members included — the audiences nullify each other

Details

This feature improves functionality and the user experience for Console users with access to ES Loyalty Boost, Audience Builder, and Audience Recommender. Audiences can be assigned to Offer Templates via the Excel Offer Pool to automate and optimize marketing tactics through self-service.

This feature also enables use of more advanced Recommender Audiences that were not previously possible in Boost. Note the following:

  • Only audiences in the same business unit are available.
  • Audience names are listed in alphabetical order in the dropdown.
  • Audiences generated using Gen AI through Audience Builder are marked with a star icon (✦) next to the audience name.

See also:


Global and Offer-Level Control Configuration Feature

The treatment for the global control group and offer-level control group is now configurable per client. Configuration is set at the account level rather than the campaign level and is managed by Exchange Solutions rather than the client directly.

Global control group — supported modes:

ModeBehavior
NO_OFFERNo offer is generated in the campaign_targeted_offer table for the account
RANDOMMembers receive randomly assigned offers

Offer-level control group — supported modes:

ModeBehavior
NO_OFFERNo offer is generated in the campaign_targeted_offer table for the account. The normal control offer used for measurement is still created but is not visible to the member.
REPLACEMENTA replacement non-competing offer is provided when an offer is selected as a control offer

See also:


Infrastructure

Scaling — Qualify Performance Envelope

A new ES Loyalty environment was created under a DevOps account to establish a production-like baseline for capturing accurate performance metrics. All associated backend stacks — including Snowpipe, Analytics, and Boost — were deployed to this environment.

Test configuration:

Performance tests were run using the following parameters to simulate realistic production conditions:

  • Large-scale member audience
  • Large product catalog
  • Up to 50 transactions per member
  • A realistic number of generated offers
  • Validation that offers were correctly created from the campaign for a sample set of users meeting the offer selector criteria

Tests began with a smaller audience and scaled up to conditions representative of the largest clients by member base and offer volume. Each member received 10 offers (10 slots) per iteration.

Additional factors accounted for in testing included sharding policy, batch size, payload size, and replacement offers.

Sample audience breakdown (Phase 1 — 3,000,000 total members):

EmployeeLinkedCherrypickerAgeProvinceTotal
YY< 30, 31–50, > 50ON, BC4,500
YN< 30, 31–50, > 50ON, BC25,500
NL< 30, 31–50, > 50ON, BC300,000
NNY< 30, 31–50, > 50ON, BC3,000
NNN< 30, 31–50, > 50ON, BC2,667,000
Total3,000,000

See also: