Customer Segmentations: Bloomreach vs. Braze and Klaviyo
Explore the key differences in customer segmentation between Bloomreach Engagement and leading alternatives like Braze and Klaviyo.
About the Author:
Co-founder of Datacop, a digital agency that fulfills marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.
Email and CRM marketers spend a significant amount of time identifying meaningful segments within their customer base, based on the available data they have about them. The identification and use of these audiences in marketing channels are necessary for successful targeting and personalization. To address this need, many digital marketing automation platforms provide no-code interfaces that enable users to build and target specific audiences based on available customer data.
At first glance, the Segmentation features of Klaviyo, Braze, Attentive and Bloomreach seem similar. All enable marketers to build dynamic logical rules that define customers' membership into a particular “list of users” or a segment. These logical rules are recalculated as the platform's database updates with data such as user behavior events and user customer attributes. Once built, the customer segments become available in the campaign builder sections of each tool to specify the target audience for emails, SMS, banners, etc.
(Note that the recalculation speed between platforms may vary)
Common features of Segmentations across all platforms include
an easy-to-use “customer filter logic builder”
the use of logical combinations “Logic A AND / OR Logic B”
the ability to use negation logic “user has not done A”
the use of parenthesis “(A OR B) AND (C OR D)”
re-calculations of the size of segment membership based on current rules
It is not easy to distinguish between the platforms at first sight. The platforms also often look alike. To see for yourself, check out these screenshots from the documentation of Klaviyo/Braze and Attentive, which show the interface of their respective Segment features.
At Datacop, over the years, we have worked with all 4 of the platforms. We have found there are two key differences between the Bloomreach Engagement Segmentations and the more generic Segmentations offered by platforms like Klaviyo / Braze / Attentive, which are that…
1. Klaviyo / Braze / Attentive segments can only “hold one segment”
In Klaviyo / Braze / Attentive, the segmentations feature only supports Single-lists!
That means that to maintain an Email Health Segmentation with 10 segments, it would require building and maintaining 10 separate segmentations.
If these segments have a risk of overlap, they must be manually identified and resolved by adding exclusion logic, which adds mental workload on the marketer. We will go into more detail about an example of what this looks like later on in this article.
For e-commerce companies that have scaled beyond basic segmentations, the limitation of a single-list audience presents a lot of extra effort to implement a campaign and increased risk of error compared to segmentations that can hold multiple segments at once.
2. Bloomreach Segmentations are More Sophisticated and Flexible
Unlike typical marketing automation players, such as Klaviyo, Braze, and Attentive, Bloomreach Engagement enables businesses to create mutually exclusive, multi-segment audiences. Mutual exclusivity means that a single user within a Segmentation can only belong to one of its sub-segments. That means a database of, say, 1 million subscribers can be divided into groups that are both mutually exclusive and cumulatively exhaustive. For larger e-commerce brands that cater to diverse customer types and have extensive data to segment them, having access to multi-segment audiences can simplify the building of complex segmentations.
The functionality to host multiple segments in a single segmentation means that:
Building audiences with overlapping customer membership is far easier to execute for marketers than using “single-list” segmentations.
Building mutually exclusive and cumulatively exhaustive segmentations enables building important user analytics such as CLV and A/B test Reports directly in the platform.
In the rest of this article, we will dive deep into the above and illustrate why we prefer to work with Bloomreach Segmentations in Datacop over the typical “single-segment” Segmentations available on the mark-tech e-commerce market.
What are the advantages of having multi-segment Segmentations?
1. Reduced Effort & Mutual Exclusivity
Audiences can often have multiple overlapping outcomes for a customer, or there can be multiple outcomes for a customer. Consider the following examples:
Email Health Segmentation (e.g. New, Active, Passive, Lapsing, Lapsed, Inactive)
Gender-affinity Segmentation (e.g. Male / Female / Kids)
Category-affinity Segmentation (e.g. Shoes / Tops / Hats )
When building such segmentations, having the ability for a single Segmentation to host multiple segments means that, while normally for each possible outcome it would be necessary to build a separate segment object in the platform, in Bloomreach Engagement there will be one object to maintain “housing” all of the possible outcomes of an audience in one Segmentation. This saves time and reduces the scope of error when changing and maintaining their Segmentations and their application in campaigns.
Without the Bloomreach Engagement’s feature of mutual exclusivity in multiple segments, when building audiences where overlap is possible, additional conditions need to be added to resolve possible conflicts in overlap.
Consider the example of the Email Health Segmentation, which divides subscribers based on their recency of subscription and user engagement into audiences. To build a “Passive” segment would not only include filtering on users that haven’t engaged in the last 90 days, but also include a rule that they are not “New” - meaning they haven’t subscribed in the last 30 days. In Bloomreach, you would simply make sure the “New” segment is placed ahead of the “Passive” segment within the same Segmentation.
In Bloomreach Engagement, Segmentations automatically have a “conflict resolution” feature. If a user satisfies the conditions of more than one Segment in the entire segmentation, they will “fall” into the one that is calculated first, i.e., the segment that is most to the left. This saves time and makes it easier to work with more complex audiences. Without this feature, in Klaviyo, Braze, and Attentive, it is necessary to maintain far more segments and logic rules, spend a lot more time manually identifying and resolving logical conflict between the audiences and run a much higher risk of a customer falling into two mutually exclusive audiences at once, such as New and Passive.
2. User Analytics
Segmentations in Bloomreach Engagement are not only very useful for marketing campaigns. This is because Bloomreach automatically stores a user's segment membership within a customer attribute in each user's customer profile.
That has great utility. It can be used for checking if the segmentation worked as intended and for debugging purposes. The most common use of this feature, however, is in building user-based analytics and reporting.
In Bloomreach Reports - essentially an advanced PIVOT table that allows you to quickly build a wide range of reports based on Bloomreach data - Segmentations are available to be selected as a customer attribute. As a result, it is possible to quickly filter the whole report by a specific custom stored segment of users. An example of this would be excluding Employee and Tester profiles from user analytics and campaign reporting, or focusing only on the metrics of B2B customers.
Additionally, as shown in the screenshot above, it is possible to split the calculation of metrics in a Report by Segments within a Segmentation. In combination with the multi-segment feature of Segmentations, this enables valuable reports where the mutual exclusivity can be guaranteed, and each segment represents a distinct “user cohort”. This is particularly useful when dealing with A/B test results, for example, in situations where a user has seen both versions of the Variant, which are classified into the “mixed” segment to prevent “polluting” the final A/B test results.
Additional examples of relevant user analytics that take advantage of this feature are:
CLV (Customer Lifetime Value) Customer Report
Comparing B2B vs B2C users’ KPIs
Comparing users’ KPIs based on their home market (UK vs US vs EU)
Comparing users’ KPIs based on the category of their 1st Purchase
3. Usage in Jinja
Furthermore, Bloomreach enables the use of Segmentations as a Jinja object directly when building personalization logic in Scenarios, Emails, on the Website, in SMS, or Push campaigns.
Consider the example of a collection-affinity email block product recommendation. The marketing team has six different types of product collections and would like to surface products from one of them to the customer. If the customer has expressed interest in one of the collections, based on on-site behaviour and transactional data, the products from that collection would be surfaced for that customer. If the customer expressed interest in more than one of the collections, the team wanted to make sure there is a “pecking order” among them due to their varying quality and price points, so they can select which collections are “more important” in a conflict than others.
{# example of a jinja snippet using the segmentations object #}
{% set preferred_collection = segmentations['667a12cbffbbcd991416fd88'] %}
{% set products = recommendations('669a40479382cf6a199631a9', fillwithrandom = false, size = 4, category_names = preferred_collection) %}To enable this “pecking order” logic of product collection affinities, the segmentation functions very well as a dynamic attribute setting the user’s collection preference for the recommendation engine. If you only had single-segment lists available, to replicate this functionality, you would require longer and more manual code.
Thanks to Bloomreach Engagement Segmentation’s out-of-the-box conflict resolution and multi-segment capabilities, more complex segmentations for website or email campaigns are easier to build. Furthermore, thanks to the flexible Jinja abilities, it is possible to build easier and lighter code to implement the logic in email campaigns and experiences on the website.
If you found this post valuable
We hope you found this article valuable. If so, please consider subscribing (for free!) to receive updates on our latest publications.
If you'd like to explore any topics mentioned in the article further, or if there's anything else we can assist you with, feel free to schedule a meeting with us by clicking the button below: