The primary audiences framework for scalable personalization

As the concept of tailoring digital experiences has become ubiquitous, personalization technology has made it easier for marketing teams to run complex queries for accurately segmenting and targeting users with relevant experiences.

However, on the hunt to better understand their audiences, many organizations have developed a habit of building hundreds of different segments, each designed to serve a different need. While microsegments can be beneficial in certain scenarios, prioritizing too many presents several challenges, including an inability to effectively manage the number of segments, consistently target users, and pinpoint trends and patterns in user behavior.

Thankfully there is another way to cozy up to customers, and it’s rooted in a few simple macro-segmentation guidelines that we refer to as the Primary Audiences framework. Below, find out how you can use it to not only improve your targeting strategy but also scale your larger personalization program.

Micro-segmentation in moderation

In theory, dozens (if not hundreds) of microsegments, would allow businesses to deliver on the value of true 1:1 personalization. But in practice, microsegments are difficult to accurately target given they typically only make up a small percentage of the site’s total traffic. Monitoring and optimizing the performance of each segment can also quickly become untenable, which is further complicated by the need to track user movement for the avoidance of segment overlap.

Until recently, a major American retail brand counted a whopping 422 segments as part of its personalization program. Overwhelmed and unsure of how to derive value from each, the team cited:

1. Time wasted identifying and evaluating segments
2. A lack of connective tissue between various tests
3. Inability to compare performance across segments
4. Missed opportunities to institutionalize learnings

While problematic under the circumstances outlined above, microsegments do serve an important purpose in pinpointing unique behaviors and characteristics. In turn, they can allow teams to target very specific cohorts of users with the most relevant offers and messaging. For example, an experience tailored to users that click on a particular ad campaign.

Though, organizations looking to establish a holistic approach to personalization will need to adopt a wider and more consistent methodology for segmentation.

Introducing the Primary Audiences approach

In its simplest form, segmentation represents the process of understanding what an audience represents and then using the information collected to personalize experiences accordingly. But what ultimately separates mass micro-segmentation from macro-segmentation is the ability to effectively scale that process, which is why the Primary Audiences framework is based on the latter.

Empowering teams to get closer to their customers in a functional and repeatable manner, when put in motion, the Primary Audiences framework creates a waterfall effect of richer and more consistent audience learnings, improved targeting and testing strategies, uplifts in important business KPIs, and incremental revenue growth.

And it all starts with four core guidelines:

• Audiences must be based on a single segmentation principle
• Only 3-4 audiences should be prioritized at the maximum
• Those audiences should cover 100% of the site’s traffic
• Of which all ideas, execution, and analysis is meant to stem

Breaking down the framework further…

Simplifying segmentation for clarity

Crucial to unlocking personalization, teams must come to know what each of their segments represents as well as any major differences between them – only then can they generate a focused plan for targeting.

That is why it is suggested to build an audience based on a single segmentation principle, which is ideally behavioral in nature (e.g. current actions the user is taking) as opposed to static concepts that are assigned from past activity and may no longer hold true.

A few examples of segmentation principles

• Intent level
• Propensity to purchase
• Customer lifetime value
• Loyalty status

This not only provides teams with clarity on who and what separates their audiences, enabling them to craft and target the right experiences, but also results in greater overall alignment and buy-in from executive sponsors, which can massively impact the ongoing support a personalization program receives.

Limiting audiences to enable scalability

Keeping in mind that less is more, despite a number of potential segmentation principles to choose from, their application should be limited to 3-4 audiences maximum.

Scalability is key in a world where resources are never infinite and only so much marketing, development, and analyst support are available. Teams can overcome these common resource constraints and be more effective simply by consistently targeting, analyzing, and optimizing towards a few primary audiences.

And with the combination of machine learning, algorithms can be used to do the manual labor of deciding which variation of an experience should be shown on a 1:1 level so teams can focus on how to more meaningfully connect with their core customers.

Personalizing to everyone for maximum efficiency

The goal of website personalization is to deliver greater relevance on the whole versus enhancing experiences for small pockets of traffic, which can also minimize the incremental revenue gain associated with these efforts.

Therefore, it is important when choosing a company’s 3-4 primary audiences to consider the percentage of traffic each segment represents. Maximum impact and optimization opportunities can only be achieved if audience segments cover 100% of the site’s total traffic when combined and share little to no overlap with one another.

Let’s say, for example, a team decides to move forward with 3 primary audiences, one for each intent level (low, medium, and high). While together, each bucket may account for close to 100% of visitors, the high-intent audience could make up only 3-5% of the overall traffic – while considered a microsegment, it still rolls up to the larger segmentation principle, making it more strategic in nature.

Operationalizing customer centricity

With 3-4 primary audiences in place using the guidelines outlined above, the next step to guaranteeing personalization success lies in assessing KPIs at the segment level.

Let’s take a look at the following campaign results:

• Overall CVR = 2% | AOV $5

We may understand how the experience performed overall, but it’s unclear what should be done with the information to improve conversion rate or average order value. Unless we break down the results across our primary audiences:

• Audience A CVR = 1% | AOV $5
• Audience B CVR = 5% | AOV $3
• Audience C CVR = 3% | AOV $10

Here, we can see that each audience has exhibited different buying behaviors, which opens the door to new personalization opportunities based on audience need. In fact, audience performance should be incorporated into the creation of every single campaign – an easy goal to achieve if the company’s segmentation strategy flows from the same 3-4 primary audiences.

Eventually, that process would look something like this:

• Clearly state the business objectives and goals
• Identify data-led opportunities per audience but geared towards the overall goal
• Execute experiences targeted at and for each primary audience
• Report on each experience at the audience level
• Evaluate performance per campaign and audience (and ask why / what it means)
• Use insights collected to inform and better target future campaigns
• Track trended KPIs at the audience level as well as overall

How a leading golf retailer put personalization into practice with Primary Audiences

Specializing in designing, manufacturing, and selling golf equipment, a U.S. retailer wanted to craft an exceptional experience for its online shoppers. After identifying a correlation between conversions and the number of pageviews, its team decided to build and analyze a set of audiences based on low, medium, and high intent. Low was defined as the user visiting fewer than a dozen pages, and high was when a user visited over two dozen.

Breaking down site visitors according to intent level revealed a few key insights about how different degrees of knowledge impacted purchase decisions for the company’s golf equipment:

• Higher intent users knew the drill, preferring more technical details and brand-specific product messaging
• Low intent users often browsed according to “player type,” and needed more education on specific products

Using this info, the golf retailer was able to optimize its campaigns at the audience level, more deeply understand each segment’s user behavior, and yield even greater returns.

In the example above, specific category page messaging was designed and served to educate and drive purchases among low-intent users.

The golf retailer plans to incorporate learnings from additional on-site experiments as well as expand its use of the framework within other channels like email.

Macro segmentation as a means to deeper personalization

While some of the biggest brands in the world work with hundreds of segments, when not used in parallel with a proper macro-level approach, personalization efforts often end up not being tied together in a consistent way. But as consumer expectations for relevance grows and businesses seek to maximize the resources invested in meeting their demands, the time for prioritizing efficiency, scalability, and the institutionalization of customer insights is now. Thankfully the Primary Audiences framework allows teams to do this, not only bringing them closer to their customers but also unlocking incremental revenue because of it.