Segmentation Doesn’t Come Close To Personalisation SwiftERM

David Swift
4 min readJan 31, 2024

The ecommerce marketing emails that perform best across all measures (think all the traditional metrics: opens, engagement, CTR, conversions) have one thing in common: they are extremely relevant to their recipients, meaning the products, offers and content of the email actually matter to that shopper specifically.

Retailers can achieve relevance through personalisation or coincidence (e.g. an email arrives timed to when that individual is in the market at the exact moment with exactly the product they want).

Traditionally, brands have used audience segments to try to achieve relevance without having to make hard decisions around products, offers, content, priority and timing. However, this approach is suboptimal as it leaves money on the table for brands and ends with most consumers feeling frustrated after receiving content and offers about which they don’t care.

Instead, the retailers that see the biggest returns focus on relevance by going beyond simple segmentation to deliver true personalisation.

Segmentation vs. Personalisation: The Difference Matters

Is there a difference between segmentation and personalisation? And does it even matter? Yes and yes. The difference is massive, can you afford to be hunting down 5–6% returns when 10,000% is available?

Segmentation involves dividing your customers into audiences based on broad factors like brand or product interest. It usually requires a CRM or CRM-type system, normalised data and attributes tied to a targetable ID, as well as some broad-based understanding of different buyer types that coordinate to different product or offer affinities.

Personalisation involves creating unique experiences for individual customers based on factors like product, content or offer recommendations and priority and timing of your communications. Personalisation is a tactic to achieve relevance and to achieve a specified business outcome. With the advent of AI, and specifically machine learning (ML) within that, systems such as SwiftERM who offer a wholly autonomous solution, maximising the ROI without emailing marketing’s single greatest cost after product purchase, namely staff.

Segmentation can be (a manual) way to develop a personalisation strategy, but segmentation does not equal personalisation, or even come close.

For example, if you build an audience of people who have bought shirts in the past and send them a static email featuring shirts at a single point in time, that is segmentation. But if you (1) build an audience of people who have bought shirts in the past, (2) send them an email with personalised product recommendations that populate based on their individual browsing patterns and (3) apply send-time optimisation so the email sends at the best time for each recipient, that is personalisation.

While both examples have the same audience, the second example provides each recipient with unique recommendations based on their past behavior and predicted interests and it ensures they receive the email at a time when they are most likely to engage. As a result, two people in this audience can receive different emails at different times, creating truly personalised experiences.

The Power of Personalisation in Email Marketing

Time and again benchmark data underscores the power of email personalisation.

Consider the case of one athletic retailer that recently ran a test pitting non-personalised emails against emails featuring personalised 1:1 product recommendations. In this head-to-head test, the personalised emails delivered performance gains at every level compared to the segmented emails, including a 57% increase in revenue per email.

While audience segmentation used to be a good step toward achieving this goal, it’s was only ever a first step. With the advent of AI now it typically just creates more work, and lower returns. It requires cutting and managing lists, building out offer and product strategies with merchandising and taking on highly cumbersome optimisation tactics.

Simply segmenting your audience to find a group of shoppers interested in shirts helps narrow down which products to promote, but shirts is still a broad category and then requires more work with analytics or guesswork. Even segmenting shirts by gender only narrows it down slightly, and it leaves out cross-category shoppers like a mom shopping for her son. Wouldn’t you think the site you most like would take more care about your custom instead of you being sent any old thing?

To ensure relevance for every recipient and see significant increases in engagement as a result, you need to take personalised to the next level. This personalisation can include any or all of the following:

  • Dynamic product recommendations that populate for each individual based on factors like co-purchase patterns, next best purchase recommendations or predicted affinities
  • Dynamic content and offers that populate for each individual based on factors like discount preference, customer lifecycle stage and purchase history
  • Dynamic send times to optimize for engagement with each individual

Moving Beyond Segmentation to Personalisation

Understanding the power of personalisation in email marketing is one thing. Being able to move beyond segmentation to deliver that personalisation is quite another.

Historically, brands have had difficulty scaling personalisation. These challenges typically result from using decades-old email marketing technology that was built to solve decades-old problems where the developers now have to gaslight you to sell it. Given how these technologies were built, they make moving beyond segmentation difficult.

However, a new generation of retail personalisation software has arrived, giving us modern ESPs built to solve modern problems (like personalisation). These modern ESPs support the move from segmentation to personalisation by providing marketers with an wholly autonomous solution without need of human involvement whatsoever, freeing up much needed time, with guaranteed avoidance of errors and omissions.

Originally published at on January 31, 2024.



David Swift

SwiftERM hyper-personalisation SaaS for ecommerce email marketing.