Don’t oversegment your subscriber list. Here’s how.


Michael Bromley

Market segmentation and personalization – dividing your audience into common groups to deliver more relevant sales messages – works well.  

Donald Trump used it to win the 2016 election, and Campaign Monitor says that those who use a segmentation strategy on their email lists can experience as much as a 760% increase in revenue.

Today, we have tables loaded with analytics about our customers, containing all sorts of granular information about what, when, and how they buy our products.  And the perceived wisdom is to carve, split, and divide your customers into ever smaller groupings.

But how small should you make your segments?  Should you aim for a segment of one and strive for personalized marketing?  Or should you keep it simple, and segment based on a broader picture?

This article will take a look at common segmentation strategies and some of the potential dangers in over-segmenting your email list.

Email Segmentation Strategies.

How might you approach segmentation?  Some of the common ways are:

Demographics.  A/S/L, as we used to say – what’s your age, sex, and location?  Marketers have used basic demographics to target their advertising for over 100 years.

There are still use cases today, like gender for clothing and jewelry, or location for bricks and mortar businesses.  But, as the cliché of boomers vs. millennials informs us, demographics won’t help you understand the individual. You need to go deeper.

Customer Persona.  A customer persona is an avatar of your ideal customer. Demographics play a part, but a persona also includes market research, analytics, and intuition.  You can also have more than one customer persona for different aspects of your business.

For example, Ford will have a separate customer persona for the F-150 Truck and the Escape small SUV.  Think no-nonsense blue-collar worker vs. soccer-parent. Ford uses these customer personas to create more relevant advertising for the different types of people who buy their cars. 

You might think that you could skip creating a customer persona, but learn a lesson from Amazon.  They failed in a big way when they sent out emails for a baby registry to people who weren’t having babies.  A ‘pregnancy persona’ and proper segmentation would have helped a lot here. 

Sales Funnel Position.  Another way to segment your email list is by a customer’s position in your sales funnel.  A new lead who knows little about your offering needs a different sales message than those who are already customers. See more in our piece about what flows and automations you should have.

Below is the process flow Snappa uses to encourage new customers to upgrade to ‘Pro’ level and annual billing.

Snappa has separate campaigns to up-sell and promote increased customer lifetime value for each step taken down the funnel.

Customer Engagement.  With today’s technology, you can segment your audience based on how subscribers interact with your marketing messages. 

If a subscriber clicks on an email link or fails to open several emails, you can assign them to a segment with more appropriate messaging. You can even tie email engagement with website activity, such as pages visited or products purchased, to further hone your campaigns. 

The aim is to keep your subscribers engaged with useful emails.  If your subscribers don’t engage with your emails, then you run the risk of ISPs judging your emails irrelevant and delivering them to the spam/promotions folder. 

So, we have the technology and statistical support for dividing our email lists into ever smaller segments. The consensus is that hyper-targeted messaging works best. It then follows that you should aim to have as many segments as possible to be most effective, right?

Well, there are dangers in creating too many segments in your email list also.

The dangers of over-segmentation.

Internet activist Eli Pariser coined the term ‘filter bubble’ to describe the effect of customized content feeds chosen by algorithms on individuals.  The algorithm ‘learns’ a person’s interests and beliefs, then customizes a content feed to show correlated information, lowering exposure to conflicting ideas and opinions.  

Over time a person gets trapped in an echo chamber of content, which only appeals to, or reinforces, a particular kind of viewpoint.

Over-segmentation of your list similarly risks trapping your customer in a particular moment in time. Debra Kaye writing in Inc. argues that people, as individuals, change, adapt, and grow.  She continues, ‘People are situational animals. We use products differently depending on our mood and the circumstances in which we find ourselves.’

In short, what interests us today may not interest us tomorrow.

If you over-segment your list, because the data suggests that you should, then you could be insulating your customers from the rest of your products. You are boxing them in.

Cost implications.

Each additional segment created for an email list also costs extra money. Whether it is planning the segments, creating a new campaign, or crunching the results; segments add expense. 

A further problem with over-segmentation is the reduction in the size of data pools. Optimizing a campaign requires a sufficient level of traffic to effectively A/B test. Several larger datasets are more useful than a dozen small ones here.

And overreliance on analytics to drive business decisions can also have profound cost implications.  The Harvard Business Review explained how Tesco’s trust in marketing via customer analytical data was not enough to prevent a $3billion write-down, and an embarrassing pull-out from the US market. 

They noted, ‘data-rich loyalty programs and analytics capabilities can’t stave off the competitive advantage of slightly lower prices and a simpler shopping experience.’   

So, where does the balance lie?

Good Segmentation Practices.

Segments should be as big as possible while remaining unique and serve the intended purpose of increased profitability through relevant marketing.

Don’t aim for one characteristic per segment; it’s too granular. Instead, think of combinations of attributes. People who like motorcycles, football, and BBQ, or those that do CrossFit, keto, and get tattoos. This is the kind of information to build your customer personas with.

A perfect way to assign subscribers to segments is by getting them to self-select their interests. Here is an example from Bespoke Post via Really Good Emails.

The ability to have more than one preference is good, and it creates a win-win for the seller and consumer.  Remember, though, that people change, a hiking and camping holiday might introduce the consumer to watersports, and a new interest has been born.  So, periodically give your subscribers the chance to reselect interests. 

Hyper-personalization and Predictive Analytics. 

When your business reaches a larger scale, say, 500k subscribers, there is a place for leveraging data science and analytics.  Predictive analytics looks at the past behavior of customers and uses AI to deliver marketing messages that are hyper-relevant and timely. 

You can see examples of predictive analytics in practice when you shop online at Amazon. Amazon recommends to you items that are ‘frequently bought together’ or that ‘customers also bought.’ Amazon also uses this type of AI in their marketing emails. 

Here is an example.  On an email confirming a purchase’s dispatch, Amazon displays previously purchased items they predict you have possibly used up and may need to repurchase.

It’s all about keeping ahead of a customer’s needs and delivering the right buying prompt at the right time.


Email segmentation is something you should be doing with your list. There is compelling data that demonstrates increased profitability when you follow the practice. 

But, over-focusing on ever-smaller segments can fail to serve your customers well. You run the risk of trapping your customers in an analytical box, leaving no room for change and growth. 

Try to keep segments as broad as possible while remaining relevant.  Ask your customers their interests regularly, and when you are the right size, look to technology to drive more personalized predictive messaging.

Join our mailing list

Get your fix of marketing, design and automation tips, all written by industry experts. +1 510-394-2442San Francisco, CaliforniaMexico City, MexicoCopenhagen, Denmark