Let’s walk through an example: Imagine that you’re an email marketer for a sporting goods company (think REI). Your website offers clothing articles for men and women. You also have surfing gear and hiking gear.

Now, you could send the same email to everyone promoting a few articles in each category. But imagine that you’re the recipient and you’re a female. Which email would you rather receive? One with a little bit of everything or an email with female clothing, as well as some surfing and hiking gear? We can take this a step further and, if we know that the recipient is mainly interested in surfing gear, focus the email just on surfing gear for women.

I’d much rather receive an email that is tailored to me. The same goes for each and every one of your subscribers. Gender segmentation is often an easy one. Ask gender details during the signup process (explicitly asking) or look at a user’s browsing behavior and analyze which gender’s clothing they’re looking at (implicit data).

Gilt Groupe sends more than 3,000 variations of its daily email, for example, each tailored based on past user click-throughs, browsing history, and purchase history. Creating and sending 3,000 emails a day is very different from sending one mass email blast. As you might have imagined, Gilt’s segmentation and email variations are algorithmically generated. While this is one extreme, some companies do send many email variations manually—specially when running A/B tests across different segments (each A/B test doubles the workload since an additional email variation needs to be created, QA’ed, and monitored).

Although it’s a lot of work, it drives real returns: One financial institution increased revenue from target segments by 20 percent by using life-cycle events to trigger personalized emails to existing customers; home-goods retailer Williams-Sonoma reported a tenfold improvement in response rates by adopting personalized email offerings based on individuals’ on-site and catalog shopping behavior.

Quick aside: There is such a thing as over-segmenting your list. The segments you create should make enough of a difference performance-wise to be worth the extra effort of managing multiple emails.

Divide your subscribers into groups that are large enough so that the effort is worth it. Let’s say our subscriber list breaks down as such:

  • Males: 22,000
  • Females: 23,000
  • Likes hiking gear (that we know about): 11,000
  • Likes surfing gear (that we know about): 500

email list segmentation visual 1

Visualization for the example data 

Based on the above data, we may conclude that we’ll use the following three segments:

  • Likes hiking gear (includes males and females)
  • Males that don’t like hiking gear
  • Females that don’t like hiking gear

Here are our three segments visualized:

email segmentation example 2

As you can see, we’re ignoring the “likes surfing gear” segment because it’s too small.

You may be wondering: Should we separate the “likes hiking gear” segment into sub-segments for male and female? If you have the resources, go for it. Otherwise, keep them grouped. Another way to determine this is by running a test. How much lift did you gain with the added segmentation? Is it worth the extra effort or not?

Adding segments creates overhead—you need to draft additional email copy, come up with more subject lines, implement these campaigns in your ESP, track the performance of the email, and QA all the work you’ve done before hitting send. Each segment represents extra work. How you segment your list really depends on your available bandwidth and the returns the added segmentation provides.

Now that we have a good understanding of segments and why they’re important, let’s talk about a few popular segments you should consider implementing.


As mentioned earlier, this is one of the most popular segments. This is particularly applicable for online fashion retailers.

Recent purchasers

It is always less expensive to get previous buyers to buy again than to acquire brand-new customers. Segment your list and identify recent purchasers. Incentivize them to buy again by offering a special promotion or by making them feel exclusive by offering priority access to new items before they become available to everyone else. If you’re interested in this topic, you should read more about recency, frequency, monetary value (RFM)—a popular method used for analyzing customer value.

Highest-value purchasers

These are the big spenders. Every few months, create a campaign targeted specifically at them to promote products they’re likely to be interested in.

Most engaged users

This one is very useful for SaaS businesses. Find out who is most actively using your app and tailor your communication to them. If you offer a “pro” subscription, highlight the benefits of upgrading along with a few customer testimonials.

Events and milestones

Many websites ask for a milestone date, be it your birth date, wedding anniversary, or maybe even the date of an upcoming baby shower. Of course, these milestones and events are great times to send emails.

As an example: Imagine that you’re Charity: Water. One month before my birth date, prompt me to “donate my birthday,” where I ask my friends for donations instead of gifts.

Dormant and disengaged subscribers

A 2013 study showed that 63% of email subscribers aren’t engaged or active. Use this to your advantage. Find out why these subscribers are not engaging with your emails by asking them through a survey—offer an incentive such as a $25 Amazon gift card to increase the response rate.

Once you have a hypothesis as to why they’re not engaging with you, segment your list accordingly and tweak your messaging. It could be that you offer them steeper discounts, or email them less frequently so they don’t feel overwhelmed, or add more “relationship building” emails (emails that create goodwill because they provide value and don’t try to sell you something).

Note: How you segment your list will vary based on which ESP you’re using. If you have any questions on how to do this, you can reach out to your ESP’s support team.

Segmenting beyond your emails

Email marketing + online ads = a killer combo

As an email marketer, you should know which subscribers are customers (they have bought from you), which ones have churned (if you’re working on a SaaS business), and which ones have high purchase intent (e.g., because they saw your pricing page or abandoned the signup flow). Use this knowledge to inform different marketing initiatives, such as social media paid campaigns.

Here are a couple of examples of how email and paid media can play together in tandem:

Create ads for the following segments on Facebook, Twitter, and other social media platforms:

  • Previous buyers
  • Most active users
  • Shopping cart abandoners
  • Content promotion; then remarket to these visitors
  • Dormant subscribers

On Facebook specifically, you can create something called a lookalike audience (learn more by reading Appendix A). Here are some lookalikes you should build:

  • Similar to previous buyers
  • Similar to (active) website visitors
  • Similar to a specific audience (e.g., webinar attendees, blog subscribers, ebook downloaders, etc.)

Interested in learning more? Take a look at Appendix A.

Attribution models

Attribution models help you understand how a specific marketing channel is performing—or how different marketing channels are working together to get a customer to convert. Using different attribution models helps you surface insights that may be hidden within your data.

Google Analytics (GA), which is used by many marketers, provides a great example of the types of attribution models. To follow along in your GA account, go to Conversions > Attribution > Model Comparison and explore the different attribution models.

Let’s walk through the different models offered in GA—with help from their support page.

Here’s the scenario: “A customer finds your site by clicking one of your AdWords ads. She returns one week later by clicking over from a social network. That same day, she comes back a third time via one of your email campaigns, and a few hours later, she returns again directly and makes a purchase.

  • Last Interaction attribution
    • The last touchpoint—the Direct channel—receives 100% of the credit for the sale.
  • Last Non-Direct Click attribution
    • All direct traffic is ignored, and 100% of the credit for the sale goes to the last channel from which the customer clicked through before converting—the Email channel.
  • Last AdWords Click attribution
    • The last AdWords click—the first and only click to the Paid Search channel—would receive 100% of the credit for the sale.
  • First Interaction attribution
    • The first touchpoint—the Paid Search channel—would receive 100% of the credit for the sale.
  • Linear attribution
    • Each touchpoint in the conversion path—Paid Search, Social Network, Email, and Direct channels—would share equal credit (25% each) for the sale.
  • Time Decay attribution
    • The touchpoints closest in time to the sale or conversion get most of the credit. In this particular sale, the Direct and Email channels would receive the most credit because the customer interacted with them within a few hours of conversion. The Social Network channel would receive less credit than either the Direct or Email channels. Since the Paid Search interaction occurred one week earlier, this channel would receive significantly less credit.
  • Position-Based attribution
    • 40% credit is assigned to each the first and last interaction, and the remaining 20% credit is distributed evenly to the middle interactions. The Paid Search and Direct channels would each receive 40% credit, while the Social Network and Email channels would each receive 10% credit.

The challenge is that if you ask five marketers for their favorite attribution models, you’ll get five different answers. Unfortunately, when it comes to attribution modeling, there isn’t a one size fits all.

Important note: ESPs that offer “conversion” tracking often use cumulative tracking. Here’s how this works: Imagine you send Email_Campaign_A today to 100 subscribers. If 3 of these subscribers buy immediately, your conversion rate will be 3%. Now, imagine it’s 2 months later and, by then, 10 additional subscribers (which we originally emailed) bought something from you. The conversion rate will now appear to be 13%. Your spidey senses should be tingling, and you’d be right! The 13% conversion rate shown two months later is deceiving (accurate, yes, but deceiving). Keep this in mind when digging through the data.

If you have the ability to run your data through different attribution models, I recommend you do so with your entire marketing funnel. To get started, compare the results of last click vs. time decay vs. position based and document your findings. Are your ad campaigns performing better than you expected or worse? Does email marketing play a critical role in the middle of your funnel? When looking purely at email campaigns, compare last click attribution to email open paired with a 7-day window (or even 30 days).

It’s important to note that email conversion rates are often calculated based on click-throughs. While this makes sense, subscribers can be influenced to purchase without necessarily clicking through the email (and sometimes even without opening the email). An example would be if you received a 20% off email from your favorite store. You may open the email and then go directly to the store’s homepage without clicking on the email. Assuming you complete a purchase, the email would get no credit even though it did influence your decision to buy. Keep this in mind when calculating your email program’s performance—you are likely understating the reach and impact of email.

We’ve covered the most popular marketing attribution models. Let’s switch gears and talk about the different email marketing metrics to track and see how we can optimize each one.

Back to Table of Contents