Segmentation and Personalisation

Personalisation efforts boost revenue by 10-15% on average, and up to 25% depending on the industry. But here’s the uncomfortable truth: only 35% of businesses feel they can actually deliver personalised experiences across channels. Most are still sending the same email to everyone.

The gap between knowing segmentation matters and actually doing it well is where the money lives. And it’s a wide gap. Most brands know they should personalise. Most brands know segmentation drives better results. But the majority are still sending the same email to their entire list because ‘we don’t have time to set up segments’ or ‘our data isn’t clean enough’. Both excuses miss the point. Even rough segmentation (customers versus non-customers, engaged versus unengaged) dramatically outperforms no segmentation at all. You don’t need perfect data to start. You need to start.

Beyond First-Name Personalisation

Kath Pay (founder of Holistic Email Marketing, co-founder of Holistic Email Academy) has been saying this for years: personalisation that stops at “Hi {first_name}” can actually hurt performance. Her research finding has become widely cited: emails personalised with just the recipient’s first name in the subject line, where the body content is not personalised, can perform worse than emails with no personalisation at all. The first name sets an expectation of personal relevance that generic content fails to deliver, creating a disconnect.

Real personalisation means the content itself changes based on who’s receiving it. Dynamic content blocks that show different products to different segments. Subject lines that reference actual behaviour. Send times optimised to individual habits. Product recommendations based on purchase history, not random bestsellers.

The numbers back this up: powering your email campaigns with customer data increases your open rate by 29% and your click-through rate by 41%. Eighty percent of customers are more likely to purchase from brands offering genuinely personalised experiences. Product recommendations based on purchase history outperform first-name personalisation by 10-20x in terms of revenue impact.

I’d suggest this hierarchy for personalisation, ordered from most to least impactful:

1. Behavioural personalisation. Recommend products based on browsing and purchase history. Reference their last purchase. Acknowledge their loyalty tier. This is the highest-impact personalisation because it’s based on what someone actually did.
2. Lifecycle personalisation. Different content for new subscribers, active customers, VIPs, and at-risk customers. Each stage needs fundamentally different messaging and offers.
3. Dynamic content blocks. Show different images, products, or content sections based on segment membership within a single email template. One send, many versions.
4. Send-time personalisation. Deliver at the time each individual is most likely to engage. Most major ESPs offer this.
5. Location-based personalisation. Local weather references, local events, nearby store locations, timezone-appropriate content.
6. Name and basic demographic personalisation. Using someone’s name, acknowledging their birthday. Fine as an addition to deeper personalisation, but not meaningful on its own.

Work your way down the list. Each level adds value, but the top three deliver the vast majority of revenue impact.

Types of Segmentation

Demographic. Age, gender, income, location. The basics. Useful for broad targeting but not enough on its own. Geographic segmentation lets you localise messaging, run location-specific promotions, and send at the right time zone. For a global audience, time-zone segmentation alone can meaningfully improve open rates. An email sent at 10am New York time arrives at 3am in Sydney, which is how you end up buried under fourteen other emails by the time someone checks their phone. Time-zone-adjusted sending is a simple fix that many brands overlook.

Behavioural. What people actually do. Purchase history, email engagement, website browsing, cart abandonment. This is where the real advantage is. Behavioural trigger emails are timed to specific actions, which makes them inherently relevant. They consistently produce higher conversion rates than any other type of segmentation because the email arrives when the behaviour is fresh.

Lifecycle. Where someone is in their journey with your brand. A new subscriber needs different content than a loyal customer of three years. Customer lifecycle segmentation recognises this and tailors emails accordingly. New subscribers get onboarding. Active customers get cross-sells and loyalty rewards. Lapsing customers get win-back campaigns. Churned customers get one last attempt before suppression. Each stage requires fundamentally different messaging, tone, and offers.

The lifecycle stages most brands should track:

1. Prospect (signed up but hasn’t purchased)
2. New customer (made first purchase in last 30 days)
3. Active customer (purchased in last 90 days, more than once)
4. VIP (high frequency and/or high monetary value)
5. At-risk (previously active, engagement declining)
6. Lapsed (no purchase or engagement in 90-180 days)
7. Churned (no activity in 180+ days)

Map your email flows to these stages and you’ll naturally create a more relevant experience for every subscriber.

Psychographic. Lifestyle, interests, values, attitudes. Harder to capture but powerful when you have it. If you know a subscriber cares about sustainability, you can highlight your eco-friendly practices rather than just pushing discounts. If you know another subscriber is purely motivated by price, lead with your best deals. Zero-party data (more on this below) is the best way to collect psychographic information. Quizzes, welcome surveys, and preference centre selections all provide psychographic signals that are more reliable than inferring them from behaviour.

RFM (Recency, Frequency, Monetary). A framework borrowed from direct marketing that works brilliantly for email. Score customers on how recently they purchased, how often, and how much they spend. This gives you a structured way to treat different customer types differently. Val Geisler (founder of Fix My Churn) has built an entire practice around using behaviour-based segmentation to reduce churn and increase retention.

RFM Implementation Guide

RFM analysis sounds complex but the implementation can be straightforward. Score each customer on three dimensions, each from 1 to 5.

Recency. How recently did they last purchase? A customer who bought yesterday gets a 5. A customer who last bought eight months ago gets a 1.

Frequency. How often do they buy? Someone who buys monthly gets a 5. Someone who’s made a single purchase gets a 1.

Monetary. How much do they spend? Your highest spenders get a 5. Your lowest get a 1.

Combine these scores and you get a profile for each customer. Here’s how to treat the key segments:

RFM Score	Customer Type	Treatment
5-5-5	Champions	VIP treatment, early access, exclusive offers, referral requests
5-1-1	New customers	Nurture with onboarding, educate about product range, build the habit
4-4-4 to 5-4-4	Loyal customers	Cross-sell, upsell, loyalty rewards, ask for reviews
1-5-5	At-risk champions	Win-back urgently. These were your best customers and they’re slipping away
1-1-1	Hibernating	Sunset flow or heavy discount. Don’t invest heavily unless they respond

the honest truth: simple RFM captures 80% of the value with 20% of the effort. You don’t need a sophisticated scoring model to start. Just segment by recency of last purchase into 3-4 groups:

1. Purchased in last 30 days (active)
2. Purchased 31-90 days ago (warm)
3. Purchased 91-180 days ago (cooling)
4. Purchased 180+ days ago (cold)

Treat each group differently and you’ll see results immediately. Add frequency and monetary dimensions when you’re ready for more granularity.

For ecommerce brands on Klaviyo, predictive analytics can do much of this work automatically. Klaviyo calculates predicted next order date, predicted lifetime value, and churn risk for each customer based on their purchase history. If your ESP doesn’t offer this, the manual four-group recency segmentation described above captures the vast majority of the value.

One more practical note: RFM doesn’t have to be complex to be effective. I’ve seen brands overthink this with elaborate scoring models and weighted formulas. Start with recency alone. If that improves results (it will), add frequency. If that improves results further, add monetary. You can build sophistication over time, but the simple version works right now with no additional tools or integrations required.

Dynamic Content

Dynamic content lets you create a single email template that displays different content to different recipients based on data points. One email, but a hundred different versions. Segment A sees Product X, Segment B sees Product Y, and Segment C sees a case study.

This is one of the most powerful tools in email marketing, and most people aren’t using it. Seventy-one percent of US consumers expect brands to personalise their experiences. Seventy-six percent feel frustrated when they don’t.

Backstroke’s customers see 31% more revenue per send on average by using advanced segmentation and dynamic content. Brennan Dunn (founder of RightMessage and author of This Is Personal) has shared specific examples where implementing dynamic content blocks (showing different products or services to different segments within the same email) increased email revenue by 15-30%. The key insight: it’s not just about sending different emails to different people. It’s about making every element within a single email relevant to the reader.

Most modern ESPs support dynamic content through conditional blocks. In Klaviyo, you can use Show/Hide blocks based on profile properties. In ActiveCampaign, conditional content blocks achieve the same thing. In Mailchimp, merge tags with conditional logic work, though the setup is less intuitive. If your ESP doesn’t support dynamic content natively, you can approximate it by creating separate segments and sending slightly different versions of the same campaign to each. It’s more work but the performance lift justifies it.

A practical starting point: create two versions of your product recommendation section. Show bestsellers to non-customers and personalised recommendations based on purchase history to existing customers. This single dynamic block, applied to all your promotional emails, will improve relevance for both groups with minimal ongoing effort.

Waterfall Segmentation

A technique worth knowing: waterfall segmentation prioritises segments based on importance, so customers move through segments sequentially rather than falling into multiple overlapping campaigns. This prevents the “three emails in one day” problem that makes subscribers reach for the unsubscribe button.

Here’s how it works. You define a priority order for your segments. A customer who qualifies for multiple campaigns gets enrolled in only the highest-priority one. For example:

1. Abandoned cart (highest priority, most time-sensitive)
2. Post-purchase follow-up
3. Browse abandonment
4. Win-back campaign
5. Regular promotional campaign (lowest priority)

If a customer abandoned their cart and also qualifies for your weekly promotion, they get the cart email, not the promo. Once the cart sequence completes, they become eligible for the next campaign they qualify for.

Jay Schwedelson consistently emphasises that over-contacting is one of the biggest destroyers of email performance. Waterfall segmentation is one practical solution.

Most ESPs don’t have a built-in waterfall feature, so you need to implement it through flow logic. The basic approach: before enrolling someone in a new flow, check whether they’re already active in a higher-priority flow. If they are, exclude them. When they exit the higher-priority flow, they become eligible for the next one they qualify for. It takes some setup, but it prevents the subscriber experience from feeling chaotic.

A simpler version of the same idea: set a global frequency cap. No subscriber receives more than one automated email and one campaign email in a 24-hour period, regardless of how many flows they qualify for. Some ESPs (Klaviyo, Braze) support this natively. Others require you to build the logic manually with conditional flow steps.

Engagement Scoring

Engagement scoring assigns points to subscriber actions and decays those points over time, giving you a rolling measure of how engaged each subscriber is with your brand.

Here’s a simple model to start with:

Action	Points
Reply to email	15 points
Purchase	10 points
Click a link	5 points
Open an email	1 point
Visit website (tracked)	3 points

Apply a decay rate of 10% per week. An action from last week is worth 90% of its original points. An action from four weeks ago is worth roughly 65%. An action from three months ago is worth almost nothing.

This creates a dynamic score that reflects current engagement, not historical behaviour. Use the score to determine:

• Send frequency. High-score subscribers get every campaign. Low-score subscribers get only your best content.
• Content type. High engagement? Cross-sell and upsell. Low engagement? Re-engagement and value-heavy content.
• Flow eligibility. Only trigger certain automations for subscribers above a minimum engagement score.
• Sunset timing. Subscribers whose score drops to zero get moved into the sunset flow.

Most ESPs like Klaviyo and ActiveCampaign have engagement scoring built in. If yours doesn’t, you can approximate it with segment rules based on recency of last click.

The key thing about engagement scoring is that it accounts for recency in a way that simple segments don’t. A subscriber who clicked five links six months ago but nothing since is not engaged, even though their total click count is high. A subscriber who clicked one link yesterday is highly engaged, even though their total count is low. The decay mechanism captures this distinction. Without decay, you’re measuring historical interest, not current engagement.

Engagement-Based Sending

This is one of the easiest and highest-impact optimisations most brands can make. Instead of sending every campaign to your entire list, tier your sends by engagement level.

Tier 1: Clicked in last 30 days. Your most engaged subscribers. They get every campaign you send.

Tier 2: Clicked in last 60 days. Still engaged, but not your daily readers. They get most campaigns, maybe 75% of your sends.

Tier 3: Clicked in last 90 days. Showing signs of disengagement. They get your best content only, perhaps 50% of sends.

Tier 4: No engagement in 90-180 days. Move them into a re-engagement flow. Don’t send regular campaigns.

Tier 5: No engagement in 180+ days. Sunset flow. Reduce frequency, attempt re-engagement, then suppress.

Note: I’ve used click-based engagement here deliberately, because of Apple MPP’s impact on open-rate reliability.

The results from engagement-based sending are consistently strong:

• 15-30% improvement in open rates (because you’re sending more to people who open)
• 20-40% reduction in spam complaints (because you’re sending less to people who don’t want it)
• 0-5% change in total revenue (often neutral or even positive, because improved deliverability for your engaged segments more than compensates for the reduced sends to unengaged ones)

That last point is the one that surprises people. You send fewer total emails and your revenue stays the same or goes up. The mechanism is simple: better engagement signals lead to better inbox placement, which means more of your emails actually reach the inbox for the people who matter.

I’ve seen this pattern across many SmartrMail customers. A brand switches from ‘send everything to everyone’ to engagement-based tiers, and within 4-6 weeks their overall domain reputation improves, their inbox placement rate goes up, and their revenue either stays flat or increases. The only cost is a small amount of setup time to create the engagement segments and adjust their sending workflows.

If you’re going to implement one thing from this chapter, make it engagement-based sending. It’s the single easiest optimisation with the most reliable payoff.

Zero-Party Data Collection

Zero-party data is information that subscribers give you voluntarily and proactively. Unlike inferred data (guessing what someone likes based on their clicks), zero-party data comes directly from the source. It’s more reliable, and subscribers appreciate that you asked rather than assumed.

Welcome survey questions. In your welcome series (email 2 or 3), ask one segmentation question. Brennan Dunn’s signature technique: ask new subscribers to self-identify their role, biggest challenge, or what they’re looking for. Use the responses to tag and segment them. He’s reported that this simple step can double the conversion rate of subsequent email sequences because the content becomes specifically relevant.

Preference centres. Let subscribers choose which content topics interest them and how often they want to hear from you. Twenty to thirty percent of people who click ‘unsubscribe’ will instead adjust their preferences when given the option. That’s a meaningful number of subscribers retained.

Quizzes. “What type of [X] are you?” followed by email capture for personalised results. Tools like Interact and Typeform make these straightforward to build. The quiz format has high completion rates because people are naturally curious about how they’ll be categorised.

Post-purchase surveys. “What made you decide to buy?” or “What will you use this for?” gives you psychographic and use-case data that powers better recommendations and content.

The advantage of zero-party data over inferred data is accuracy. Someone who tells you they care about sustainability definitely cares about sustainability. Someone who clicked on one sustainability-related product might just have been browsing. The self-reported data is more reliable for personalisation.

Zero-party data also has a trust advantage. When you ask a subscriber directly, they feel in control of their data. When you infer from behaviour without telling them, it can feel invasive. The ask itself builds trust: “We want to send you relevant content, so we’re asking what you care about.” That’s a message most people respond positively to.

Preference Centres

I want to expand on preference centres specifically because they’re one of the most underused tools in email marketing.

A preference centre is a page where subscribers can adjust what they receive from you, rather than just unsubscribing entirely. It typically lets them choose:

• Content topics (product updates, educational content, sales and promotions, company news)
• Email frequency (daily, weekly, monthly, only the essentials)
• Format preferences (HTML vs plain text, though this is less common now)

The data on preference centres is compelling. When subscribers click ‘unsubscribe’ and see a preference centre instead, 20-30% will adjust their preferences rather than fully unsubscribing. That’s a direct reduction in list churn.

But the bigger benefit is the data you collect. When a subscriber tells you they only want product updates and not promotional emails, you now have zero-party data you can use to segment them permanently. Their experience improves (they only get what they want), your engagement metrics improve (they’re more likely to open and click), and your relationship strengthens (they feel in control).

Subscriber Lifetime Value

Understanding the lifetime value of a subscriber helps you make better decisions about acquisition spending, content investment, and retention efforts. The basic calculation: average revenue per subscriber per month multiplied by average subscriber lifespan in months. Simple, but most brands never do it.

Track LTV by acquisition source. Subscribers from organic search might have a completely different LTV than those from a paid Facebook campaign. I’ve seen businesses reallocate 40% of their acquisition budget after doing this analysis for the first time. Chapter 9 covers LTV calculation, acquisition cost benchmarks, and the LTV:CAC ratios you should target in detail.

---