Segmentation That Isn't Just "Engaged vs Unengaged"

Ask most brands how they segment their email list and you'll get some version of the same answer: "We send to engaged subscribers." Usually that means people who've opened or clicked in the last 30, 60 or 90 days.

And look, that's a good start. Sending to engaged people protects your deliverability and stops you shouting at an inbox that stopped listening months ago. But if engagement is your only segment, every person on that list gets the same message, whether they've bought from you ten times or never at all.

Segmentation is where email starts to feel personal. Here's how I think about going beyond the basics in Klaviyo, without building 50 segments you'll never use.

Start with the foundation: engagement tiers

Before anything fancy, get your engagement segments right. I like having a few tiers rather than one:

Your biggest campaigns go to the top tiers. The cooling-off group gets fewer, higher-value sends. The unengaged group goes into a sunset flow (more on that below) instead of every campaign.

Segment by purchase behaviour

This is where it starts getting good. How someone has bought from you tells you more than whether they opened an email.

Just splitting a campaign between "never purchased" and "purchased" and changing the offer or message can make a big difference.

Segment by what they're interested in

If you sell more than one type of product, sending everyone the same new arrival email is a missed opportunity. Klaviyo lets you build segments based on what people have viewed, added to cart or bought.

A few ways to use this:

• People who've bought from one category get campaigns featuring that category first.
• People who browsed a product range but never bought get content that helps them choose.
• People who bought a product that runs out get a reminder around the time they're likely to need more.

If your sign-up forms ask a simple question ("What are you shopping for?"), store the answer as a profile property and you've got another easy segment from day one.

Use Klaviyo's predictive analytics (if you have the data)

Once a brand has enough order history, Klaviyo can predict things like customer lifetime value, churn risk and the expected date of someone's next order. These are really handy for:

• High churn risk segments: catch people before they drift away, with a winback-style message or a reminder of what's new.
• Expected next order date: time a replenishment or "running low?" email around when they're likely to buy anyway.
• Predicted high CLV: spot future VIPs early and look after them.

If you're a newer brand and these aren't available yet, don't stress. Purchase-based segments do a lot of the same job manually.

Don't forget your exclusion segments

Segmentation isn't only about who gets a message. It's also about who doesn't. A few exclusions I'd almost always use:

• Exclude people who just bought from heavy promotional campaigns for a few days.
• Exclude people who are currently in your welcome or abandoned checkout flows from overlapping campaigns, or at least use Smart Sending to avoid doubling up.
• Exclude unengaged profiles from regular campaigns.

Getting exclusions right is one of the easiest ways to make your emails feel less spammy.

Have a sunset flow for the people who've gone quiet

Instead of sending forever to people who never open, build a sunset flow. Send one or two emails asking if they still want to hear from you, maybe with a reason to stay. If they don't engage, stop sending to them. It feels scary to "lose" subscribers, but a smaller list that actually reads your emails is far more valuable than a big one that drags down your sender reputation.

Keep it manageable

The trap with segmentation is going overboard. Fifty segments sounds impressive until you have to write a different version of every campaign.

My rule: start with engagement tiers and the purchase segments (never purchased, one-time, repeat, VIP). Get those working. Then add interest-based segments for your biggest product categories. Only build a new segment if you're actually going to send it something different.