One of the hottest topics related to segmentation with brands leading up to the holidays is manual list suppression. It makes sense given we’re currently living in a deliverability-infused world mixed with brands feeling like Klaviyo is bleeding you dry.

You spend a lot of money acquiring a customer, so to the extent you don’t have to pay for them again because they’re already in your list, that’s like free money. That’s also where list suppression comes in.

After lots of questions from operators about whether we could detect a profile that should be un-suppressed… we finally ran some tests. Turns out, we can predict them quite well.

We ran our behavioral models across a massive list of profiles manually suppressed based on engagement (leave anything suppressed based on bounces, spam complaints, etc. alone). If you’re not familiar with manual list suppression, these are people who didn't unsubscribe, but are often swept up in manual engagement suppression automations so many brands run.

I honestly wasn’t sure how we would do, but our AI found thousands of users showing high-intent signals like visiting the site, viewing collections, even searching for specific products, but receiving zero emails because they were in "ESP Jail."

Now, I know your first question: “But Nathan, I don’t have your AI because Raleon was acquired. So why are you telling me this?”

Because, dear operator, I put together a set of rules you can drop into Klaviyo or Mailchimp that don't need a neural network to find those hidden customers. You just need to reconsider your "never un-suppress anyone" rule.

I took some of what we saw behaviorally work, and tried to create rule based segments that covered most of the low hanging fruit. You might have to adjust it a little to match how your Klaviyo it setup, so if it doesn’t straight copy & paste don’t panic.

And as always, with these segment suggestions, make sure you layer in your common rules like “allowed to receive marketing communication” and the like. I’m just covering how to hit those hidden behaviors.

These people were in the process of giving you money, but gave up. Because they are suppressed, they may not have even get your Abandoned Cart flow. They are probably wondering if you went out of business.

We see this every year. You cleaned your list in September to "warm up IPs," and accidentally cut off the guy who only buys gifts in December.

The AI loves this group. They aren't buying yet, but they are lurking. They are the digital equivalent of someone pacing back and forth in front of your store window, thinking about what to purchase (In other words, what I do on 5 different watch brands’ sites monthly).

Even suppressed users get Order Confirmations. If they are clicking those, they are alive.

If you find you have a large number of these (for instance, 5,000+ profiles):

This is just an extra precaution to make sure that, while you’re seeing all the indications that it’s worth sending to them again, you don’t inadvertently spike unsubscribes. Again, with these tactics, you can absolutely email them. It just pays to be considerate of deliverability ahead of BFCM.

Our AI model predicted which of these users was showing us conversion and engagement like behaviors, saving us from un-suppressing the wrong profiles. But in the absence of AI, use the rules above. You’ll know it’s working based on the profile count of the segments you have.

What we’ve found is, depending on how aggressively you suppress, you may be sitting on a goldmine of people who know your brand, have visited recently, and are legally contactable. Don't let "best practices" cost you actual revenue.

And that's it for this week's edition.

Make the most of all the profiles you have this holiday season. Leverage AI in an actually useful way to find those profiles, and most importantly - have a wonderful thanksgiving for those celebrating it.

Be on the lookout for next week’s email. I’ve got a killer Claude skill I’ll be sharing.