Most Personalization Is Fake
if the data is shallow, the message will be too
Hi Team!
It’s been a great week. We made it through the vicious heatwave in NY that made it feel like a European summer, and I restocked points and miles big time as well.
Chase Sapphire Reserve 150k bonus (current offer)
Chase Business Ink 100k (current offer)
Amex Gold 150k (my wife and I both snagged a few months back)
We wiped through 300k points on our St. Kitts and Panama trip, but were able to recoup another 445k since then, with another 100k coming down the pipe :)
Sicily and Malta are booked for August.
Croatia, Montenegro, and Bosnia are booked for October.
The points-and-miles gods are shining on me this year. ☀️
A few weeks back, we talked about segmentation and why most brands are grouping customers by what they bought instead of who they are.
This week, I want to get a bit deeper and more practical: what useful personalization data actually looks like and what a few brands are doing differently when they build around it.
Most brands have gotten very good at using data that doesn’t actually tell them much about the customer. A customer viewed a hoodie, so the hoodie now follows them around the internet like a Victorian ghost.
That’s behavioral recall, and the issue isn’t that the data is wrong. It’s that the data is too thin to build anything meaningful on top of.
Clicks, purchases, abandoned carts, PDP views, discount usage, and maybe a quiz response from 2021 that nobody’s touched since. Helpful signals, but a very shallow foundation.
I want to talk about what doing this better actually looks like.
Let’s get into it.
This week’s newsie is brought to you by OuterSignal!
Their order tells you what they bought. OuterSignal shows you who they are.
OuterSignal tells brands 10x more about their customers than they know right now — age, occupation, property value, social following, family status, interests, education, and more — and uses that to identify VIP customers, build precision segments, and personalize outreach at scale.
Whether it’s flagging celebrities, athletes, investors, podcasters, buyers at large retail chains, or simply surfacing micro-influencers within a target cohort, OuterSignal finds, flags, and activates the customers that matter most with unprecedented precision.
OuterSignal will research 1,000 customers for free, and is offering friends of Eli 50% off their first two months when they upgrade.
Tell them Eli sent you!
Shallow signals make shallow personalization
Most brands personalize off whatever data is easiest to grab: viewed product, purchased product, clicked email, added to cart, browsed collection. That data tells you what happened inside your store or email program. It doesn’t tell you much about the customer’s life outside of it.
Your behavioral data sees the same action no matter who’s behind it, so the brand sends everyone the same ‘Still thinking it over?’ email and calls it personalization.
The data is technically accurate. It’s also too thin to make the message feel like it was written for a human.
This is how brands end up with lifecycle programs that look sophisticated in Klaviyo and feel completely generic to the person reading them. The logic tree is impressive.
The message is still basically, “You looked at this. Please buy it.” Very moving stuff.
What Gratsi figured out
I recently came across Gratsi, a DTC wine brand selling boxed wine that does not look like something you’d hide under the sink before guests come over.
They had the same problem most brands have: people were buying, subscribing, clicking, and coming back, but the customer file still didn’t explain much. “Wine buyer” is technically true, but it doesn’t get you very far.
A retiree hosting friends, a health-conscious professional looking for a lower-sugar option, a founder sending client gifts, and a parent trying to get through Friday night without reading 19 tasting notes from a guy named Sebastian can all look identical in Shopify.
Gratsi used OuterSignal (this week’s sponsor) to understand their core personas. OuterSignal enriched their customer data and surfaced different identity-based groups inside the customer base: Active Retirees, Health-Conscious Professionals, Creative Entrepreneurs, and a few others.
Then Gratsi took one campaign and changed the message based on those groups. The send time, audience size, subject line, and product stayed consistent. The email itself changed based on the reason someone might care.
Active Retirees got messaging around slowing down and enjoying the moment.
Health-Conscious Professionals got messaging about zero sugar, zero guilt, and feeling better the next morning.
Creative Entrepreneurs got a different angle entirely.
The persona-matched messaging drove a 47.2% lift in revenue, 54.1% higher click rate, 56.1% more subscription conversions, and 41.7% more orders. Open rates were basically flat, which is the part I care about most. The lift happened after the open, when the message matched the person reading it.
Purchase history is a lazy creative brief
This is where lifecycle teams get stuck. The brief says “past wine buyers,” which could mean dinner party hosts, weekday drinkers, collectors, gift buyers, wellness shoppers, empty nesters, suburban parents, founders, retirees, or someone who bought one box because an ad caught them at 11:43 PM after two glasses of someone else’s wine.
Same thing in apparel. “Bought a dress shirt” tells you almost nothing useful. One customer is an executive wearing it twice a week. Another is a groom buying it for one occasion. Another needs something affordable and wrinkle-free for work travel.
Same SKU, completely different customer, and if those three people get the same email, one might convert, one ignores it, and one unsubscribes because the brand keeps talking like they’re shopping for a life they don’t have.
The product is the shared detail. The reason for buying is where the money is.
Better data makes the copy less lazy
A lot of weak marketing starts before the copywriter ever touches the brief. The brief is boring because the segment is boring, and the segment is boring because the data underneath it is too basic.
The “insight” is that someone purchased Product A, so now the brand would like them to purchase Product B. Groundbreaking stuff. Somewhere, a lifecycle marketer is bravely dragging a product block into Klaviyo.
Identity data gives the creative team something real to work with. A health-conscious professional should get a different wine message than an active retiree. A new parent should get a different apparel message than a founder. A micro-influencer has a different value to the business than a one-time buyer at the same AOV.
This is also why brands end up over-discounting. When you don’t understand the customer, price becomes the easiest lever. So the brand sends 15% off, then 20% off, then a final final final chance email that somehow returns every month like a raccoon in the attic.
Better customer data gives you more to work with: status, use case, lifestyle, timing, identity, actual reason to buy. The discount becomes a tool instead of the whole damn strategy.
That’s it for this week!
Any topics you’d like to see me cover in the future?
Just shoot me a DM or an email!
Cheers,
Eli 💛
