First-Party Segments: Building Audiences From Your Own Orders Instead of the Pixel
The pixel is gone, the cookies are crumbling, and your best audience is sitting in your own order history.
A first-party segment is built from your own order data, not from a pixel. The customers who bought twice, the ones who spent over a threshold, the ones who bought a specific product: those are your segments, and they are more accurate than anything a third-party pixel can infer.
The data you already own
The best audience you have is sitting in your order history. The customers who bought twice, the ones who spent over a threshold, the ones who bought a specific product, the ones who have not bought in six months: those are your segments, and they are built from facts, not from inferences. An order is a fact. A pixel is a guess.
The difference matters because the pixel is going away. Third-party cookies are crumbling, the pixel's data is getting thinner, and the audiences built on it are getting less accurate. A segment built on your own orders does not depend on the pixel, and it does not get worse as the pixel gets worse. It gets better as your order history grows.
The fields you need are simple: tags, order count, total spent, last order date, and product purchased. Those five fields are enough to build most segments, and they are all in the order data. A customer who bought twice and spent over $200 is a different segment from a customer who bought once and spent $30, and the difference is visible in the data without any platform.
Building the segments
Start with the segments that map to a decision. A segment of customers who bought a specific product and have not bought the complement is a segment you can email with the complement. A segment of customers who spent over a threshold is a segment you can invite to a loyalty tier. A segment of customers who have not bought in six months is a segment you can win back with an offer.
The segment should be built from the order data, not from the email engagement. Email engagement tells you who opened an email. Order data tells you who bought. A customer who opens every email and never buys is not a good segment. A customer who never opens an email and buys every month is a great segment, and the email engagement data would have missed them.
The sync matters as much as the segment. A first-party audience synced to Meta or Klaviyo should sync weekly, in full, so the audience always reflects the current order data. The sync replaces the audience each time, which means a customer who just bought moves into the right segment within a week, and a customer who just lapsed moves out of it. A sync that runs monthly will be stale by the time it runs again.
The minimums that matter
There are minimums, and they are real. Klaviyo's CLV export needs at least 500 customers with orders, 180 days of history, and some customers with 3 or more orders. Below that, the model does not have enough data to be reliable, and the export will be thin. Meta requires a minimum of 100 people in a source audience for a custom audience. Below 100, the audience is too small to use, and Meta will not let you target it.
The minimums are not a reason to wait. A store with 200 customers can build a segment of its top 50 spenders and email them. A store with 50 customers can build a segment of the customers who bought twice and thank them. The segment does not have to be large to be useful. It has to be accurate, and accuracy comes from the order data, not from the size.
The mistake is to build a segment from the pixel because the order data feels too small. The pixel's audience is larger, but it is less accurate, and an inaccurate audience is worse than a small one. A small, accurate audience will convert. A large, inaccurate audience will not.
What to do with them
A first-party segment is useful only if it is used. A segment that is built and never emailed, never synced, and never targeted is a segment that does not exist. The test is whether the segment changes a decision: does it change who gets the email, does it change what the ad says, does it change the offer. If it does not change anything, it is not a segment. It is a list.
The first use should be the simplest one. Email the segment of customers who bought a product and have not bought the complement, with the complement. That email will convert, because the segment is accurate and the offer is relevant. The second use should be the lookalike, built from the best segment, which is the subject of the next post.
The pixel will keep getting worse. Your order history will keep getting better. Build the segments from the data that is getting better.
Keep reading
Related guides on this site: how RFM segmentation scores recency, frequency, and monetary value lifecycle segments from new to lapsed to lost how to audit the segments nobody activates More from Creative that adapts to every customer segment.
Frequently asked questions
What is a first-party segment?
A segment built from data you own, such as your order history, your customer list, and your email engagement. A third-party segment is built from data a platform infers about people, usually from a pixel or from browsing behavior.
Why are first-party segments better?
Because the data is yours and it is accurate. An order is a fact. A pixel inference is a guess. A segment built on orders will be right about who bought what, and a segment built on a pixel will be right about who looked at what, which is not the same thing.
How much data does a CLV export need?
Klaviyo's CLV export needs at least 500 customers with orders, 180 days of history, and some customers with 3 or more orders. Below that, the model does not have enough data to be reliable.
What is the minimum size for a Meta custom audience?
Meta requires a minimum of 100 people in a source audience for a custom audience. Below 100, the audience is too small to use, and Meta will not let you target it.
How often should a first-party audience sync?
Weekly, in full. The sync replaces the audience each time, so the audience always reflects the current order data. A sync that runs monthly will be stale by the time it runs again.
What fields do I need to build a first-party segment?
Tags, order count, total spent, last order date, and product purchased. Those five fields are enough to build most segments, and they are all in the order data.