Hello @Mag3me,

Thanks for sharing this question with our community.

You can simply build a segment of users and export the properties for predicted CLV and Initial Referring Domain (as long as you meet the requirements for predictive analytics outlined here).

You may also find value in reviewing the solution in the following thread on segmentation:

I hope that helps! 

Hi @Dov ,

thank you for your message!

I tried what you suggested, but there are almost no customers inside.

If I filter e.g. all Customers who placed order at least once it’s 7000. If I use Properties - Initial Referring Domain equals (or contains, made no difference for me) - facebook.com OR m.facebook.com, Instagram etc. I have 24 people in it.

Even if I use “Initial Source” equals referral, I’m only at around 900 people. My client just has facebook / IG ads turned on, so their traffic (and therefore the customer base), is mainly consisting of that.

Is it possible to track which ads (or at least detailed which source) they were clicking when they first became customers? Via custom code on the website or similar? As those existing properties don`t seem to be sufficient!

Really looking forward  to your answer!

Hi @Mag3me.

I’m afraid we can only work within the parameters of the existing properties. Aside from exporting or segmenting the source or CLV, you can look to our integration with Facebook, specifically the lead ads feature. Using Facebook lead ads with Klaviyo will allow you to funnel in users who fill out a Facebook lead ad into Klaviyo, provide a Lead Ad metric with a corresponding Lead Ad ID. You can use this ID to identify where someone came from within Facebook by leveraging it in a flow, campaign, report etc. Here’s an example screenshot of what creating a flow based on the “Filled Out Lead Ad” metric with a filter for “AdId” would look like:

Alternatively, you can build a custom report using the “Filled Out Lead Ad” metric to keep track of information pertaining to this event.

You may want to investigate our new developer portal and reach out to one of our terrific partners for any further custom solutions.

I hope that is helpful! 

We have 10 tiktok accounts… and we want to create 1 landing page each (since they’re different “brands”). When people sign up, we intend to put it in 1 masterlist .

My question is -- will we be able to segment them according to source (like source 1 = Tiktok 1, source 2 = Tiktok2 etc)? Does Klaviyo have that feature? Or will it all have “tiktok” as a source (since it’s coming from only 1 platform)? 
What would be the best route to take to be able to segment it according to source?

Hi @Bobomatcha2!

May I ask how you are integrated with TikTok? Klaviyo doesn’t have a native integration with TikTok, so if you’re integrated, I’m guessing it’s via a custom integration using our APIs. If that’s the case, then you can pass a custom $source property value via API:

Once you have that setup, you will be able to segment profiles by this source property:

• https://help.klaviyo.com/hc/en-us/articles/115005062847#properties-about-someone3

I hope this helps!

- Brian

The Initial Referring Domain and Initial Source properties in Klaviyo rely on browser-side tracking and cookies, which are increasingly limited and prone to being overwritten by subsequent visits or simply failing to capture the initial click due to things like Facebook's internal browser, ad-blockers, or Intelligent Tracking Prevention (ITP) on platforms like iOS.

This is especially common for Facebook and Instagram traffic, which often drops into (direct) or (referral) without the rich detail you need for accurate predicted CLV analysis.

To solve this and get a more robust, detailed, and accurate view of your customer's initial entry point, you need a server-side tracking solution that combines data from your e-commerce platform, your email platform, your ad networks, and a server-side container.

Specifically, a combination of the Shopify API, the Klaviyo API, Facebook Conversions API, Google Analytics Data API, Google Tag Manager, and a service like Stape or a custom solution on Google Cloud Platform is a much better approach.

The reason this combination is superior is that it moves tracking away from the easily blocked browser and into a secure, server-to-server connection.

You would use Google Tag Manager to capture the crucial first-touch data, including all your UTMs and the unique Facebook Click ID (fbc) and Browser ID (fbp), immediately upon the user landing on your site from an ad.

This data is then sent to a server-side tracking solution (like Stape or a Google Cloud Platform setup) where you can clean it up and enrich it.

Then, using server-side tagging, you can push this initial attribution data as a custom profile property into Klaviyo via the Klaviyo API and simultaneously send a more reliable conversion event (like an Active on Site or a custom Ad Clicked event) to the Facebook Conversions API.

For purchases, when an order is placed on Shopify, the Shopify API can be used to capture the purchase details.

This purchase event, which you already have a full history of, can be linked to the initial custom profile properties in Klaviyo, allowing you to segment your predicted CLV based on the accurate Initial Ad Source custom property.

This architecture ensures that the initial attribution data is locked in at the first point of contact and is not overwritten by subsequent visits.

By using server-to-server APIs like Facebook's Conversions API and the Klaviyo API, you bypass most browser-side tracking limitations, giving you a much higher match rate and greater data fidelity for the initial source, which is critical for correctly calculating predicted CLV for customers acquired via specific ads versus other traffic channels.

This level of detail and accuracy simply isn't achievable with just the standard Klaviyo web tracking properties, especially for customers coming from walled gardens like Facebook and Instagram.