Tell Meta, Google and TikTok which customers are worth paying for.
The agent predicts what each shopper will spend over the next 12 months and sends it to Meta, Google and TikTok as audiences.
- Audiences built from predicted 12-month value
- Each order sent server-side and counted once
- Your own customers tested out of retargeting
Ad platforms learn from the signals you send them.
When the pixel and a server feed both report an order, Meta can count it twice; at one bedding brand it logged 2.11 purchase events per real order (July 2026). The pixel also misses iOS and in-app buyers, so the platform learns from part of your buyers and bids on clicks, not worth.
The Meta ROAS use case puts Meta's numbers next to the store's, row by row.
Get more from your ad spend →Three signals that tell the platforms what a customer is worth
Everything goes into your own ad accounts. You keep your budgets, bids and creative.
Audiences built from customer value
Every shopper ranked by predicted 12-month value; your best customers seed the lookalike.
One order, counted once
Purchases sent from the server with your pixel's event ID and a predicted value, iOS and in-app buyers included.
Your own customers out of retargeting
An exclusion test takes your active email list out of retargeting and branded search.
What brands measured
Revenue from store orders; ROAS is Meta's own report.
Shopify revenue per ad click, 23 May to 22 July 2026 vs the same days in 2025; budget 2.1× while Meta-reported ROAS held within 8%.
Read the case study →Same brand, matched 30-day groups+65%buyer rate of shoppers the value campaign reached: 0.56% against 0.34%.
Read the case study →European bedding and sleepwear brand on WooCommerce-33.7%Meta-reported ROAS after the brand switched off personalization for a test, June vs August 2026, CPM and CTR flat.
Read the case study →Discount sportswear retailer29%of store revenue from customers already on the email list, who still clicked 40,000+ paid ads in July 2026.
Read the use case →Where this has run
Questions about ads audiences
What are value-based audiences?
Audiences built from what each shopper is predicted to spend over the next 12 months, not from one purchase. Releva ranks every shopper by predicted value, seeds lookalikes with your best customers, weights visitors by likely spend, and syncs the audiences to Meta, Google and TikTok.
What does Releva add that the ad platforms don't have?
Your store's view. Meta, Google and TikTok optimize for their own numbers, the same way for everyone. Releva tells them who is worth finding, from predicted customer value in your own orders, and who is already your customer and can stay out of retargeting.
How is this different from Meta Advantage+ or Shopify Audiences?
Advantage+ campaigns use Meta's own signals to find buyers, and Shopify Audiences is available to Shopify Plus merchants. Releva builds audiences from your own store data and ranks them by predicted 12-month value. It sends the same audiences to Meta, Google and TikTok, and runs customer exclusions as a test.
Will excluding our own customers cut sales?
That's why it runs as a test. Only your active email list comes out, and only from retargeting and branded search. Prospecting stays on, those customers still hear from you by email and push, and you decide after seeing the result.
Do you send server-side purchase events?
Yes, to Meta and Google. They share your pixel's event ID, so each order is counted once instead of twice.
Do we have to change how we run campaigns?
No. You keep your accounts, budgets, bids and creative. Releva adds the audiences and suggests exclusions, and you decide.
Bring your ad account and your store
Thirty minutes, side by side. You leave knowing what the platforms claim and what your orders say.
Book 30 minutes
You'll talk to one of us.