Zappush ranks every customer by lifetime value from your reconciled first-party data, builds a clean VIP seed of your top buyers, and syncs it to Meta and Google as a value-based lookalike source. The platforms model acquisition on the customers who are actually worth the most. The seed auto-refreshes as new VIPs emerge, the same way on Wix, WordPress, Kajabi, custom, or Shopify.
From $59/month · 14-day free trial · Cancel anytime
Nadia Rahman
Top 5% by lifetime value
+52% revenue uplift
Native lookalike audiences are built from the pixel, and the pixel never saw order value, repeat rate, or lifetime value. So "lookalike of purchasers" is really a lookalike of your cheapest, most common buyer. It gets worse because the pixel misses an estimated 30-40% of conversions after iOS 14.5 and ad-blocker attrition, and Safari's ITP caps client-set cookies at 7 days, so the seed is both incomplete and value-blind. Feed a value-based lookalike a list of "all customers" and thousands of mixed-quality records drown out the few hundred true VIPs. You end up scaling acquisition against volume, not value. You win low-LTV, one-and-done buyers at a CPA your contribution margin can't support, while your actual best cohort is never used as a modeling target. The longer the seed stays stale and unranked, the further the models drift from your best customers.
Zappush computes LTV per customer from reconciled first-party order history: total spend, order frequency, and recency across every device and session. It then ranks the whole base into percentiles. It reads real orders joined server-side on a shared event_id, not a decaying cookie, so a VIP who bought last year on another device still ranks correctly.
Take the top percentile, the top 10-20% by LTV, and assemble it as a tight, high-intent seed instead of an all-customers dump. Zappush hashes and normalizes the identifiers with SHA-256 and sizes the seed into Meta's documented 1,000-50,000 range, so the models train on quality, not noise.
Zappush pushes the VIP seed to Meta Custom Audiences and Google Customer Match with each member's historical LTV attached as customer_value, so both platforms build value-based lookalikes off it. As customers cross into or out of the top percentile, membership updates automatically. The seed stays current instead of a monthly CSV re-upload.
Score and rank every customer by lifetime value from reconciled first-party orders, using spend, frequency, and recency. Slice by percentile (top 10%, top 20%) instead of the blunt "purchased once" flag the pixel can offer.
Assemble the top-percentile cohort as a tight, high-intent lookalike seed sized into Meta's 1,000-50,000 range, so a few hundred true VIPs aren't drowned out by thousands of low-value records.
Attach each member's historical LTV as customer_value and sync to Meta Custom Audiences and Google Customer Match. Both build value-based lookalikes and bid toward high-value prospects. One seed, every platform.
Choose how tight the seed is against how far the lookalike reaches. A narrower top percentile yields a more similar, higher-intent audience. A wider one trades precision for scale, so you tune the quality/reach balance deliberately.
Every new reconciled order recomputes LTV and percentile, and membership updates within minutes. Emerging VIPs are added and demoted customers removed automatically, so the model stays trained on who is valuable now.
E-COMMERCE LTV
A lookalike of all purchasers is really a lookalike of your cheapest, most common buyer. Zappush ranks every customer by lifetime value from reconciled order history and assembles the top 10-20% as the seed, sized into Meta's documented 1,000-50,000 range. Each member ships with historical LTV attached as customer_value.
HIGH AOV
One $4,000 repeat client and one $40 one-time buyer fire the same Purchase event, so the pixel cannot tell them apart. Ranking on total spend, order frequency, and recency separates them, and only the top percentile enters the seed. Modeling then runs against the cohort your contribution margin can afford to win.
SUBSCRIPTIONS
VIP cohorts turn over: last quarter's top spenders are not this quarter's. Every reconciled order recomputes LTV and percentile rank, and membership updates within minutes, so emerging VIPs are added and demoted customers drop out. The seed reflects who is valuable now instead of a monthly CSV re-upload.
Start free in minutes, or see it live on a call.
From $59/month · 14-day free trial · Cancel anytime
