Stop paying Meta to guess. Start feeding it proof.
When 65% of your conversion events go unmatched, Meta’s algorithm optimizes on a fraction of your buyers. Upstack sends identity-resolved events through the Conversions API so match rates hit 90%+ and CAC drops 15–30% in 60 days.
Perfect White Tee saw 90x ROI and a 33% CAC reduction in 60 days.
Trusted by DTC brands spending $50K\u2013$5M/month on Meta ads
4.9/5 on Shopify App Store
1,000+
Brands
actively running Upstack
40%+
Matched events
average signal uplift
<20 min
Setup
most brands go live
Meta can’t optimize on events it can’t match to real people.
Your CAC keeps climbing and nothing in Ads Manager explains why. The creative is strong, the audience is right — but Meta’s algorithm is training on 35% of your actual buyers because it can’t match the rest. You’re not bad at media buying. You’re feeding the machine broken data.
It’s not a pixel problem. It’s not a CAPI problem. Without identity, the algorithm is guessing.
What’s happening
| What Meta Sees | Reality | Impact |
|---|---|---|
| 35 purchases matched | 100 purchases happened | Algorithm training on 35% of data |
| CAC looks like $85 | True CAC is $55 | Decisions based on wrong numbers |
| Campaign A “outperforming” | Campaign B drove more sales | Scaling wrong campaigns |
| “New customer” | Returning customer | Paying to acquire existing customers |
“I’m spending well into six figures a month. When you can get an edge on your competitors, it’s a big deal.”
— Johnny Hickey, CMO, Perfect White Tee
Identity-resolved events. 90%+ match rate. EMQ 7.5+.
Most CAPI solutions just move broken data from browser to server. Upstack Signal resolves identity first — so every event Meta receives can be matched to a real person.
Capture
Server-side event capture bypasses ad blockers, Safari ITP, and iOS restrictions. Events Meta’s pixel misses, Upstack catches. 99.9% event capture rate vs. 40–60% for browser pixels.
Resolve
Upstack ID builds a persistent identity graph for every visitor, lasting up to one year. “Four anonymous sessions” becomes “one known customer.”
Signal
Enriched CAPI events carry hashed email, phone, fbclid, external ID, and device fingerprint. Meta matches 90%+ of events to real users instead of the industry-average 35%.
How Data Flows
From first visit to activated signal
Reduce CAC 15–30% in 60 days.
| Metric | Without Signal | With Signal |
|---|---|---|
| Match rate | ~35% | 90%+ |
| EMQ score | 4–5 | 7.5+ |
| Attribution coverage | 35% of conversions | 90%+ |
| Algorithm training | Partial data | Complete data |
| Time to results | — | 30 days |
Perfect White Tee
−33% CAC in 60 days
8-figure brand. CAC dropped 33% in 60 days after installing Upstack Signal.
Paire
−20% CAC in 60 days
Revenue was up 25% but spend was up 47%. After Signal, MER improved 24% and blended NET ROAS jumped 40%.
Montreal Weights
+15% ROAS in 90 days
Upstack Signal sent cleaner data to Meta, resulting in a 15% increase in overall ROAS and $55K in recovered abandonment revenue.
“We didn’t change anything. Not our creative. Not our spend. We just plugged in Upstack — and Meta started working again.”
Nathan Yun
Founder, Paire
Paire tried Instant (nearly 2x cost, no meaningful lift). Then they plugged in Upstack — MER improved 24% and blended NET ROAS jumped 40% in 60 days.
Not all CAPI is created equal.
Basic CAPI solutions send anonymous events server-side. Signal enriches every event with resolved identity parameters before delivery.
| Parameter | What We Send | Why It Matters |
|---|---|---|
| Email address | Hashed, verified | Primary matching signal |
| Phone number | E.164 format, hashed | Strong secondary signal |
| Click ID (fbclid) | Captured at ad click | 100% match when present |
| Client IP | First-party | Network-level matching |
| User agent | Browser/device info | Device-level matching |
| Browser ID (fbp) | Meta’s first-party cookie | Session matching |
| External ID | Upstack ID | Consistent cross-session |
What is Event Match Quality (EMQ)?
EMQ is Meta’s score (1–10) for how well your conversion events can be matched to Meta users. Most brands score 4–5. With Upstack Signal, brands consistently hit 7.5+. At that level, Meta’s algorithm has enough data to actually optimize your campaigns.
Server-side tracking is table stakes. Identity is the breakthrough.
Identity-resolved CAPI, not raw CAPI
Most CAPI solutions just move the same broken data from browser to server. Signal enriches events with resolved identity before they reach Meta.
One platform replaces two vendors
Most brands pay for a tracking tool AND an attribution tool — two contracts, two dashboards that never agree. Signal includes tracking, identity, and attribution in one platform.
Specific proof, not vague promises
We don’t say “better accuracy.” We say “EMQ 7.5+ in 30 days, CAC reduction 15–30% in 60 days.” Every claim is measurable.
Results in days, not months
Perfect White Tee saw CAC drop 33% in 60 days. Paire saw 20% CAC reduction in 60 days. Enterprise solutions take 6 months. Signal takes under 20 minutes to set up.
Signal is part of the Upstack stack.
Upstack Pixel captures events. Upstack ID resolves identity. Signal activates that identity by sending enriched events to your ad platforms.
Frequently asked questions about Meta CAPI optimization
Stop feeding Meta broken data.
Every unmatched event is a missed lesson for the algorithm. Fix the signal foundation and watch CAC drop as Meta finally learns what your buyers look like. Start a free trial — see results in 30 days.
Cancel anytime
Not ready for a trial? Book a 15-minute walkthrough or read our DTC tracking guide.