Meta Ads Management
Meta Ads built on signal, not on reported ROAS.
Server-side CAPI, value-based Advantage+ campaigns, and cohort LTV analysis, so Meta's algorithm finds buyers, not clickers.
Post-iOS 14, the brands that are winning on Meta are the ones that restored signal quality through server-side CAPI and rebuilt their campaigns around LTV-weighted audiences rather than broad interest targeting. The brands that are losing are still relying on a browser pixel that is missing 30–40% of conversions and reporting a ROAS that makes their CFO happy but does not match revenue in the bank.
What iOS 14 changed about Meta Ads measurement and what it did not
The iOS 14 AppTrackingTransparency framework reduced Meta access to browser and app event data for users who opt out of tracking on iOS devices. For most advertisers, this meant a portion of purchase and lead generation events that occurred on iOS devices became invisible to Meta through the browser pixel alone, resulting in under-reported conversions, degraded audience match rates, and reduced signal quality for campaign optimisation. What iOS 14 did not change is the ability to send conversion data from server-side infrastructure. The Conversions API sends event data directly from the advertiser server to Meta API without any dependency on the user device or browser, bypassing the restrictions that the iOS privacy framework introduced. Advertisers who implemented CAPI with high-quality first-party data enrichment recovered a substantial portion of the signal quality lost through browser pixel degradation, and in some cases improved on pre-iOS 14 measurement accuracy because the server-side implementation was more rigorous than the original pixel deployment.
Server-side CAPI implementation: what the build actually involves
Server-side CAPI implementation routes conversion events from the advertiser server directly to Meta Conversions API, with first-party customer data included to maximise event match quality. The events most valuable to route through CAPI include purchase completions, subscription initiations, lead form submissions, and high-value engagement events where the underlying data is available at the server layer. The technical implementation requires a server-side GTM container, a hosting environment for that container such as Google Cloud Run or AWS, the CAPI connection configured with the correct event schema, first-party data enrichment using hashed customer identifiers such as email and phone number, and deduplication logic to prevent the same conversion event from being counted twice when it is received from both the browser pixel and the server. Match rate is the validation metric: a well-implemented CAPI connection produces an event match quality score that confirms the first-party data is being received and matched to Meta user records accurately.
LTV-weighted audience strategy: finding buyers, not just converters
The audience strategy that consistently produces better acquisition cohorts on Meta is one built from first-party customer data segmented by lifetime value. The standard approach: export the customer list from the CRM, segment by LTV quartile, upload the top quartile as a Custom Audience using hashed email and phone data, and create a Lookalike Audience from that seed. Configure value-based optimisation in the campaign using CAPI purchase value signals so the algorithm optimises for revenue per acquisition rather than acquisition count. The algorithm learns to find buyers who resemble the highest-value customers, not buyers who simply complete a transaction. The practical difference between value-based and event-count optimisation is the composition of the acquired cohort: when the algorithm is trained on LTV signals, the resulting customers tend to have better retention and higher long-run value than cohorts acquired through volume optimisation. The data to validate this is available in any business with sufficient purchase history and a CRM that records customer revenue over time.
Server-side Conversions API via GTM server container with deduplication logic, first-party enrichment, and signal quality optimisation to 8+/10.
Rebuild into three layers: CAPI-optimised prospecting (Advantage+), sequential retargeting, and CRM-seeded retention suppression.
LTV-weighted lookalike audiences from CRM purchase data, value-based CAPI optimisation, and interest targeting elimination where data allows.
Creative brief framework, hook testing, and format rotation based on audience stage, not the same creative across all campaigns.
Geographic hold-out tests to measure true incremental ROAS, so you know what Meta is actually driving versus claiming credit for.
Contribution margin-based budget allocation with regular hold-out testing to eliminate spend on audiences with negative incrementality.
- 01Signal audit: check current pixel match rate, signal quality score, and identify every deduplication gap between browser and server events.
- 02CAPI build: server-side GTM container, Meta CAPI connection, first-party data enrichment, deduplication keys, and match rate validation.
- 03Campaign rebuild: restructure campaign hierarchy, audience strategy, and bid signals once clean data is flowing.
- 04Hold-out test: 3-week geographic or audience hold-out to measure true incrementality before scaling spend.
- 05LTV optimisation: Klaviyo or CRM purchase history seeded into Meta as custom audience for LTV-weighted prospecting.
Ready to get started?
Book a 30-min call