Server-side attribution · 6-platform CAPI · Multi-location retail

Your POS shows revenue your Meta and Google reports don’t. Rebuild attribution server-side per location.

You run ad spend across 50-1,500 retail locations. iOS 14 ATT cut Meta attribution 30-40%. Chrome cookie deprecation is cutting Google + ITP attribution further. Your POS shows revenue your platform reports do not credit, your bid algorithms underspend on cohorts they cannot see converting, and the per-location join key was never in the platforms to begin with — so even the attribution you do recover does not reconcile to per-location P&L. Server-side per-location attribution event emission to Google Ads + Meta CAPI + TikTok + LinkedIn + Snap + Pinterest with consent-mode v2 and the per-location join key closes both gaps.

Audit your attribution gap (Tier 1)Explore Fractional CMO operation

Published May 30, 2026

Two attribution gaps, one server-side rebuild

The platform-attribution gap. iOS 14 ATT broke Meta’s client-side pixel attribution by 30-40%. Chrome cookie deprecation is doing the same to Google + ITP attribution. Ad-blocker users were already breaking it. The platforms still serve ads; they just can’t credit conversions back to the click that produced them. Your bid algorithms see fewer conversions than actually happened and underspend on cohorts that were converting. Or overspend on cohorts that were not. Either way the optimization signal is wrong.

The per-location reporting gap. Even if every server-side event landed cleanly, the per-location join key was never on the platform side. Google + Meta + TikTok + LinkedIn + Snap + Pinterest report conversions at the campaign/ad-set/ad level, not at the per-store level. Per-location P&L cannot reconcile to per-location ad spend without the join key on every event. Per-market budget allocation defaults to brand-level rollups that hide per-location performance distribution.

The server-side rebuild closes both gaps. Server-side emission ships events from your infrastructure carrying first-party identifiers (email, phone, address, click ID) the platforms match against logged-in user data. The per-location join key ships with every event so per-location reporting joins ad spend to ad-attributed revenue per market.

We’ve built the 6-platform server-side rebuild for multi-location operators. Here’s what we know.

You probably already use Google Tag Manager, Segment, Tealium, Heap, RudderStack, or a similar event-collection platform. Each is good at the event-collection primitive. The gap is the 6-platform server-side emission layer above the event collector: per-platform CAPI integration with deduplication keys, consent-mode v2 integration with your CMP, per-location join key on every event, and per-platform match-rate monitoring. We bring the per-platform CAPI maintenance runbook and the consent-mode v2 integration playbook.

We have built this for multi-location retail operators across verticals. We know which platforms produce the highest match-rate per first-party identifier set (email + phone + address + click ID is the standard stack; per-platform match-rate differs). We know which consent signals are required per platform under consent-mode v2. We bring the per-platform match-rate benchmark so the first 30 days of emission produces calibrated optimization rather than degraded signal.

How we get from broken pixels to server-side per-location

Step 1 — Tier 1 AI Readiness Assessment (2-3 weeks). We audit your current pixel + CAPI surface. We sample 30-60 days of POS-recorded revenue against platform- reported conversions per location. Output: the per-platform attribution-gap baseline, the per-location join-key map, the consent-mode v2 integration plan, and the per-platform CAPI rollout sequence.

Step 2 — Tier 2 AI Swarm Setup Sprint (4-8 weeks). We build the attribution layer end-to-end: per-platform CAPI integration (Google Ads, Meta, TikTok, LinkedIn, Snap, Pinterest), deduplication-key generation, consent-mode v2 integration with your CMP, per-location join-key wiring, per-platform match-rate monitoring. Your engineering + media team receives the running system, all source code, all credentials.

Step 3 — Tier 3 Fractional CMO with AI Swarm ( 6-month minimum, 1-2 days/wk). We operate the attribution layer in production. Per-platform match-rate reviewed weekly. Per-platform CAPI maintenance as the platforms ship API updates. Per-platform attribution-gap audit monthly. Per-location attribution rollup published weekly to marketing + finance leadership.

What changes for you

You stop watching the bid algorithms underspend on the cohorts your POS shows are converting. The server-side events ship the conversions back to the platforms; optimization signal returns.

You stop having the same conversation with finance about why marketing reporting and POS reporting do not agree. The per-location join key + consent-mode v2 produce numbers both teams can read against each other.

You can answer the question your CFO asks every monthly business review: which platforms produced revenue at which locations, and at what marginal acquisition cost. The per-platform per-location attribution record is the answer.

You can launch in a new market with the per-platform CAPI integration pre-configured and the per-location join key auto-populated from your inventory of stores.

Frequently asked

What does server-side attribution event emission actually do?

Server-side attribution event emission sends conversion events directly from your server to ad-platform APIs (Google Ads Enhanced Conversions, Meta Conversions API, TikTok Events API, LinkedIn Conversions API, Snap Conversions API, Pinterest Conversions API) rather than relying on client-side pixels. The client-side pixels are the layer iOS 14 ATT broke (Meta lost 30-40% of attribution); the layer Chrome cookie deprecation is breaking further (Google + ITP); the layer your ad-blocker users were already breaking. Server-side emission survives all three because the event ships from your infrastructure carrying first-party identifiers (email, phone, address, click ID) the ad platforms can match against logged-in user data — server to server, no pixel required.

Why is per-location attribution a separate problem from server-side attribution?

Server-side attribution closes the platform-attribution gap. Per-location attribution closes the per-location reporting gap. A single national ad campaign drives a conversion at the Phoenix store; the conversion ships server-side to Google + Meta + TikTok; the platforms credit the campaign. Without the per-location join key on the event, the per-location P&L cannot reconcile to per-location ad spend at the per-market budget allocation step. The join key (location_id, derived from the order address or the in-store transaction record) ships with every event so the per-location reporting layer can join ad-spend to ad-attributed revenue per market.

What is consent-mode v2 and how does it integrate?

Consent-mode v2 is Google’s framework for honoring user consent state in ad-platform integrations. Each event carries the consent signal (ad-personalization, ad-user-data, ad-storage) the user actually granted. The ad platforms degrade gracefully to modeled conversions when consent is partial rather than over-counting non-consented conversions or under-counting consented ones. Server-side emission integrates consent-mode v2 by reading the consent signal from your CMP (OneTrust, TrustArc, Iubenda) and attaching it to every outbound event. Without consent-mode v2 integration the server-side rebuild creates a different kind of attribution gap.

What does Completions commit to on Tier 3 if we run the attribution layer in production?

Tier 3 process commitments include: per-event server-side emission to all 6 ad platforms with deduplication keys; per-location join key attached to every event; consent-mode v2 integration maintained as the ad platforms ship API updates; per-platform match-rate reviewed weekly with media leadership; per-platform attribution-gap audit monthly comparing server-side reported conversions to POS-recorded revenue; per-location attribution rollup published weekly to marketing + finance leadership. We commit to the operating discipline. Per-platform match-rate precision is tuned per stack and recorded as engagement KPIs.

Who owns the event store, the CAPI credentials, and the CMP integration post-engagement?

Your team owns the event store, the per-platform CAPI credentials, the CMP integration, the per-location join key map, the deduplication key store, and the engineering credentials. Completions owns the orchestration knowledge: the per-platform CAPI maintenance runbook, the deduplication-key tuning history, the consent-mode v2 integration playbook. At engagement end we transition operational ownership back to your team over 30-60 days with documented handover.

How does the attribution layer connect to the rest of the marketing + revenue stack?

The attribution layer subscribes upstream to the change-event-emission stream (per-customer state changes that fire conversion events), to your POS (in-store transactions), to your ecommerce platform (online conversions), and to the routing-rules engine (lead-state outcomes). It publishes downstream: server-side events to 6 ad platforms via CAPI, per-location attribution events into the multi-location reporting layer, and per-platform match-rate signals into the per-state action-decisioning layer (low match-rate locations route differently). Four upstream feeds, three downstream consumers, one attribution contract.

Start with the audit

Tier 1 AI Readiness Assessment (2-3 weeks): we audit your current pixel + CAPI surface, sample 30-60 days of POS-recorded revenue against platform-reported conversions per location, and produce the per-platform attribution-gap baseline + per-location join-key map + consent-mode v2 integration plan + per-platform CAPI rollout sequence. If you decide to build, Tier 2 ships the attribution layer. If you decide to operate it with us, Tier 3 runs the weekly match-rate monitoring + monthly attribution-gap audit in production. You choose the next step at each gate.