30 min read

Checklist: Deploy Attribuly Retargeting Stack (30–90 Days)

Practical 30/60/90-day checklist to deploy the Attribuly retargeting stack on Shopify—tracking, Consent Mode v2, Meta & TikTok dedup, Klaviyo flows, SLAs & KPIs.

Checklist: Deploy Attribuly Retargeting Stack (30–90 Days)

If you run a Shopify brand and need server-side tracking, consented data collection, and fast retargeting activation, this checklist gives you a pragmatic path to go live in 30–90 days with clear acceptance criteria. We assume Stack A: Shopify + GTM server-side (sGTM) + Meta + TikTok + Klaviyo + GDPR/CCPA via Cookiebot or OneTrust. All milestones map to non‑negotiable 30/60/90 KPIs, and each item includes inputs, expected output, a test method, and an owner.

Assumptions and 30/60/90 acceptance KPIs for the Attribuly retargeting stack

  • Stack A: Shopify + sGTM + Meta + TikTok + Klaviyo + CMP (Cookiebot/OneTrust) respecting GDPR/CCPA.

  • Default onboarding model and SLA: Standard — 30‑day go‑live, business‑hours support, weekly check‑ins.

Milestone

Target KPI

Acceptance Criteria

Day 30

Purchase event match rate 85–90%

Meta EMQ ~8/10 for Purchase; TikTok Events API + Pixel dedup working; hashed identifiers present

Consent opt‑in >70%

CMP banner live; Consent Mode v2 signals visible (gcs/gcd); consent-based tag firing verified

Day 60

First retargeting ROAS ≥1.5

DPA audiences live; complete server Purchase payloads (value/currency/order_id); ROAS validated in Ads Managers

Cart abandoner list ≥2,000

Klaviyo abandoned cart/browse flows live; list growth tracked; dynamic product blocks enabled

Day 90

Blended retargeting ROAS ≥2.0

Retargeting scaled; diagnostics clean; attribution windows aligned

CAC ↓15%

Acquisition efficiency tracked month-over-month; redundant spend trimmed

Email revenue from abandoners +25%

Lift measured against pre‑onboarding baseline; flow optimization documented

Quick prerequisites and access

  • Shopify: Admin access; theme or Customer Events permissions; Checkout Extensibility as applicable.

  • GTM: Web and Server containers created; server endpoint on verified custom subdomain (first‑party cookies); Preview access.

  • Platforms: Meta Business Manager + Events Manager; TikTok Ads Manager + Events; Klaviyo account with Shopify integration.

  • CMP: Cookiebot or OneTrust configured; categories mapped to Google Consent Mode v2; banner deployed.

  • Governance: Owners assigned (Marketing Ops, Engineer, Agency); weekly check‑ins scheduled; rollback plan outlined.

For server-side container setup and custom domain verification, follow trusted resources such as the Stape guides on Shopify server-side tagging and the Analytics Mania introduction to server-side tagging.

Day 0–30: Tracking, identity, and consent go-live (target: match rate 85–90%; consent opt‑in >70%)

Tracking & identity

Item

Inputs

Expected Output

Test Method

Owner

Acceptance

sGTM endpoint on custom subdomain

DNS, SSL, GTM server container

First‑party cookies and stable identifier routing

sGTM Preview; DNS verification; cookie inspection

Engineer

Endpoint verified; cookies set

Meta Pixel + CAPI hybrid with dedup

Pixel installed; CAPI via sGTM; event_id generated once

Browser + server Purchase merged (no double count)

Meta Events Manager “Received Events”; dedup flag; timing window

Engineer

Dedup confirmed on Purchase — see Meta’s guidance on the Conversions API and dataset quality

Advanced Matching (hashed PII)

Email/phone capture; SHA‑256 hashing

Higher EMQ for Purchase (~8/10)

EMQ Diagnostics score; warnings resolved

Marketing Ops

EMQ trend ≥8/10 — according to Meta’s dataset quality scoring, hashed identifiers lift match quality

TikTok Pixel + Events API

Pixel + server payloads with event_id

Deduped Purchase; DPA‑ready params

TikTok diagnostics; dedup warnings cleared

Engineer

Dedup stable; payload complete — TikTok’s Events API docs describe required identifiers and parameters

Preserve click IDs (fbc/fbp; ttclid)

First‑party cookies; checkout propagation

Better match and attribution continuity

Network log checks during checkout

Engineer

IDs captured end‑to‑end

Revenue payload completeness

value, currency, order_id on Purchase

ROAS calculable in Ads

Ads Manager validates revenue on server events

Marketing Ops

Payload fields present

Practical primer: see the “validate multi‑touch in 30 days” process to dual‑run browser+server events and verify dedup and diagnostics using daily checks.

Consent playbook and privacy signals

Item

Inputs

Expected Output

Test Method

Owner

Acceptance

CMP banner (Cookiebot/OneTrust)

CMP installed; categories defined

Visible banner; logs of consent decisions

UI check; CMP logs

Marketing Ops

Banner live, localized

Google Consent Mode v2 wiring

Map CMP categories to ad_storage, analytics_storage, ad_user_data, ad_personalization

Consent signals propagate (gcs/gcd) to sGTM

Inspect requests; GA4 DebugView; sGTM client logs

Engineer

gcs/gcd present and respected — see Google’s Consent Mode overview and server-side implementation guidance

Shopify Customer Privacy API sync

Read/set consent per region

Shopify scripts honor consent defaults

Theme/Customer Events test; consent states reflected

Engineer

Region logic validated — per Shopify’s Customer Privacy API documentation

Consent-based tag firing

Suppress ad tags on deny; enable on allow

Privacy-respecting behavior

GTM Preview with varied consent states

Engineer

Tags obey consent

Opt‑in rate tuning

Copy/UX variants; regional rules

>70% opt‑in without dark patterns

A/B CMP banner; weekly rate review

Marketing Ops

≥70% by Day 30

Orientation for beginners: review the consent and identifiers starter to understand Consent Mode v2 and Shopify Customer Privacy basics.

Day 31–60: Retargeting activation and Klaviyo orchestration (target: ROAS ≥1.5; ≥2,000 abandoners)

Item

Inputs

Expected Output

Test Method

Owner

Acceptance

DPA setup (Meta/TikTok)

content_ids, contents, catalog feed

Dynamic product retargeting audiences

Catalog diagnostics; audience population

Marketing Ops

Audiences populated

High‑intent segments

ViewContent, AddToCart, InitiateCheckout

Warm audiences in Ads Managers

Events diagnostics; audience size trends

Marketing Ops

Segments > minimums

Attribution lookback harmonization

Align windows across platforms

Consistent performance readouts

Compare attribution windows; document

Marketing Ops

Windows aligned

Klaviyo abandoned cart flow

Shopify events; 2–4h first touch; dynamic blocks

Growing abandoner list; revenue attribution

Flow analytics; benchmarks

Marketing Ops

≥2,000 abandoners — Klaviyo’s best practices recommend 2–4 hour first touch and subsequent sends at 24–48 hours and 3–7 days

Dedup health checks

event_id across pixel+server

Clean diagnostics; no double counts

Weekly Events Manager review

Engineer

Dedup warnings resolved

ROAS validation

Complete Purchase payloads

Ads ROAS ≥1.5 by Week 8

Ads dashboards vs. Shopify revenue

Marketing Ops

ROAS threshold met

Day 61–90: Scale, optimize, and govern (target: blended ROAS ≥2.0; CAC ↓15%; email revenue +25%)

Item

Inputs

Expected Output

Test Method

Owner

Acceptance

Creative iteration & audience scaling

New creatives; broadened lookbacks

Sustained ROAS and audience growth

Creative tests; audience overlap checks

Marketing Ops

Blended ROAS ≥2.0

CAC reduction plan

Budget reallocation; exclude waste

Lower CAC vs. baseline

CAC trend analysis month‑over‑month

Marketing Ops

CAC ↓15%

Email/SMS optimization

Subject lines; offers; channel mix

Higher abandoner flow revenue

Flow A/B tests; revenue lift vs. baseline

Marketing Ops

+25% revenue

Monthly QA cadence

Diagnostics review; payload audits

Early detection of drift

EMQ, dedup, consent, payload scorecard

Engineer

KPI stability documented

Governance & SLA adherence

Weekly check‑ins; change windows

Fewer incidents; faster recovery

Incident logs; rollback drills

Ops Lead

SLA compliance recorded

Documentation & handoff

Playbook, owners, versions

Operable stack post‑onboarding

Repo/wiki updates; sign‑off meeting

Ops Lead

Handover complete

Troubleshooting matrix (symptom → test → fix)

  • Duplicate Purchase events: Check event_id parity and time windows in Meta/TikTok diagnostics; fix by generating event_id once client‑side and passing to both pixel and server.

  • Low EMQ on Purchase: Confirm hashed email/phone in Pixel/CAPI; ensure Advanced Matching enabled; verify consent allows ad_user_data; see Meta’s dataset quality guidance for scoring.

  • Missing revenue in server events: Ensure value, currency, order_id in Purchase payloads; test via sGTM Preview and Ads Manager logs.

  • Lost click IDs (fbc/fbp; ttclid): Verify custom sGTM domain sets first‑party cookies; inspect checkout redirects retaining IDs; for background on privacy impacts, see the iOS 17 attribution accuracy guide.

  • Blocked tags under GDPR: Inspect gcs/gcd signals; confirm CMP fires before tags; map CMP categories correctly to Consent Mode v2; reference Google’s Consent Mode documentation.

  • Small retargeting audiences: Validate event coverage (ViewContent/AddToCart/InitiateCheckout); check DPA catalog integrity; confirm consent allows marketing.

  • ROAS below threshold: Diagnose audience quality, creative fatigue, missing revenue parameters; iterate creatives and ensure payload completeness.

  • Klaviyo list growth stalls: Confirm triggers (Added to Cart/Browse Abandon); review send‑times and dynamic product blocks; test opt‑in capture; consult Klaviyo benchmarks.

Practical example: unifying identifiers during Day‑30 onboarding

Disclosure: Attribuly is our product.

During Day‑30 onboarding, you can use Attribuly to unify Shopify, Meta, TikTok, and Klaviyo identifiers without changing your growth playbook. For instance, generate and persist a single event_id client‑side, pass it to both browser and server streams, and ensure hashed email/phone are captured where consent allows. Then validate in diagnostics: Meta’s Events Manager should show Purchase deduped with EMQ trending toward ~8/10, while TikTok’s logs confirm overlapping events are deduped. This practical workflow helps teams reach the Day‑30 targets (85–90% purchase match rate; >70% consent opt‑in) by making identity and payload checks part of daily QA rather than a one‑off task. For a pilot outline, see the multi‑touch validation guide: Validate multi‑touch attribution in 30 days.

Next steps and resources

Owner cadence reminder: Maintain weekly check‑ins (Standard SLA), run a monthly QA scorecard across EMQ, dedup, consent, and payload completeness, and document changes with versioned releases.