ChatGPT Prompts to Build a Complete Ecommerce Tracking Plan (2025)

What You’ll Build (in about 90–120 minutes)

You’ll ship a working, end-to-end tracking plan for a web shop: GA4 ecommerce events, a clean data layer, GTM mappings, optional CDP routing (Segment/RudderStack/Snowplow), consent gating (GDPR/CCPA, Consent Mode v2), QA/validation, reporting dashboards, and governance. Every section gives you copy‑paste prompts and verification checklists.

Prerequisites

• Access: GA4 property + admin, GTM container (or equivalent), your CMP (consent platform), and read access to your ecommerce backend.
• Roles: 1 analyst/PM, 1 web developer, optional CDP/BI owner.
• Difficulty: Intermediate. Some dev help needed for data layer and CMP hooks.

Before you start, skim Google’s current specs for ecommerce events and parameters so your mental model matches reality: see Google’s developer overview in the Set up ecommerce (GA4) guide and the authoritative GA4 event parameters reference (Google Developers, 2024–2025).

How to use this guide

• Copy a prompt, paste into ChatGPT (or your LLM), provide your context in the placeholders, and ask for the exact outputs/tables/code you need.
• After each module, run the “You’re done when…” checklist to verify.

Prompts: Discovery & KPI Definition

Time: 10–15 minutes | Owner: Analyst/PM | Difficulty: Easy

Copy‑paste prompt

You are an ecommerce analytics architect. Using the context below, draft a one-page tracking plan charter.

Context
- Business model: {{D2C | marketplace | B2B}}
- Catalog size: {{N SKUs}}
- Average order value (AOV): {{amount + currency}}
- Key channels: {{paid search, social, email, affiliates, organic}}
- Markets & consent: {{EEA, US-CA/CCPA, ROW}}
- Platforms: {{Shopify | Magento | Custom}}, tag manager: {{GTM | Tealium | Adobe Launch}}, analytics: {{GA4}}, CDP: {{None | Segment | RudderStack | Snowplow}}, data warehouse/BI: {{BigQuery/Looker Studio | Snowflake/Tableau | Power BI}}

Deliverables
1) Goals and KPIs (conversion rate, revenue, AOV, ROAS, retention, LTV, refund rate) with target ranges.
2) Funnel definition (impressions → product views → add to cart → checkout → purchase → refund), with key questions per stage.
3) Event scope (GA4 ecommerce recommended events) and identity strategy (anonymous_id → user_id after auth; cross-domain if needed).
4) Non-functional: consent policy baseline (default denied in EEA), data retention setting ({{2 or 14}} months), QA cadence (weekly smoke + monthly audit).
5) Risks & assumptions.

Verification

• Goals/KPIs listed with owners and target ranges.
• Funnel stages clear and traceable to events.
• Identity and consent assumptions documented.

Citations to confirm scope and terminology: Google’s About ecommerce in GA4 (Help, 2024–2025) and Recommended events for GA4 (Developers).

Prompts: GA4 Ecommerce Event Taxonomy (+ examples)

Time: 15–20 minutes | Owner: Analyst/PM | Difficulty: Medium

Why this matters: GA4 ecommerce relies on correctly structured events and an items[] array. Currency and transaction_id are critical for revenue/refund integrity, as shown in Google’s Set up ecommerce (Developers, 2024–2025).

Copy‑paste prompt

You are a GA4 solutions architect. Create a GA4 ecommerce taxonomy for our website using Google’s recommended events. Output a Markdown table and JSON payload examples.

Inputs
- Currency: {{USD | EUR | GBP}}
- Tax & shipping: {{tax_included=true/false}}, {{shipping strategy}}
- Promotions: {{we use onsite banners/UTMs}}

1) Table with columns: Event name, When to trigger, Event-scoped params, Item-scoped params, Notes.
2) Include these events at minimum: view_item, view_item_list, select_item, add_to_cart, remove_from_cart, view_cart, begin_checkout, add_shipping_info, add_payment_info, purchase, refund, view_promotion, select_promotion.
3) Provide example JSON payloads for add_to_cart, begin_checkout, purchase, refund. Use an items array with item_id, item_name, item_category, item_brand, item_variant, price, quantity, coupon, discount as applicable. Include currency, value, tax, shipping, transaction_id where relevant.
4) Highlight required/strongly recommended fields per Google’s docs.

You’re done when

• The taxonomy table covers all core events above.
• JSON examples include items[] with correct nesting and currency/value/transaction_id as applicable.

For parameters and items[] structure, cross‑check the GA4 event parameters (Google Developers) and the Recommended events catalog (Google Developers).

Prompts: Data Layer Specification (copy‑paste JSON)

Time: 20–30 minutes | Owner: Developer + Analyst | Difficulty: Medium–Hard

Tip: Build a single, predictable dataLayer schema that feeds both GA4/GTM and your CDP. Google’s validation advice in the Validate ecommerce with DebugView (Developers, 2024–2025) helps you check payloads quickly.

Copy‑paste prompt

You are a web analytics implementation engineer. Produce a data layer specification for our ecommerce site. Output:
1) A JSON schema for PDP (product view), PLP/list view, cart, checkout steps, purchase, and refund events.
2) For each event, provide a push example to window.dataLayer with event name and an ecommerce object compatible with GA4 items[].
3) Mark required vs optional fields and data types.
4) Include user_id if authenticated (no PII) and a session_id/anonymous_id placeholder.

Assumptions
- Currency: {{USD}}
- Prices stored {{with_tax | without_tax}}
- Platform: {{Shopify | Magento | Custom}}

Deliverables
- JSON schema and example dataLayer pushes for: view_item, view_item_list, select_item, add_to_cart, remove_from_cart, view_cart, begin_checkout, add_shipping_info, add_payment_info, purchase, refund, view_promotion, select_promotion.

You’re done when

• Each event has a concrete dataLayer.push example with an ecommerce.items array.
• purchase includes transaction_id, value, currency, tax, shipping; refund includes transaction_id and item-level array for partial refunds.

Prompts: GTM/GA4 Tag Mapping

Time: 20–25 minutes | Owner: Analyst/Developer | Difficulty: Medium

Reference: GA4 tags via GTM require a GA4 Configuration tag and event tags per event. See Google’s Ecommerce events setup overview (Help) and Enhanced measurement (Help) for non‑ecommerce events.

Copy‑paste prompt

You are a GTM implementation lead. Map our data layer to GA4 tags in GTM.

Inputs
- GA4 Measurement ID: {{G-XXXX}}
- Domains for cross‑domain: {{example.com, checkout.example.com}}
- User ID variable name (if any): {{dlv.user_id}}

Deliverables
1) List the GA4 Configuration tag settings, including Fields to Set for allowLinker=true and linker domains for cross‑domain. Include instructions to set user_id when available.
2) A mapping table: dataLayer key → GA4 event name → parameter name (event‑scoped vs item‑scoped) → trigger.
3) Trigger logic per event (DOM events or custom events), and sequencing to ensure GA4 loads before ecommerce pushes.
4) QA notes for DebugView and Realtime verification.

Check cross‑domain and user ID with Google’s Cross‑domain measurement docs (Help, 2024–2025) and User‑ID setup (Help). If using GTM for linker, set Fields to Set as described in Google Tag Manager Help’s linker setup article.

You’re done when

• GA4 Configuration tag is configured; _gl linker parameter appears when you navigate between allowed domains.
• DebugView shows all ecommerce events with items[] details.

Prompts: CDP Routing (Segment/RudderStack/Snowplow)

Time: 20–30 minutes | Owner: CDP/Developer | Difficulty: Medium

Segment mapping aligns its Ecommerce Spec to GA4 via the Actions destination; see the Segment GA4 Actions destination docs and the Segment Spec overview.

Copy‑paste prompt (Segment)

You are a Segment implementation specialist. Convert our GA4 ecommerce taxonomy into Segment calls and configure the GA4 (Actions) destination.

Deliverables
1) For each GA4 event (view_item, add_to_cart, begin_checkout, purchase, refund, etc.), output the equivalent Segment analytics.track calls using the Segment Ecommerce Spec names (e.g., Product Viewed, Product Added, Checkout Started, Order Completed, Order Refunded) with properties that include a products array.
2) Provide a mapping table: Segment event/property → GA4 event/parameter (including items array mapping and currency/value/transaction_id).
3) Include analytics.identify and alias guidance for user identity.
4) Include QA steps to validate GA4 destination mappings and observe events in GA4 DebugView.

Copy‑paste prompt (RudderStack)

You are a RudderStack implementation specialist. Produce rudderanalytics.track/identify calls equivalent to our GA4 ecommerce taxonomy.

Deliverables
1) Track calls for Product Viewed, Product Added, Checkout Started, Order Completed, Order Refunded with a products array and standard properties (product_id, name, category, price, quantity, currency).
2) Propose a GA4 destination mapping plan and call out any fields requiring manual mapping in RudderStack’s UI.
3) Provide identity rules (identify/alias) and QA steps to validate the GA4 destination and GA4 DebugView.

Note: If any mapping ambiguity exists, instruct me to consult the current RudderStack GA4 destination docs and verify payloads end‑to‑end using DebugView.

Copy‑paste prompt (Snowplow)

You are a Snowplow implementation engineer. Using Snowplow’s browser tracker and ecommerce plugin, define the events and entities to mirror our GA4 taxonomy.

Deliverables
1) Initialization snippet using newTracker with the Snowplow Ecommerce plugin.
2) Event examples for product_viewed, product_added, checkout_started, order_completed, order_refunded with attached ecommerce context entities (product, cart, order).
3) Notes on modeling (dbt snowplow‑ecommerce package) and QA steps.

Docs for parity: Snowplow’s ecommerce plugin and entities in the Snowplow ecommerce events docs and OOTB ecommerce entities. For RudderStack JS SDK basics, see RudderStack JavaScript SDK docs.

Prompts: Consent & Privacy (GDPR/CCPA)

Time: 15–25 minutes | Owner: Developer/Privacy | Difficulty: Medium–Hard

Google’s Consent Mode v2 requires four consent keys (ad_storage, analytics_storage, ad_user_data, ad_personalization). Default to denied in EEA and update after user action, per Google’s Consent Mode concepts (Developers, 2024–2025) and Ads Help enforcement updates.

Copy‑paste prompt

You are a privacy‑aware tagging engineer. Draft a consent policy and GTM implementation plan.

Deliverables
1) A consent matrix with rows (Analytics, Ads, Functional) and columns (Purpose, Consent keys used, Default state by region, Tags gated, Data retention setting).
2) GTM implementation steps: Fire CMP on Consent Initialization; set default denied via gtag('consent','default', {...}); update on user action via gtag('consent','update', {...}).
3) Example code for default and update calls, and GTM sequencing.
4) QA plan: Use Tag Assistant consent debugging; test EEA vs non‑EEA; verify that GA4 behaves with cookieless pings when denied.

Reference the Google guidance for CMP/GTM sequencing and debugging: Consent debugging guide (Developers) and Consent Initialization behavior (Ads Help). For key behaviors when consent is denied vs granted, see Google Ads Help: Consent Mode keys and behavior.

Prompts: QA & Validation (DebugView, test orders)

Time: 20–30 minutes | Owner: Analyst/QA | Difficulty: Medium

Use Google’s official validation paths: DebugView and Realtime, plus Chrome DevTools and the GA4 Measurement Protocol verification guidance. See the Validate ecommerce with DebugView (Google Developers) and Verify implementation via Measurement Protocol (Developers).

Copy‑paste prompt

You are a QA lead. Create a test plan for ecommerce tracking.

Deliverables
1) Test cases covering: view_item, add_to_cart, begin_checkout, add_shipping_info, add_payment_info, purchase, refund. Include steps, expected GA4 DebugView parameters, and acceptance criteria.
2) A checklist for consent states (denied → granted) and region simulation.
3) A cross‑domain stitching test plan (verify _gl linker, session continuity) and user_id checks.
4) A data quality monitor: daily sanity checks (events, revenue vs backend within ±5%), and alerts for drops.

Practical tip: When testing cross‑domain, confirm the _gl parameter appears in the URL as you navigate—per Google’s Cross‑domain measurement doc. Use Chrome’s Network panel to inspect GA4 requests as recommended in Google’s validation doc above.

Prompts: Dashboards & Reporting

Time: 15–20 minutes | Owner: Analyst | Difficulty: Easy–Medium

Synthesize your reporting using proven ecommerce questions (funnels, cohorts, merchandising, attribution). For inspiration and metric choices, review the Saras Analytics ecommerce dashboard guide (2025) and Optimize Smart’s tutorials on Explorations and the user purchase journey.

Copy‑paste prompt

You are a reporting lead. Design a minimal KPI dashboard + 3 Explorations.

Deliverables
1) KPI board: Revenue, CR, AOV, ROAS, Refund rate, New vs returning revenue, Top products, Top lists/promotions.
2) Explorations: (a) Funnel (product view → purchase), (b) Cohort retention (by first purchase month), (c) Merchandising (list → product → cart efficiency).
3) Dimensions/filters: device category, default channel group, campaign, country, consent state (if available), user type (new/returning), promo exposure.
4) Definitions: state precisely which GA4 metrics/dimensions to use and which event parameters feed them.

Prompts: Documentation & Governance

Time: 15–25 minutes | Owner: Analyst/PM | Difficulty: Easy–Medium

Include data retention and deletion processes, plus ownership. For background, see GA4’s data retention controls (Help) and User Deletion API overview (Developers), and privacy controls in GA4 user data help.

Copy‑paste prompt

You are the analytics program owner. Create a documentation and governance pack.

Deliverables
1) Tracking plan doc template with sections: Charter, Event taxonomy, Data layer spec, GTM/CDP mappings, Consent policy, QA plan, Reporting catalogue.
2) RACI: roles for tagging, QA, reporting, privacy, and CDP/warehouse.
3) Change control: versioning, GTM container IDs/versions, data layer schema versions, rollback plan.
4) Maintenance cadence: weekly QA, monthly audit (consent/cross‑domain), quarterly deep‑dive, annual policy review.
5) Data lifecycle: retention setting ({{2 or 14}} months), deletion workflow (User Deletion API, BigQuery/DW deletion notes), and who approves.

You’re done when

• Ownership is assigned, versioning is written down, and retention/deletion are codified with approvers.

Troubleshooting Appendix (symptoms → root causes → fixes)

Use these as seeds for your own LLM prompts.

• Missing purchase revenue

  • Likely causes: no transaction_id or currency; purchase fires before GA tag; consent denied in EEA; wrong property.
  • Fix: Verify event contains transaction_id and currency/value per Google’s purchase spec; check DebugView and network; confirm Consent Mode updates before tags.

• Item arrays not populating

  • Likely causes: items placed at wrong level; malformed JSON; dataLayer push races with tag load.
  • Fix: Ensure items[] is nested under ecommerce; validate with Google’s DebugView guide; delay tags until data present.

• Cross‑domain session breaks

  • Likely causes: linker not configured; navigating via target=_blank without linker; domains missing.
  • Fix: Add domains via GA admin or GTM linker fields; validate _gl parameter as in cross‑domain docs.

• Refunds not matching backend

  • Likely causes: partial refund items omitted; wrong transaction_id; currency mismatch.
  • Fix: Send item‑level quantities/values in refund; confirm transaction_id matches order; include currency as per refund spec.

From experience: Partial refunds often fail because order item IDs don’t match storefront SKUs; map to a stable item_id that exists in both analytics and your backend.

Tools/Stack (neutral parity)

Use whatever fits your team’s skills and governance. A typical stack includes:

• Analytics: GA4; alternatives/complements: Mixpanel, Amplitude.
• Tag management: Google Tag Manager (web/server); alternatives: Tealium iQ, Adobe Launch.
• CDP/collection: Segment, RudderStack, Snowplow.
• Data warehouse/BI: BigQuery, Snowflake; Looker Studio, Tableau, Power BI.
• Commerce/ERP source of truth for orders/inventory: Shopify, BigCommerce, Magento; ERPs like NetSuite or Odoo; and WarpDriven as an AI‑first ERP option. Disclosure: WarpDriven is our product.

Next Steps

• Run each prompt with your context. Paste outputs into your tracking plan doc and GTM/CDP tickets.
• Implement data layer pushes on staging, wire up tags, and validate via DebugView before going live.
• Schedule the governance cadence you defined and set up daily/weekly monitors.

If you follow the prompts and hit the checkpoints, you’ll have a GA4‑first, privacy‑aware ecommerce tracking plan that stands up to audits and scales with your stack.