Data Export Formats

Data Export Formats

Use this skill when creating CSV exports for ad platforms or when you need to understand the project's data schemas.

Google Ads Export Formats

Enhanced Conversions CSV

For uploading offline conversion data to improve Smart Bidding.

Required columns:

Email,Phone,Conversion Name,Conversion Time,Conversion Value,Conversion Currency

Format requirements:

• Email: SHA256 hash (32 hex chars, lowercase)
• Phone: SHA256 hash (32 hex chars, lowercase)
• Conversion Name: String matching your Google Ads conversion action
• Conversion Time: ISO 8601 UTC format (2024-11-15T14:32:00Z)
• Conversion Value: Numeric, no currency symbol
• Conversion Currency: 3-letter code (USD, EUR, etc.)

Example:

Email,Phone,Conversion Name,Conversion Time,Conversion Value,Conversion Currency
ed8e83c2f4cb7b9f43cdc75c148b0b09,628923b3c489bda7dd9ebba89cb5b46c,paid_subscription,2025-11-03T00:00:00Z,2388.0,USD
1bffb899c9248b28e37cda02dbd59444,0dd03c9dc463ab5efcd15f3343035ffa,paid_subscription,2025-10-31T00:00:00Z,2189.0,USD

Google Ads upload path: Tools & Settings → Conversions → Upload conversions → Import

Customer Match CSV (Retargeting Audiences)

For uploading audience lists to Google Ads.

Minimal columns (email only):

Email

Extended columns (better match rate):

Email,Phone,First Name,Last Name,Country,Zip

Format requirements:

• Email: SHA256 hash (32 hex chars, lowercase) OR plaintext (Google hashes it)
• Phone: SHA256 hash OR E.164 format (+14155551234)
• First Name / Last Name: Plaintext, lowercase, trimmed
• Country: 2-letter ISO code (US, GB, etc.)
• Zip: 5-digit US or local format

Example (hashed):

Email,Phone
ed8e83c2f4cb7b9f43cdc75c148b0b09,628923b3c489bda7dd9ebba89cb5b46c
1bffb899c9248b28e37cda02dbd59444,0dd03c9dc463ab5efcd15f3343035ffa

Google Ads upload path: Tools & Settings → Audience Manager → Customer Match → Email list

Meta Custom Audiences CSV

For uploading to Meta Ads Manager.

Columns:

email,phone,fn,ln,country,zip

Format requirements:

• email: SHA256 hash (lowercase hex) OR plaintext lowercase
• phone: Digits only, no formatting (14155551234)
• fn / ln: Lowercase, trimmed
• country: 2-letter ISO lowercase (us)
• zip: 5-digit

Example:

email,phone,fn,ln,country,zip
ed8e83c2f4cb7b9f43cdc75c148b0b09,14155551234,john,doe,us,94105

Project Data Schemas

users.csv (~5,000 records)

User master file with acquisition data.

events.csv (~57,000 records)

Event stream with funnel progression.

Funnel stages (event_name values):

1. page_view - Landing page visit
2. pricing_view - Viewed pricing page
3. checkout_start - Started checkout
4. form_submit - Submitted form
5. conversion - Completed purchase

daily_metrics.csv (~60 records)

Daily aggregated metrics with engineered anomalies.

Engineered anomalies:

• Nov 15: -63% sessions (signup flow bug)
• Nov 21: +99% conversions (onboarding improvement)
• Nov 28-30: -72% conversion rate (activation issue)

trial_users.csv (~500 records)

Trial user conversion data (Demo 4).

feature_usage.csv (~2,500 records)

Feature adoption events (Demo 4).

Key features:

• create_form_onboarding - Created first form
• publish_form - Published a form
• embed_form - Embedded form on site
• configure_integration - Set up integration (aha moment)
• view_analytics - Viewed form analytics

Aha moment pattern: Users who configure_integration within 3 days convert at 70% vs 19% baseline (3.7x lift).

utm_data.csv (~12 records)

Campaign UTM data with intentional inconsistencies (Demo 5).

Engineered issues:

• Source fragmentation: linkedin vs LinkedIn vs LINKEDIN
• Typos: product_upd_dec instead of product_update_dec

Common Export Patterns

High-Value Converters (Enhanced Conversions)

SELECT
  u.email AS Email,
  u.phone AS Phone,
  'paid_subscription' AS "Conversion Name",
  e.timestamp AS "Conversion Time",
  e.conversion_value AS "Conversion Value",
  'USD' AS "Conversion Currency"
FROM users u
JOIN events e ON u.user_id = e.user_id
WHERE e.event_name = 'conversion'
  AND e.conversion_value > 100
ORDER BY e.conversion_value DESC

Retargeting Audience (Customer Match)

WITH pricing_views AS (
  SELECT user_id, COUNT(*) as view_count
  FROM events
  WHERE event_name = 'pricing_view'
  GROUP BY user_id
  HAVING COUNT(*) >= 2
),
checkout_starters AS (
  SELECT DISTINCT user_id FROM events WHERE event_name = 'checkout_start'
),
converters AS (
  SELECT DISTINCT user_id FROM events WHERE event_name = 'conversion'
)
SELECT u.email AS Email, u.phone AS Phone
FROM users u
JOIN pricing_views pv ON u.user_id = pv.user_id
JOIN checkout_starters cs ON u.user_id = cs.user_id
LEFT JOIN converters c ON u.user_id = c.user_id
WHERE c.user_id IS NULL

File Locations

BigQuery Location

Project: agents-webinar-2025 Dataset: webinar_demos Tables: users, events, daily_metrics, trial_users, feature_usage