Marketing Analytics & Attribution

GA4 event taxonomy for fintech is not the same as ecommerce. A loan application is not a purchase. The funnel has stages: lead capture, qualification check, credit decision, funding, post-funding monitoring. Each stage has different business value and different measurement requirements.

We structure GA4 events around your actual business milestones, not around what Google's default ecommerce schema suggests. For a lending client, this means separate events for "application_submitted", "credit_decision_received", "funds_disbursed", and "loan_funded_status". Each event carries conversion_value reflecting the stage, contract probability, or expected customer lifetime value. The CRM is your source of truth for whether that application moved to funded status, and the offline conversion tracking pipeline ensures that real outcome flows back into Google Ads.

Consent handling in fintech measurement requires precision. GDPR and CCPA place restrictions on how data flows from your CRM into advertising platforms, but offline conversion uploads under Google Ads policies and the Conversions API are designed to comply when implemented correctly. We configure GA4 consent mode so that data collection behavior adapts based on user consent status: if a user has not consented to advertising, their conversion events still fire locally for analytics, but do not flow to ad platforms. This preserves your ability to measure performance while respecting consent signals and regulatory guardrails.

Standard Google Ads conversion tracking optimizes for "conversions" (form submissions, clicks, page views). Value-based bidding optimizes for the revenue that each conversion actually generates. The 536% YoY growth case started because the lending client's CRM held the truth about which applications turned into funded loans and at what APR. That value signal was not reaching Google Ads.

ConversionUploadService is the Google Ads API endpoint that accepts offline conversion data: user click ID or customer ID, conversion time, conversion value, and conversion action. Once that data is uploaded, Google's algorithm learns which ad groups, keywords, and audiences drove high-value conversions, not just any conversion. The lag between click and funding is typically 3-7 days in lending; ConversionUploadService accepts uploads for up to 90 days post-click, ensuring you capture the full loan funding window.

Implementation requires three data points: first, a reliable join between your CRM records and Google Ads click IDs (stored in gclid or hashed customer email matched to your first-party CRM ID). Second, clean event schema in your CRM: conversion time must be when the true business outcome occurs (loan funded date, not application date). Third, conversion value must represent actual customer value: either the loan amount, the expected lifetime commission, or a blended score if you have multiple revenue streams. We build the ETL pipeline that extracts this data daily from your CRM, validates it against duplicate prevention rules, and uploads to Google Ads. Upload lag is typically within 24 hours of funding, giving Google enough real-time signal to adjust bidding within the same campaign day.

Regulated financial services operate under GDPR (EU), CCPA (California), or local equivalents (Chile, Mexico, Brazil). These laws govern how personal data flows from your website to ad platforms. Naive attribution setups that pipe every user interaction to Google Analytics and then to Google Ads often create compliance friction.

GA4 consent mode enables you to measure performance while respecting user consent choices. When a user rejects advertising consent, GA4 still fires analytics events but truncates identifiers and does not send data to Google Ads. When a user consents, the full event stream flows to Google Ads and the Conversions API. This is not approximate measurement; it is precise, segmented measurement: you see campaign performance for consented users separately from non-consented users, and your bidding algorithm only trains on consented data.

In fintech specifically, offline conversion tracking has a compliance advantage. When a loan is funded, that event triggers in your CRM independent of the user's original advertising consent status. The offline upload to Google Ads then happens under first-party, CRM-to-CRM server API contracts, not through the browser. This separation means your measurement does not depend on third-party cookies or pixel-based tracking; it is based on authenticated CRM records. Regulators view this architecture more favorably because the data flow is transparent, controllable, and auditable. We implement consent mode configuration at initial GA4 setup and document the consent enforcement logic so your privacy team can verify compliance.

Offline conversion tracking is only as reliable as your CRM data. Common issues we encounter: applications with null funding dates, duplicate funding records (because a loan was re-funded or adjusted), revenue tied to wrong customer IDs (if your CRM does not normalize person-to-account relationships), or conversion values that change after upload (when an interest rate is adjusted and you now want to re-upload with new value).

We run a conversion audit before live uploads begin. The audit queries your CRM to identify data quality patterns: what percentage of applications have a funded_date within 30 days, which fields are reliably populated, whether your database normalizes person and company records consistently, and what latency exists between funding and when it appears in the database. For a lending client in Mexico, this audit revealed that 12% of funded loans lacked a matched advertiser click ID because the customer applied on a secondary domain. We implemented cookie tracking on that domain to capture the click ID, then backfilled historical CRM records with the click ID via a one-time ETL job.

Conversion window timing matters. Google Ads expects you to upload a conversion within 90 days of the click. Lending funnels can stretch to 21 days or longer (application, underwriting, disbursement). You must define in advance: is your conversion event "loan approved" or "loan funded"? If you choose "approved", bidding optimizes for approval rate; if you choose "funded", bidding optimizes for actual revenue-generating customers. Most fintech clients choose "funded" because approval without funding is not revenue. This means your upload window must span from click to funding, often 15-45 days. We configure ConversionUploadService with daily or near-real-time uploads and track upload status in a monitoring dashboard so you know if a batch failed or uploaded late.