Google Ads conversion tracking: native tags vs GA4, enhanced conversions, and consent mode

Google Ads only optimizes as well as the conversion data you feed it. This guide walks through the measurement architecture choices for native tags versus GA4 imports, enhanced conversions, and consent mode v2 so you can wire bidding data correctly before spend starts, then extend it later with CRM feedback via offline conversion tracking if you are a lead business.

Choosing between native Google Ads tags and GA4 key-event imports

Native Google Ads conversion tags and GA4 key-event imports are not interchangeable. Per Google’s Analytics help, view-through conversions require native Google Ads conversion measurement, and you lose them when you import conversions from Analytics. If you care about display, YouTube, or discovery assist value, this alone pushes you toward native tags for your primary bidding signals.

When you import GA4 key events as conversions, Google’s documentation states that those imported actions are set to secondary automatically if the Google Ads account already has its own goals. This prevents double counting for Smart Bidding, because primary conversions feed the Conversions column and bidding, while secondary conversions are observation-only. A common assumption is that running both native and GA4 actions will double-count conversions; the platform avoids this by defaulting GA4-based actions to secondary when native goals exist.

GA4-imported web conversions also follow GA4’s channel credit rules. Google’s help explains that when you create Google Ads conversions based on GA4 key events, you choose whether credit is attributed to Google Paid Channels (the default for new links since June 2023) or to Paid and Organic Channels. That choice also determines where you configure conversion windows: if you use Google Paid Channels, you manage windows in Google Ads, and if you use Paid and Organic Channels, you manage them in Analytics.

Use native Google Ads tags when you need view-through conversions, maximum control of attribution and windows inside Google Ads, and minimal dependency on Analytics configuration. Use GA4 imports when you want a single Analytics-centric view of performance and are comfortable with GA4’s channel credit logic and the imported actions being secondary by default if native goals already exist.

How deduplication behaves with native and GA4-based actions

Deduplication between native and GA4-based conversion actions in Google Ads is handled structurally at creation time. Google’s Analytics help states that when you create Google Ads conversions based on GA4 key events in an account that already has its own goals, those GA4-based actions are created as secondary. Since primary conversion actions are what feed Smart Bidding and the Conversions column, this prevents the same business outcome from being counted twice for bidding.

This means you can safely run both native and GA4-imported actions in parallel for comparison, as long as you keep the native actions as primary and leave the GA4 imports as secondary. If you later promote a GA4-imported action to primary, you should demote or remove the overlapping native action to avoid intentional double counting. Window changes are not retroactive according to Google Ads help, so if you adjust conversion windows while you are comparing actions, expect only future conversions to follow the new windows.

Enhanced conversions for web and leads, and how hashing works

Enhanced conversions are an add-on to existing conversion tracking that send hashed first-party identifiers to improve matching. Google Ads help explains that enhanced conversions supplement existing conversions by sending first-party data that has been normalized and then hashed with SHA-256, which is a one-way hashing algorithm. A common fear is that this sends your customer list in plain text; the documentation is explicit that normalization and hashing happen before anything is sent.

Normalization rules before hashing are specific. Per Google’s documentation, whitespace is trimmed, all characters are lowercased, phone numbers are formatted in E.164 format, and periods are removed before the domain in gmail.com and googlemail.com addresses. The resulting value is then hashed using SHA-256, and the output is a hexadecimal SHA-256 string. Google uses these hashed identifiers to match against signed-in Google accounts that interacted with your ads.

There are 2 variants of enhanced conversions. Enhanced conversions for web apply to sales and events that occur directly on the website. Enhanced conversions for leads apply to offline sales that originate from website leads, where you later send hashed first-party data that can be matched back to signed-in Google accounts that engaged with your ads. For lead businesses, this sits on top of the base tracking described here and is extended further when you send CRM outcomes back through offline conversion tracking.

Google documents 3 setup paths for enhanced conversions: using the Google tag, using Google Tag Manager, or using the Google Ads API. In all cases, the same normalization and SHA-256 hashing rules apply before data is transmitted.

Consent mode v2 parameters, modes, and EEA enforcement

Consent mode v2 is Google’s framework for adapting tag behavior based on user consent choices, and it adds 2 advertising-specific parameters. Tag platform documentation lists 4 consent parameters: ad_storage, ad_user_data, ad_personalization, and analytics_storage. Consent mode v2 is defined by the addition of ad_user_data (consent to send user data to Google for advertising) and ad_personalization (consent for personalized advertising); if these parameters conflict, Google’s documentation states that personalization is disabled.

Basic and advanced consent mode differ in what gets sent when consent is denied. In basic consent mode, Google tags are blocked until consent is granted, and when a user denies consent, nothing is sent, not even the consent status. Tag platform docs explain that modeling in this case uses a general model. In advanced consent mode, tags load immediately and send cookieless pings even when users deny consent, which enables advertiser-specific modeling.

Consent mode v2 also controls enhanced conversions. Tag platform documentation states that when ad_user_data is denied, personal data collection for advertising is disabled, which includes the hashed first-party data used by enhanced conversions. For EEA users, Google Ads help explains that to keep using measurement, personalization, and remarketing, you must collect consent and share consent signals with Google. Since March 2024, Customer Match lists in the EEA require both ad_user_data and ad_personalization consent fields to be granted.

Current attribution models and what was removed

Google Ads attribution models changed significantly in 2023. Google Ads help states that data-driven attribution is now the default model for most conversion actions. The same documentation explains that first click, linear, time decay, and position-based attribution models were removed starting in June 2023, and remaining conversion actions using those models were switched to data-driven from September 2023.

Last click attribution remains supported, as does external attribution. This means your practical choices inside Google Ads are data-driven attribution, last click, or an external attribution model. Any belief that you can still choose time-decay or position-based attribution inside Google Ads is out of date, because those models were removed in 2023 according to Google’s published timeline.

Configuring conversion windows and primary versus secondary actions

Conversion windows define how long after an interaction a conversion can be credited to that interaction. Google Ads help documents that click-through conversion windows are configurable from 1 to 90 days depending on the source, with a default of 30 days and a recommendation of at least 7 days. Engaged-view conversion windows can be set from 1 to 30 days, with a default of 3 days, and view-through conversion windows can be set from 1 to 30 days, with a default of 1 day.

Google’s documentation also states that conversion-window changes are not retroactive. When you adjust a window, only conversions that occur after the change respect the new setting; historical data remains attributed using the old window. For GA4-imported conversions, Google’s help explains that if you choose Google Paid Channels when creating the import, you manage windows in Google Ads, and if you choose Paid and Organic Channels, you manage them in Analytics.

Primary and secondary conversion actions control what Smart Bidding optimizes toward. Google Ads help defines primary conversion actions as those that feed the Conversions column and bidding, while secondary actions are observation-only and do not influence Smart Bidding. Smart Bidding strategies such as Target CPA, Target ROAS, Maximize conversions, and Maximize conversion value optimize for conversions or conversion value using auction-time bidding, and they optimize toward the conversions your account measures as primary.

In practice, you should configure your main business outcomes as primary conversion actions and keep diagnostic or supporting metrics as secondary. When you import GA4 key events into an account that already has native goals, Google’s documentation notes that those imported actions are created as secondary by default, which avoids accidental double counting for bidding.

Measurement before spend and HIPAA/PHI-constrained architectures

Temper And Forge’s documented practice is that conversion measurement is the precondition for everything the firm does on Google. ROAS is the keep-or-kill metric, and ROAS is only as real as the tracking underneath it. TF requires the measurement stack configured before spend starts, because an untracked test can only measure clicks.

The architecture choice follows the business: TF uses native tags where view-through and bidding fidelity matter, then adds a CRM-feedback layer for lead businesses via offline conversion uploads. TF also works with HIPAA-compliant and PHI-related clients where traditional conversion tracking is off the table entirely. For those advertisers, measurement is built around compliant alternatives and honest proxies, the quality-signal consequences are priced in (this is one reason TF ignores Quality Score), and the ROAS read is adjusted rather than faked.

FAQ

The following FAQs cover adjacent mechanics and edge cases that often come up when implementing Google Ads conversion tracking.