Enhanced conversions is a Google Ads measurement feature that sends hashed first-party data — email address, phone number, name and address — along with your conversion tags. When cookie restrictions break the link between an ad click and a later conversion, Google matches that hashed data against signed-in Google accounts and restores the connection. The hashing is SHA-256 and happens before the data leaves your page: Google receives a fingerprint, not the address book.
The stakes are simple. Browsers and privacy features have been severing click-to-conversion links for years, which means unenhanced accounts undercount — and Smart Bidding treats undercounted as underperforming. Bidding on systematically missing data is quietly expensive.
How enhanced conversions work
At conversion time, your tag captures user-provided data from the page — the email typed into the checkout or lead form — hashes it, and sends it with the conversion event. Google matches the hash against its signed-in users; where a match connects the conversion to an ad interaction the cookie could not, the conversion is attributed and reported inside your existing conversion actions. No new columns, no separate line — just counts that drift closer to reality, gradually, as matches accumulate.
The feature comes in two flavours:
| Enhanced conversions for web | Enhanced conversions for leads | |
|---|---|---|
| Solves | On-site conversions the cookie lost | Offline deals that close days or weeks later |
| Captures | User data on the conversion page | Hashed email or phone at form submission |
| Then | Matches and attributes directly | You import closed deals keyed by that hash instead of a click ID |
| Fits | Ecommerce, sign-ups, bookings | Sales teams, CRMs, lead generation |
Setup runs through one of three paths: the Google tag detecting fields automatically, Google Tag Manager with explicit variables, or the API for server-side setups. After launch, the conversion action's diagnostics surface whether enhanced data is actually arriving and matching — the difference between configured and working.
What an agent owns here: the before and after
Enhanced conversions is a case where the honest AI answer is: the fix is mostly on your website, not in the account. Tags and consent flows live outside what an ads connector writes to. What an agent does own is the before and after.
Before: diagnosis. An agent reading your account can surface the symptoms of undercounting — conversion counts that trail your CRM, lead campaigns whose reported cost per lead disagrees with what sales receives, patterns consistent with lost attribution rather than lost performance. That analysis is what turns "should we set this up?" from a hunch into a decision.
After: verification. Once enhanced conversions launch, the useful questions are empirical — are diagnostics showing data arriving, has the gap against the CRM narrowed, did the trend change where the setup went live. An agent can run that comparison on a schedule and report the drift, which is how you find out whether the feature is recovering anything or merely installed.
The framing to keep: this is measurement infrastructure. It makes the data contract truer; what bidding does with truer data is the next conversation. Start with conversion actions — the contract itself — and see where measurement checks sit in an AI-assisted audit.