Auditing Google Ads Enhanced Conversions for Leads
A field guide to auditing Google Ads Enhanced Conversions for Leads, diagnosing match rate issues, and benchmarking what agencies charge to implement it.

Enhanced Conversions for Leads recovers an incremental 8 to 12 percent of otherwise unattributed leads, below Google's marketed 10 to 15 percent ceiling and only when match rate holds above 40 percent.
That match rate threshold decides whether ECL earns its implementation cost on a lead gen account or becomes another tracking surface charging rent.
Auditing ECL means checking three things that break in predictable ways: a defensible match rate, a clean diagnostics report, and conversion values Smart Bidding can trust. Fixing those breaks costs more than the lift ECL alone delivers, so the economics have to line up before the implementation is worth signing.
The September 2025 diagnostics overhaul made these failure modes easier to spot, and harder to hide from anyone willing to read the report.
The 40 Percent Match Rate Floor
ECL matches SHA-256 hashed email and phone against signed-in Google sessions. Three bands matter for bidding. Below 40 percent, Smart Bidding treats the enhanced signal as too thin to incorporate. Between 40 and 60 percent, the enhanced target carries partial bid weight. Above 60 percent, it carries full weight into the auction.
Match rate itself is a function of three variables: how much first-party data the form captures, how it is normalized before hashing (Gmail domains require period and plus-suffix stripping; other domains do not), and the share of users signed into a Google product when they convert. B2B accounts with corporate domain emails sit lower than consumer accounts because fewer recipients have Google accounts attached to their work address. EU audiences sit lower because fewer users clear the consent gate before data collection begins.
This matters for the audit sequence. A match rate below 40 percent is not an ECL problem — it is a data capture problem. Adding ECL on top of a form that does not require email, or a CRM pipeline that exports only hashed first names, means the implementation runs without a signal Smart Bidding can act on. Fix the capture layer first; turn the feature on second. Compare against the account's own 30 day trend, not an industry average that does not resemble the traffic.
What the New Diagnostics Report Shows
Google updated the leads diagnostics report in September 2025 to show alerts at every severity level simultaneously instead of hiding them behind the first critical issue. The update adds a high-priority alert for imports carrying user-provided data without a GCLID, validates GCLID-plus-UPD setups that skip the Google tag with a Good status, and retires the CLICK_NOT_FOUND informational error. It prioritizes import-related issues as more critical and downgrades tag-related issues to secondary priority.
The practical read: the report now surfaces where attribution breaks before it surfaces where tags misfire. If an account shows Urgent, work backward from tag firing and conversion action configuration. If it shows Needs Attention with the "Importing limited user-provided data" flag, the CRM is likely sending leads with hashed email but no click ID. That pattern shows up in HubSpot offline conversion integrations built before Google required user-provided data for Smart Bidding on lead gen.
Reading the report once a month is not sufficient on an account spending above $50,000 monthly. Configuration changes take roughly three weeks to fully propagate through the diagnostic status. Agencies managing lead gen at that spend tier should be reviewing weekly during optimization windows, not quarterly.

Failure Modes That Cost You Signal
Three audit findings show up most reliably on lead gen accounts running ECL.
- Value-scaling mismatches. A CRM field exporting "$100,000" as a string versus 100000 as a number can cause Smart Bidding to treat an average lead as a six-figure opportunity, over-bidding on anything that resembles the training data.
- Double counting from parallel tracking. Lead form submissions that fire both a native Google Ads tag and a GA4-imported conversion count twice. Properly handled, the CRM import is the primary conversion action and all client-side events are observational secondary actions.
- Parameter stripping. Cross-domain redirects on booking widgets, payment pages, and embedded forms drop click IDs, leaving the system to match on hashed PII alone. The enhanced signal then carries more weight than the diagnostics assume, and match rate reads artificially low because the GCLID denominator shrinks.
These failures interact. One financial services case study reported conversion volume dropped by roughly half once duplicate and low-quality conversion actions were removed, with qualified lead CPA improving as Smart Bidding retrained on signal that reflected reality. The sequencing lesson for an enhanced conversions implementation audit is unambiguous: fix counting rules before pushing match rate higher. Chasing match rate on an account with double counting optimizes against the wrong number faster. See the related CRM-to-Google-Ads discrepancy patterns for the variance range on counting fixes.
What Agencies Charge to Implement and Maintain ECL
Implementation pricing breaks into tiers by data path.
- Google tag with automatic collection. $1,500 to $4,000 one-time if forms are standard and the Google tag already fires correctly across the funnel.
- Google tag with manual user-provided data mapping. $3,000 to $8,000, depending on form count and whether consent handling needs a rebuild to pass EU audiences.
- Offline upload via API or CRM connector. $6,000 to $15,000 to build, plus an ongoing data pipeline retainer of $500 to $2,000 monthly for the mapping, hash normalization, and error-queue monitoring.
- Server-side GTM routing. Add $800 to $2,500 monthly on top of implementation for infrastructure and tagging oversight. The server-side GTM cost breakdown is worth reading if the quote came in above that range.
The maintenance line is where agencies often bury margin without naming it. If the contract is percent-of-spend and does not specify ECL maintenance separately, confirm who is responsible for weekly diagnostic review and monthly match rate reporting. A retainer of 12 to 15 percent of ad spend without a named tagging engineer on the account typically means ECL is maintained reactively rather than proactively.
Match-rate drift is the clearest indicator of neglect. A 10 point drop over a quarter, with no form or consent changes in the release notes, suggests the diagnostics report has not been opened at all.
What a Healthy ECL Deployment Actually Reads Like
Three numbers separate a clean deployment from one needing rework: match rate trend over 30 days, count of Urgent or Needs Attention alerts open past 72 hours, and the delta between Google Ads reported conversions and CRM-logged leads for the same window. Those are the figures an agency should produce before scoping work.
The CRM delta is where accounts leak money in either direction, overstating lift and overbidding, or understating lift and starving winning campaigns of signal. The expected spread is wide: implementation quality separates a top-decile account from a bottom-decile one by more than 40 percentage points of measured lift.
The question worth putting to an agency is how many lead gen accounts on their roster they have audited at that depth in the past 90 days. A good answer names account spend tiers, the match rate distribution across them, and at least one account where they walked away from implementing ECL because the data capture layer was not ready. Anything less specific suggests ECL is being sold as a feature, not operated as an ongoing signal-quality discipline.