Modeled conversions
Modeled conversions are conversions a platform estimates statistically rather than observes directly. When direct measurement is blocked — by missing consent, cross-device journeys, or privacy protections — ad and analytics platforms model the likely conversions from observable trends and aggregated data, and report them alongside observed ones. Understanding which conversions are modeled is essential to reading attribution honestly.
What this means
A modeled conversion is one the platform did not directly tie to an ad interaction but infers happened. Modeling kicks in where direct observation fails: a user who did not consent to tracking, a journey that crossed devices, or a browser that blocked the identifier. Rather than report zero, the platform estimates how many conversions likely occurred based on patterns from comparable, observable traffic.
Google Ads, for example, documents using conversion modeling to estimate conversions it cannot observe directly, and reports them within the same conversion columns.
How to read them
Modeled conversions are most defensible in aggregate and over reasonable volumes — they are estimates of totals, not assertions about specific users. Problems arise when teams treat modeled numbers as exact, segment them too finely, or compare a modeling platform's totals against a tool that only counts observed events without acknowledging the difference.
The honest posture is to know the share of modeling involved, avoid over-interpreting small or sliced figures, and reconcile against an independent observed baseline where possible.
- Estimated, not directly observed
- Used where consent/cross-device/privacy block measurement
- Most reliable in aggregate, not at fine granularity
How it appears in analytics and logs
A conversion count that includes modeling means part of the total is estimated; the estimate can be reasonable in aggregate but should not be treated as exact per-user truth.
Diagnostic use case
Distinguish observed from modeled conversions when reading reports, so you know which figures are measured and which are statistical estimates filling measurement gaps.
What WebmasterID can help detect
WebmasterID reports first-party, observed events; it does not model conversions, so you can use its directly-measured counts as a grounded baseline against platforms that include modeled figures.
Common mistakes
- Treating modeled conversions as exact per-user facts.
- Slicing modeled totals too finely to stay reliable.
- Comparing modeled totals against observed-only counts blindly.
Privacy and accuracy notes
Modeling exists largely to report performance without identifying individuals — it uses aggregated, consented signals to estimate totals rather than tracking specific people across the gap.
Related pages
- Consent and attribution
Consent is upstream of attribution: under frameworks like the EU's GDPR and ePrivacy Directive, storing or reading identifiers for tracking generally requires the user's consent. When consent is declined or withheld, the touchpoints those identifiers would have recorded never enter the data, so attribution operates on partial paths. Understanding consent is therefore inseparable from reading attribution honestly.
- SKAdNetwork attribution
SKAdNetwork (SKAN) is Apple's framework for attributing app installs and post-install conversions to ad campaigns without identifying the user or device. Instead of a deterministic identifier, it sends the ad network an aggregated, delayed 'postback' confirming a conversion happened, with deliberately limited campaign granularity and a conversion value of restricted resolution. It is the privacy-preserving backbone of iOS install attribution after ATT.
- Enhanced conversions
Enhanced conversions is a Google Ads feature that supplements cookie-based conversion measurement by sending hashed first-party customer data — such as an email address the user provided — to match conversions that cookies alone would miss. The data is hashed (SHA-256) before transmission. It is one industry response to the decline of third-party identifiers, with its own consent and configuration requirements.
- Privacy-first analytics
Observed first-party events as a grounded baseline.
Sources and verification notes
- Google Ads Help — About conversion modelingDocuments how conversions are modeled when they cannot be observed directly.
Last reviewed 2026-06-24. Facts are checked against primary/official sources where available; uncertain specifics are marked “Data not yet verified” rather than guessed.