Skip to content
whitead

Data-Driven Attribution (DDA)

glossary google ads updated 2026.04.29 5 min read
// contents · 9 sections
  1. What Data-Driven Attribution is
  2. DDA vs Last Click
  3. Why other models were sunset
  4. Common misconceptions
  5. "Does DDA always outperform Last Click?"
  6. "Why did my conversions shift after enabling DDA?"
  7. "Does DDA work cross-channel like GA4 attribution?"
  8. Related concepts
  9. Sources

Data-Driven Attribution (DDA) is Google Ads' ML model splitting conversion credit across ad touchpoints — since 2023-09, the only non-Last Click option.

What Data-Driven Attribution is

DDA is an algorithmic model that compares paths that converted to paths that did not, then assigns each ad interaction a fractional share of conversion credit based on its incremental contribution. Inputs include click and impression touchpoints, time between interactions, device, ad format, and query type. Google Ads Help defines it directly:

"Data-driven attribution distributes credit for the conversion based on your data for each conversion action. It's different from the other models in that it uses your account's data to calculate the actual contribution of each interaction across the conversion path."
— Google Ads Help: About data-driven attribution, retrieved 2026-04-29

Mechanically, DDA is a counterfactual model: it asks "would this conversion have happened without this touchpoint?" and weights credit by the lift each touchpoint produces.

DDA vs Last Click

Both models are still available per conversion action. Choice matters because Smart Bidding (tCPA, tROAS, Maximize Conversions) optimises against the credit distribution your conversion action exposes — not raw conversion counts.

Model Status Credit distribution
Data-Driven Attribution Default since 2021-09 ML, fractional across touchpoints
Last Click Available 100% to last clicked ad
First Click / Linear / Time Decay / Position-Based Sunset 2023-09 Auto-migrated to DDA

Conversion-volume note: DDA no longer enforces a minimum data threshold (previously 3,000 ad interactions and 300 conversions over 30 days). Below roughly 300 conversions per 30 days, however, the model has lower confidence and shifts can look noisy — not wrong, just less stable.

Why other models were sunset

Rule-based attribution was retired because Google's usage data showed it was vestigial. Per Google's March 2023 announcement, First Click, Linear, Time Decay, and Position-Based combined accounted for less than 3% of Google web conversions. From June 2023 these models could no longer be selected for new conversion actions; from September 2023 any action still using them was auto-converted to DDA in both Ads and GA4. The product rationale was simplification: a single ML default that scales to all account sizes, plus Last Click as the parity escape hatch (Google Blog, 2021-09-27). By 2024 the migration was complete — every new conversion action defaults to DDA.

What we see in Whitead audits: the most common DDA finding is mixed attribution across conversion actions. A primary purchase action created in 2019 still reports on Last Click while newer lead and add-to-cart actions sit on DDA. Smart Bidding optimises per conversion action, so the bid model receives one credit signal for purchases and a different one for everything else — inconsistent learning. Verify the model per action under Tools → Conversions → "Attribution model" column, not at the account header. Imported GA4 conversions inherit the GA4 property's attribution setting, not the Ads default — a separate trap.

Common misconceptions

"Does DDA always outperform Last Click?"

No — model superiority is conversion-volume dependent. Above ~300 conversions per 30 days DDA has enough signal to model touchpoint lift reliably and typically lifts conversions at the same CPA when paired with Smart Bidding. Below that, DDA still works (no threshold blocks it) but credit shifts can be noisy. Mavlers' 2025 attribution guide frames this as a confidence question, not an availability question.

"Why did my conversions shift after enabling DDA?"

Last Click concentrates credit on bottom-funnel campaigns (branded search, retargeting); DDA spreads credit upstream to generic search, video, and discovery touchpoints. Branded campaigns commonly drop credited conversions while upper-funnel ones gain. Counts also fragment because DDA assigns fractional credit (0.4 + 0.6) versus Last Click's binary 1.0. Total conversions do not change — distribution does.

"Does DDA work cross-channel like GA4 attribution?"

No. Google Ads DDA only models touchpoints inside Google Ads. Cross-channel attribution (organic, email, social, direct) lives in GA4's attribution reports with a separate DDA model. Same name, same methodology, different scope. For cross-channel decisions use GA4 attribution; for in-Ads bid optimisation use Ads DDA.

Sources

← Back to Wiki