Data-Driven Attribution (DDA) became the default for new Google Ads conversion actions in September 2023, but a long tail of legacy accounts still runs on last-click — quietly starving Smart Bidding of upper-funnel credit and pushing budget toward the wrong slices.
What "migrating to DDA" actually means
Last-click attributes 100% of conversion credit to the final ad-interaction click before the conversion. DDA uses your account's own conversion paths to assign fractional credit across the touchpoints that meaningfully shifted the probability of converting. Smart Bidding (tCPA, tROAS, Maximize Conversions/Value) optimizes against the credit signal it receives — so an attribution model change is a bidding signal change, not a reporting change.
Google's claim, in their own words:
"Data-driven attribution distributes credit for the conversion based on how people engaged with your various ads and decided to become your customers. Data-driven attribution uses your account's data to calculate the actual contribution of each ad interaction across the conversion path." — Google Ads Help — About DDA
Observed in the wild: Search Engine Land's 2024 analysis of mid-market accounts found credit shifts ranging from "barely measurable" on brand-heavy accounts to 20%+ reallocation toward upper-funnel Display/YouTube on multi-touch B2B journeys (Search Engine Land, 2024-06-18). PPC Land's 2025 follow-up notes that DDA's credit redistribution "rewards assist channels but does not magically lift total conversions" — same conversions, redistributed credit (PPC Land, 2025-03-04).
Top causes of a stuck last-click setup (ranked by audit frequency)
- Account predates September 2023 — conversion actions were created when last-click was default; no one has touched the attribution dropdown since.
- Ineligibility flag never re-checked — account once had too few conversions for DDA, was set to last-click as fallback, and is now eligible but unrevisited.
- Imported GA4 conversions on legacy model — GA4 attribution model and Google Ads attribution model are separate settings; migrating GA4 to DDA does not propagate.
- Multiple conversion actions on mixed models — one action on DDA, another on last-click, double-counted into the same Smart Bidding strategy.
- Fear of "the numbers will change" — non-issue if you communicate the 14-day model comparison baseline before flipping.
Diagnostic checklist
| # | Check | Where in UI | Threshold |
|---|---|---|---|
| 1 | Conversion action attribution model | Goals → Conversions → Summary → action → Settings | "Last click" = migration candidate |
| 2 | Eligible conversions over 30 days | Goals → action → Performance | Operator rule of thumb in the 500-conv/30d range (Optmyzr/SEL); Google has not published a hard floor since 2021. Below that, stay on Time Decay |
| 3 | Mixed-model conversion actions feeding same campaign | Goals → Conversions table sort by Model | Any mismatch = unify first |
| 4 | Model comparison report available | Insights & reports → Attribution → Model comparison | If empty: not enough data, skip |
| 5 | Smart Bidding strategy on tCPA/tROAS | Campaigns → Bid strategies | Yes = freeze targets ±0% for 14d post-switch |
| 6 | Cross-device conversions enabled | Conversion action → Include in Conversions | Required for DDA fidelity |
Fix paths
Run sequentially — skipping the baseline is the most common reason migrations get rolled back on day 3 in a panic.
- 14-day baseline capture. Pull the model comparison report for last-click vs DDA over the trailing 14 days. Export per-campaign and per-channel credit deltas. This is your "before" file — without it, day-7 leadership questions cannot be answered.
- Unify conversion actions. If two conversion actions feed the same campaigns on different models, switch the secondary first or exclude it from "Conversions" (keeping it in "All conversions"). Avoid double-flipping during one migration.
- Switch the primary action to DDA. Goals → action → Edit settings → Attribution model → "Data-driven". If grayed out: not eligible — fall back to Time Decay.
- Freeze Smart Bidding targets for 14 days. Do not retune tCPA/tROAS during this window. The bidder is relearning the credit distribution; target changes confound the signal. See Smart Bidding learning phase.
- Day-14 review. Re-pull model comparison. Expected pattern: Search brand/exact slightly down on credit, Search generic flat-to-up, Display/YouTube/Video assists materially up. Total conversions should be approximately flat ±5%; if total conv shifts >10%, you have a tracking issue, not an attribution issue.
- Day-30 retune. Once credit distribution stabilizes, retune tCPA/tROAS using the new conversion volume — typically tCPA can come down 5-10% on assist-heavy campaigns because Smart Bidding now sees credit it was blind to.
Methodology note. Whitead's attribution_model rule flags any conversion action set to "Last click" when the account is eligible for DDA (model comparison report populated; operator rule of thumb in the 500-conv/30d range per Optmyzr/SEL — Google has not published a hard eligibility floor since 2021, so DDA eligibility is computed opaquely from the model comparison report). The rule is MED severity rather than HIGH because the symptom is misallocation, not measurement loss — total conversions usually do not change, only the channel mix Smart Bidding optimizes for. Fix priority is: eligibility check → mixed-model unification → primary action switch → 14-day target freeze. We deliberately do not auto-recommend DDA on accounts running heavy brand-search-only — credit redistribution is minimal and the migration cost rarely pays back.
When to escalate
Escalate when:
- Conversion action is below the operator rule-of-thumb floor (~500 conv / 30d per Optmyzr/SEL; Google has not published a hard floor since 2021) and DDA shows ineligible — switch to Time Decay instead and revisit eligibility quarterly. Forcing DDA on thin data produces noisy credit and unstable bids.
- Total conversions shift >10% in the 14-day post-switch window — that is a tracking, consent, or Enhanced Conversions issue surfacing under the new model, not attribution. Roll back, fix tracking, retry.
- Account is >80% brand search by spend — DDA's credit redistribution has little to optimize against. Stay on last-click or Time Decay; revisit when generic/upper-funnel spend exceeds 30%.
Will my historical reports change after switching?
No. The attribution model applies to conversions counted from the switch date forward. Historical conversion data stays on the model under which it was recorded — model comparison report shows the parallel reconstruction.
What about GA4 — do I need to migrate there too?
GA4 has its own attribution setting under Admin → Attribution Settings. They are independent. Most teams align both on DDA; if you import GA4 conversions into Google Ads, the Google Ads attribution model wins for bidding purposes.
How does DDA interact with Consent Mode v2?
DDA uses modeled conversions to fill consent-denied gaps. The combination is additive — DDA needs the modeled paths Consent Mode v2 supplies to model assists for non-consenting users. Running last-click + Consent Mode v2 wastes a chunk of the modeling Google is already doing for you.
Related concepts
- Smart Bidding — what consumes the attribution signal
- Smart Bidding learning phase — why the 14-day target freeze matters
- Data-Driven Attribution — glossary entry on the model itself
- Enhanced Conversions — measurement layer DDA depends on
- Consent Mode v2 — modeled conversions feed DDA's path data
- fix-attribution-last-click — rule-tied finding for this migration
Sources
- About Data-driven attribution — Google Ads Help, 2025-09
- Choose an attribution model for your conversions — Google Ads Help, 2025-11
- Update on attribution models in Google Ads — Google Ads blog, 2023-04-06
- Data-driven attribution in Google Ads: real-world impact analysis — Search Engine Land, 2024-06-18
- DDA Google Ads 2025 review — what changed in credit allocation — PPC Land, 2025-03-04
- Attribution and Smart Bidding: how the bidder uses credit signals — Optmyzr, 2025-07-22