Google Ads is transitioning Demand Gen Lookalike segments toward suggestion mode through a phased 2026 rollout. In suggestion mode, the seed and reach choice guide optimization but may not act as a strict similarity boundary. Advertisers who need the earlier targeting-constraint behavior may have an opt-out path during the transition. Account availability and labels can change, so the lesson begins with documenting the live mode.
VISUAL LESSON
What you will learn
- 01Explain suggestion mode versus targeting-constraint behavior.
- 02Prepare a high-intent, governed seed list.
- 03Create a before-and-after reporting plan for the phased update.
ILLUSTRATIVE MODE COMPARISON
The same reach label may represent a different control model
GOOGLE ADS AUDIENCE MAP
Document the active behavior
Record seed source, membership health, reach choice, status tag, exclusions, and whether the interface describes a signal or targeting constraint.
Use a distinct ad group for the Lookalike strategy, aligned creative, stable location and language settings, and explicit customer exclusions.
Segment the transition period and compare reach, conversions, conversion value or qualified outcome, while noting every mode change.
THE LOOKALIKE TRANSITION
Seed → similarity signal → optimized reach
THE LEAD ATLAS METHOD
For a separate business-to-business campaign, Lead Atlas Data can research contacts matched to the exact market, locations, and business categories that define the prospecting brief rather than relying on a consumer-audience signal.See how custom list research works ↗Confirm the live mode
Open the Demand Gen audience report and inspect the Lookalike status, tag, reach control, and account notices. Do not assume an older narrow, balanced, or broad selection still behaves as a strict boundary.
Save the date, campaign and ad group, segment ID or name, seed source, setting, and screenshot or export. The rollout itself is a measurement change.
Build a high-intent seed
Use consented first-party data that represents the action you want more of: recent purchasers, qualified converters, valuable subscribers, or another clearly governed group. Avoid mixing prospects, customers, employees, test records, and poor-fit conversions.
Check list freshness, upload health, policy eligibility, membership scale, exclusions, and ownership. A Lookalike cannot correct a seed that represents the wrong outcome.
Understand suggestion behavior
Google says the transition lets the seed and reach level guide its models while permitting reach outside earlier similarity thresholds to optimize for campaign goals. That means the audience label should not be interpreted as a guarantee that every reached person matches a fixed profile.
If the account uses the targeting-constraint path, document that choice and its limits. Do not describe either mode as universally better; the right choice depends on control needs, data quality, and measured outcomes.
Design a readable campaign test
Keep the Lookalike strategy in a distinct ad group with its own creative rationale and reporting. Hold location, language, conversion goal, bid strategy, and major exclusions stable while comparing the mode or seed.
Use new-customer and existing-customer definitions carefully. Monitor reach, spend, conversions, value, customer status, and downstream quality rather than only click-through rate.
Govern the transition
Create a change log for rollout notices, mode tags, opt-out decisions, seed refreshes, membership issues, and reporting shifts. Rebaseline after a material behavior change instead of comparing the full period as if targeting were constant.
Deliverable: seed specification, governance check, live-mode evidence, ad-group map, exclusions, transition date, outcome table, risk note, and the next review decision.
THE TAKEAWAY
Use a clean high-intent seed, record whether Lookalike is a suggestion or constraint, isolate it in reporting, and judge downstream conversion quality rather than assuming similarity guarantees fit.OFFICIAL REFERENCES