Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsLookalike segments and Optimized targeting solve different parts of Demand Gen audience selection. A Lookalike segment starts with your first-party seed—such as purchasers, qualified leads, site visitors, app users, or YouTube-engaged people—and asks Google to find new users who resemble it. Optimized targeting starts with optional audience segments or signals, then allows Google’s systems to search beyond them for users predicted to convert. Google says the two features can be used together.
There is an important 2026 qualification: Demand Gen Lookalikes are being moved to a suggestion mode. In that mode, the seed list and Narrow, Balanced, or Broad setting guide Google’s model but do not necessarily restrict delivery to a fixed similarity band. Accounts that retain the older constraint behavior can use Google’s documented opt-out route, but availability is account-specific.
As an Amazon Associate I earn from qualifying purchases.
Lookalike segments and Optimized targeting in one sentence
Choose a Lookalike segment when you have a meaningful first-party audience whose characteristics should guide prospecting. Choose Optimized targeting when conversion performance is the priority and you want Google to explore beyond your selected audience signals. The first is seed-led discovery; the second is conversion-led expansion.
| Question | Lookalike segments | Optimized targeting |
|---|---|---|
| What starts the system? | A first-party seed, such as a customer list, website or app activity, or YouTube engagement. | Optional manually selected audience segments used as an initial signal. |
| What is the main job? | Find new people with characteristics similar to the seed audience. | Find people Google predicts are most likely to convert, including people outside the selected signals. |
| What does the advertiser control? | The seed and, where available, a reach preference; the extent of control depends on whether the account is in suggestion or constraint mode. | The initial audience signals and campaign objective, while Google’s model decides how far to expand. |
| Best strategic question | “Do we have a relevant, high-intent audience that should shape prospecting?” | “Are we willing to let Google search broadly for conversions?” |
Google’s Demand Gen FAQ describes the distinction this way: “Lookalike segments help you drive similarity based audiences and optimized targeting helps you drive performance by showing ads to people who are most likely to convert.” Google Ads Help presents that as a difference in purpose, not a guarantee that one setup will outperform the other.
#1 Best Overall
How a Demand Gen Lookalike segment works
1. You supply a first-party seed
The seed is the audience Google uses to infer common characteristics. Possible seeds include recent purchasers, high-value customers, qualified leads, people who visited a website, app users, or users who engaged with a YouTube channel or video. The seed should be closely related to the campaign’s actual goal. A list of completed purchasers is generally a more useful prospecting guide for an ecommerce acquisition campaign than a very broad list of all site visitors; a list of qualified leads is more aligned with lead generation than a list containing every casual page view.
2. Google models similarity
Google analyzes patterns in the seed and looks for additional users who appear to share relevant characteristics. Historically, Demand Gen offered three reach choices: Narrow, Balanced, and Broad. Google’s constraint guidance described those settings as aiming at approximately 2.5%, 5%, and 10% of users in the target location, respectively. Those percentages describe the legacy similarity/reach behavior, not a permanent promise about delivery in every account.
3. The 2026 rollout changes the meaning of the slider
Google is transitioning Demand Gen Lookalike segments to suggestion mode through a phased 2026 rollout. The update began becoming available to advertisers in March 2026. Under suggestion mode, the seed list acts as a signal and Google may reach qualified people outside the former similarity thresholds to pursue goals such as conversions or a target cost per action. Google’s rollout guidance says the audience reporting table identifies Lookalikes as a “signal.”
In practical terms, a Broad setting should no longer be described as a hard ten-percent boundary when the account is in suggestion mode. Narrow and Balanced likewise guide prioritization rather than guaranteeing that every impression falls inside a defined similarity percentage. Google’s guidance says advertisers who prefer the former constraint behavior may opt out through the documented route, but the opt-out controls are being introduced later in 2026 and may not be visible in every account. Check the campaign’s audience controls and reporting before describing a Lookalike as constrained.
Google’s current explanation of the change is: “Your seed list will act as a signal for Google’s models, and the system may reach highly qualified customers outside of similarity thresholds to help you reach your campaign goals.” The Lookalike rollout guidance contains the account-level qualification.
Rank #2
How Optimized targeting works
1. You provide optional audience signals
When setting up a Demand Gen ad group, you can select audience segments that describe the people you consider promising. These signals can include your own data, custom segments, affinity or in-market segments, or other available audience types. They inform the system but are not necessarily delivery boundaries.
2. Google searches beyond those signals
Optimized targeting uses Google’s models to identify users it predicts are more likely to complete the campaign’s conversion action. If the selected audience does not provide enough opportunities, the system can move outside it. That means an ad group containing an in-market segment is not guaranteed to show only to people classified in that segment. The feature is designed around performance outcomes rather than a declared similarity relationship with a particular seed.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
3. Conversion quality determines whether expansion is useful
Optimized targeting is most logically aligned with campaigns that have a clear conversion objective and enough reliable conversion data for automated bidding and audience modeling. It does not promise incremental lift, a particular volume, or a specific cost per action. Google’s documentation defines its purpose; actual delivery depends on the account, campaign settings, available inventory, conversion measurement, and bidding.
What changed for the comparison in 2026
The old mental model—“Lookalike equals controlled similarity, Optimized targeting equals open expansion”—is now incomplete for Demand Gen. A Lookalike in constraint mode can still function as an audience limit according to the account’s available controls. A Lookalike in suggestion mode supplies a seed signal that Google can use while reaching outside the former band. Optimized targeting has always been oriented toward finding likely converters beyond the selected signals.
| Account state | How to interpret a Lookalike | What you should say in reporting |
|---|---|---|
| Legacy constraint behavior | Narrow, Balanced, or Broad can act as reach constraints according to the account’s enabled controls. | Describe the selected level and its documented target percentage, while noting that the percentages are approximate aims for users in the target location. |
| Suggestion mode | The seed and reach preference guide prioritization; Google may find users outside the former similarity threshold. | Call it a signal, not a guaranteed similarity boundary. Use the audience reporting table and campaign results to understand delivery. |
Because the rollout is phased, two advertisers can see different controls at the same time. Do not infer an account’s status from the campaign name, an old screenshot, or another advertiser’s interface.
Lookalike versus Optimized targeting: the decisions that matter
Seed-data relevance and quality
Lookalikes require a seed that represents the desired customer. A seed dominated by low-value leads, old users, accidental visitors, or mixed objectives can teach the model the wrong pattern. Before using one, check that the list is large enough for the account’s eligibility requirements, recent enough to reflect the current offer, and tied to the conversion you actually want.
Free tools Windows power users keep installed
One-click scans. No signup required.
Optimized targeting does not require you to define a similarity seed, but it still benefits from accurate conversion actions. If the primary conversion is a weak proxy—such as a page view when revenue is the goal—the system may optimize toward the proxy rather than business value.
Control versus exploration
A legacy constrained Lookalike gives you the clearest audience boundary. A suggestion-mode Lookalike gives Google a first-party clue while allowing broader exploration. Optimized targeting gives the model the broadest freedom of the two approaches because its selected audiences are signals rather than walls.
Campaign objective
Use a Lookalike when the question is, “Can we find more people resembling our best existing audience?” Use Optimized targeting when the question is, “Can Google find additional people likely to complete the conversion at the campaign’s efficiency goal?” Those questions can lead to different setups even when the same seed or audience segment is available.
Reporting clarity
Separate audience strategies when you need to know which approach generated results. Google recommends putting distinct audiences with different messages or targeting strategies in separate ad groups so management and performance reporting remain clearer. The Demand Gen audiences overview explains this ad-group approach.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Can you use both features together?
Yes. Google’s Demand Gen FAQ says Lookalike segments and Optimized targeting “work well together to help you achieve your advertising goals.” A practical combination is to add a relevant Lookalike as a strong first-party signal while allowing Optimized targeting to search for additional converters. This setup is not two independent audiences with guaranteed, non-overlapping delivery; both features can influence the model, and the system may expand beyond the Lookalike signal.
If you need a clean comparison, avoid putting every strategy in one undifferentiated ad group. Create separate ad groups—or separate campaigns when budget, creative, or bidding must be isolated—with one clear audience strategy per test cell. Keep the conversion action, attribution settings, creative, geography, schedule, and bid strategy as comparable as possible. The labels alone cannot prove who received an impression, so interpret results through conversion reporting rather than assuming that a Lookalike impression was necessarily delivered to a formally defined similarity group.
How to choose the setup: a step-by-step framework
- Define the conversion. Decide whether the campaign is optimizing for a purchase, qualified lead, registration, or another measurable action. Remove or de-prioritize soft events that do not represent success.
- Audit first-party data. If you have a recent, relevant list of purchasers, high-value customers, qualified leads, site visitors, app users, or YouTube-engaged users, consider a Lookalike. Document the list’s date range, inclusion rules, and relationship to the conversion.
- Choose the control you need. If delivery must stay within a similarity band, check whether the account still offers constraint behavior and whether an opt-out is available. If the account is in suggestion mode, treat Narrow, Balanced, and Broad as guidance.
- Decide how much exploration is acceptable. Select Optimized targeting when broader model exploration is acceptable and conversion efficiency is the primary objective. Use a Lookalike signal when the seed’s characteristics are strategically important.
- Separate materially different strategies. Put audiences that require different creative, bids, budgets, or reporting in separate ad groups, following Google’s Demand Gen audience guidance.
- Verify the account’s rollout state. Inspect the current audience controls and reporting table. Do not rely on documentation written before the 2026 transition or on another account’s interface.
- Set a measurement window. Allow enough time for the campaign to accumulate conversions, and record spend, conversion volume, cost per action, conversion value, and reach. Avoid declaring a winner from a few early conversions.
- Review actual delivery. Compare the audience labels and signal status shown in Google Ads with conversion outcomes. A label indicates configuration; it is not proof that every impression stayed inside a fixed audience.
Practical scenarios
Launching a new product with a strong purchaser list
A retailer with a clean list of recent purchasers can use those customers as the Lookalike seed. The seed gives prospecting a concrete definition of a likely buyer. If the account is in suggestion mode, the retailer should describe the setup as purchaser-informed discovery, not as delivery limited to a purchaser-similarity percentage. Optimized targeting can be enabled when the retailer is comfortable letting Google seek additional converters beyond that signal.
Generating leads with mixed-quality historical data
A business whose old lead list contains many unqualified contacts should not assume that a Lookalike will improve lead quality. First tighten the seed to qualified or closed-won leads if that data is available. If no reliable seed exists, a conversion-focused Optimized targeting setup may be more defensible, provided the primary conversion action accurately represents a valuable lead.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Protecting a tightly defined prospecting audience
If contractual, geographic, or brand requirements demand a similarity boundary, verify that the account can still use Lookalike constraint behavior. In suggestion mode, the former Narrow, Balanced, and Broad percentages do not guarantee that delivery remains inside those limits. If the required control is unavailable, document the limitation before launch and use the available audience settings and reporting to monitor expansion.
Testing similarity-led and performance-led approaches
Build separate ad groups with comparable creative and conversion settings: one using the Lookalike signal, another using Optimized targeting, and optionally a third using both. Keep budgets and bids intentional; unequal budget allocation can make a raw conversion-count comparison misleading. The result will be an account-specific observation, not a universal ranking of the features.
Common mistakes and how to avoid them
- Calling a suggestion-mode Lookalike a hard audience. Check the rollout state. In suggestion mode, Google may reach outside the former threshold.
- Treating 2.5%, 5%, and 10% as current delivery guarantees. Those Narrow, Balanced, and Broad figures come from the legacy constraint guidance and are approximate aims for users in the target location; they do not constrain delivery in suggestion mode.
- Using an irrelevant seed. A broad or stale list can produce an unhelpful similarity pattern. Match the seed to the campaign’s desired customer and conversion.
- Assuming Optimized targeting respects selected audiences as boundaries. Its purpose is to look beyond signals for likely converters.
- Mixing incompatible strategies in one report. Distinct audiences and messages are easier to manage and interpret in separate ad groups.
- Claiming guaranteed performance. Google provides feature definitions and intended behavior, not a quantified head-to-head lift. Judge performance with your own conversion and cost data.
- Confusing Demand Gen Optimized targeting with audience expansion. Google documents audience expansion separately for specified Video campaign goals. The related help page is About audience expansion; do not use that terminology as if it were a precise description of Demand Gen Optimized targeting.
What the available evidence can and cannot tell you
Google’s official documentation establishes the intended roles: Lookalikes are seed-based similarity signals, Optimized targeting is performance-oriented expansion, and the two can work together. The documentation does not provide an independent, named comparative study or a universal conversion-rate advantage for either setup. It also cannot reveal an individual advertiser’s rollout state without checking that account. Treat any claim that one feature always wins, or that a label alone identifies every person reached, as stronger than the documented evidence supports.
Bottom line
Start with a Lookalike when your first-party seed is relevant and you want its customer pattern to guide discovery. Start with Optimized targeting when conversion performance matters most and you are comfortable with Google searching beyond selected audience signals. During the 2026 Demand Gen transition, verify whether your account is using suggestion or constraint behavior: in suggestion mode, a Lookalike is guidance rather than a guaranteed similarity fence. Use both when their roles complement each other, separate materially different strategies for reporting, and let measured conversion quality—not the feature name—decide what to keep.
Frequently Asked Questions
Do Lookalike segments still limit targeting in Demand Gen?
Only when the account retains the legacy constraint behavior. In the 2026 suggestion-mode rollout, the seed and Narrow, Balanced, or Broad setting guide Google’s model, and the system may reach outside the former similarity threshold. Check the account’s current controls and audience reporting.
Can I tell from the audience label exactly who saw my ad?
No. A Lookalike or Optimized targeting label describes the configured signal or strategy, not a guarantee that every impression came from a fixed audience definition. Evaluate delivery through Google Ads reporting and your conversion data.
What should I test first if I have both a good seed list and a conversion goal?
Use the seed as a Lookalike signal and consider enabling Optimized targeting if broader exploration is acceptable. If you need to compare the strategies, place the Lookalike-only, Optimized-targeting, and combined approaches in separate ad groups with comparable settings.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
