Audience Targeting

Demographic Targeting vs Optimized Targeting: Fences and Autopilot

In short: Both shape who sees the ad, but at different priority levels. Demographic targeting is a fence you draw by hand: excluded brackets are never served, full stop. Optimized targeting is autopilot: it expands delivery toward predicted converters, treating your chosen audience segments as hints it may leave. The mechanically important fact is that the fence outranks the autopilot - optimized targeting can wander beyond your segments, but it cannot serve into a demographic bracket you excluded. That makes them collaborators more than rivals. Use demographics to define who you must never pay for, and optimized targeting to find more of whoever converts inside those fences.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Demographic Targeting

Restricts or adjusts delivery by age, gender, parental status, and household income, using the profile data Google has declared or inferred for signed-in users.

You set demographics at the ad group or campaign level, either excluding brackets outright or layering bid adjustments on top of them. It earns its keep as a subtractive control, cutting bands you cannot legally or profitably serve. Two traps: a meaningful share of impressions carry unknown demographics and get excluded alongside the rest, and housing, employment, and credit categories block these controls entirely.

Full definition

Optimized Targeting

Lets Google find additional converting users beyond the audiences you selected in a Display, Demand Gen, or Video campaign, treating your segments as a starting signal.

Enabled at the ad group level, it uses your chosen audiences and live conversion data as seeds, then bids on people outside those lists who look likely to convert. It is the intended successor to similar segments. It only behaves well when conversion tracking is reliable, and advertisers routinely confuse it with audience expansion, which widens reach without optimizing toward a conversion goal.

Full definition

Side by side.

The differences that actually change what happens in your account.

	Demographic Targeting	Optimized Targeting
Direction of control	Narrows delivery by removing or reweighting brackets.	Widens delivery by adding predicted converters beyond your segments.
Who decides	You, explicitly, bracket by bracket.	The model, per auction, against the conversion goal.
Hard boundary or hint	Hard - an excluded bracket is never served, even with automation on.	Soft - your chosen segments are a starting signal it is free to outgrow.
What it needs to work	Nothing beyond the settings themselves.	Reliable conversion tracking with enough volume to learn from.
Where it reports	Per-bracket rows in the demographics tables.	An aggregate expansion line in audience reporting, with no composition detail.
What it costs you when wrong	A converting bracket sits silently excluded until someone audits the settings.	Budget scales confidently against a broken or soft conversion signal.
Interaction with the other	Its exclusions still bind when optimized targeting is on.	Expands only within whatever demographic fences remain.

What actually separates them.

Demographic exclusions survive optimized targeting - the model may leave your chosen segments but cannot enter a bracket you excluded - so the two operate at different priority levels rather than as alternatives.

Demographics are static until you edit them, while optimized targeting re-decides eligibility at every auction.

Demographic controls need no conversion data at all, while optimized targeting is only ever as smart as the conversion stream feeding it.

A demographic mistake is visible as a missing bracket in a report anyone can read, while an optimized targeting mistake hides inside an aggregate expansion line.

Under Smart Bidding most demographic bid adjustments are ignored while outright exclusions are honored, so once automation is on, your demographic lever effectively collapses to in-or-out.

Which one should you use?

Use Demographic Targeting when

You sell an age-gated product and some brackets must never be served.
Bracket-level reporting shows consistent non-converting spend you want cut, not reweighted.
A stakeholder needs a guarantee about who cannot see the ad, not a prediction.
You are running under manual bidding, where bracket bid adjustments still actually apply.

Use Optimized Targeting when

Your conversion tracking is solid and you want delivery to scale past your lists.
Your audience segments are too small to spend the budget on their own.
You are replacing a legacy similar-audiences setup with the current mechanism.
You would rather define outcomes and fences than hand-pick who qualifies.

Common questions.

Does optimized targeting override my demographic exclusions?

No. Exclusions - demographic brackets and excluded audiences - still bind when optimized targeting is on. What it treats as flexible are the audience segments you selected as targeting, which become signals rather than limits. Draw the fences with exclusions and let the toggle roam inside them.

Should I bother with demographic bid adjustments under Smart Bidding?

Mostly no. Smart Bidding sets bids from its own predictions and ignores most manual adjustments, though it honors outright exclusions. If a bracket underperforms badly enough to act on, exclude it; if it is merely soft, let the bidder price it down on its own.

My campaign is serving outside the age range I wanted - why?

The usual cause is that the other brackets were never excluded - selecting a bracket for observation does not restrict delivery, and the Unknown bracket stays eligible unless you exclude it too. Optimized targeting is rarely the culprit here, since it respects demographic exclusions. Fix it in the demographics table, then decide separately what the unknown pool is worth to you.

Can optimized targeting run with no audience segments selected at all?

Yes. Segments are optional seeds - without them it leans on the campaign's conversion data alone, which works but usually learns slower. Adding a tight list of converters or engaged visitors gives the predictions a head start, which is exactly the role seed lists play now.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint free See how it works

Related comparisons

All comparisons