How Optimized targeting works
On Display, Demand Gen and some video campaigns you pick audience segments: in-market lists, custom segments, your own site visitors. Optimized targeting treats those as a starting point rather than a fence. It watches who actually converts, reads signals from your landing page and creative, and serves to other people it judges likely to do the same, whether or not they belong to the segments you chose.
What it does not cross is the rest of the campaign. Exclusions, location and language settings and negative lists all still apply. The boundary it steps over is your audience list, not your account’s rules.
It also depends completely on your conversion data, because that is the only description it has of a good outcome. Whatever you told the account to count as a conversion is, in practice, the instruction you have given it.
Why Optimized targeting matters
It decides who sees the ads, which means it can quietly change what a campaign is. A report showing that your carefully built audience performed well may really be describing people the system found on its own, and the headline figures will not tell you which.
It is also switched on by default in several campaign types, and Google has changed which ones over the years. Most advertisers are therefore using it whether or not they decided to, and knowing it is there changes how you read every audience report the account produces.
Common mistakes with Optimized targeting
The damaging mistake is leaving it on while the conversion action is weak. If the campaign counts page views, newsletter sign-ups or a click on a phone number that nobody actually dialled, the system will diligently go and find more people who do exactly those things. Loose conversions plus automatic expansion is the quickest way to buy a great deal of cheap, worthless activity.
The second is judging it on reach. Wider is not the result you wanted; cost per real enquiry is. The third is switching it off on instinct and never checking whether the campaign did better without it, which leaves you holding an opinion where evidence was available.
How to act on it
Fix the conversion action first. Count what genuinely makes money — a submitted enquiry, a completed sale, a call that was answered — and take the soft actions out of the column that drives bidding. Everything else about this setting rests on that decision.
Then give it something worth learning from. A customer list, past converters or visitors to a high-intent page are far better audience signals than a broad interest segment, because they describe buyers rather than browsers. Keep exclusions current at the same time, since they are now doing the job your audience list used to do.
Finally, test it honestly. Run comparable campaigns with it on and off, leave both alone long enough to settle, and compare cost per acquisition rather than impressions. In a small market such as Nepal, expansion often has nowhere useful to go and the honest answer is that the setting made little difference either way.
