Google can tell you the age bracket, the gender, the parental status and, in a lot of countries, the household income bracket of the people converting in your account.
It is sitting in a report most advertisers have never opened. Not because it is hidden, but because nobody ever told them it was there, and nothing in the interface volunteers it.
This chapter is about that report, and about the far more important question of when you are allowed to act on what it says.
Where It Lives and What It Holds
Open a campaign, go to Audiences, keywords and content, then Demographics. Four breakdowns are waiting.
Age, in brackets. Gender. Parental status. And household income, which is the one almost nobody knows exists.
Household income is reported in bands rather than dollars: top 10%, then 11 to 20, 21 to 30, 31 to 40, 41 to 50, and the lower 50%. It is available in roughly twenty countries including the United States, Canada, Australia, Japan and India, so if you do not see it, your market is likely why.
One structural note before you go looking. Demographics are available in Search, Display, Video and Gmail campaigns, but not in Shopping. If your account is mostly Shopping, this chapter is a smaller lever than it is for a lead generation business.
Observation Is Not Targeting
This is the distinction that decides whether the report is useful or dangerous, and the two words look similar enough that people mix them up.
Observation watches a segment without changing who sees your ads. Your reach is identical. You are simply collecting the data broken out by group.
Targeting restricts. Turn it on and you are telling Google to show your ads only to that segment, which narrows your reach immediately.
Start with observation, always. It costs nothing, it changes nothing, and after a few months you have real evidence instead of an assumption. Targeting is a decision you make after the observation has told you something, not instead of it.
The exception is when a demographic constraint is a genuine fact about the business rather than a preference. A service with a legal age requirement is not guessing when it excludes under-eighteens. That is not optimization, it is accuracy.
The Rule That Matters More Than the Report
Here is where most demographic analysis goes wrong, and it is not a Google problem.
You open the report, you see that one age bracket has a terrible conversion rate, and the obvious move is to bid it down or exclude it. That instinct is wrong far more often than it is right, because segment data is thin data, and thin data produces confident-looking patterns that are actually noise.
We run every segment through three tiers before anybody is allowed to touch a setting.
Under 50 clicks in the segment: look, do not act. The data goes in the report for transparency and no recommendation attaches to it. Not a weak recommendation. None.
Between 50 and 200 clicks: directional only. You may note the pattern. You may put it on a watchlist. You do not change a bid. The honest phrasing here is "early indicators suggest," never "the data shows."
**Over 200 clicks and at least 5 conversions: now you can act.** Both thresholds, not either. Two hundred clicks with two conversions stays directional, because a conversion rate built on two events is not a conversion rate.
The tiers are not bureaucracy. They are the difference between an adjustment that makes you money and one that quietly removes a profitable audience because eleven people in that bracket happened not to buy in March.
Small Accounts Stay at the Bottom Tier, and That Is Fine
The part nobody says out loud: on a modest budget, narrow segments may sit below the actionable threshold for months.
An account spending a few thousand a month, split across four age brackets, two genders and six income bands, is dividing a small amount of traffic into a lot of boxes. Most of those boxes will hold a handful of clicks. That is not a failure of the account or of the analysis. It is arithmetic.
Two consequences worth accepting rather than fighting.
Analyze at the broadest useful level first. Age brackets will reach a usable volume long before age crossed with income crossed with device does. Broad splits earn confidence faster because they hold more traffic.
Waiting is doing something. The instinct on a small account is to act on whatever data exists because it feels like managing. Acting at the wrong tier is not neutral, it is worse than waiting, because you are permanently changing who sees your ads on the basis of a rounding error.
This is exactly the same discipline the device split requires, and the same tiers apply there, and to geography, and to time of day. One framework, every segment.
Zero Impressions Does Not Mean Nobody Searched
Now a specific finding, and it is the one we catch most often in inherited accounts.
You open the demographics report and a segment shows zero impressions and zero spend. The natural reading is that nobody in that bracket searched. That reading is frequently wrong.
A zero can mean three different things:
The segment is actively excluded. Somebody set it as a negative, possibly years ago, possibly by someone no longer at the business.
A bid adjustment sits at minus one hundred percent, which is an exclusion wearing different clothes.
Genuinely no auction volume, which is the only one of the three that means what people assume it means.
Those lead to completely different actions, and you cannot tell them apart from the performance report alone. You have to go and look at the targeting settings. Reporting a zero as "no activity" without checking is how an exclusion survives for years, unexamined, in an account everybody believes is fully targeted.
There is a related habit worth building. Whenever you report a segment's performance, write the click count next to it. Not in a footnote, next to it. A conversion rate with no sample size beside it is a number that cannot be argued with, which is precisely what makes it dangerous.
The Unknown Bucket Is Bigger Than You Think
One more caution about reading the report at all.
Google reports an Unknown category across these breakdowns, for people whose age, gender or income it has not identified. It is often a substantial share of your traffic, and it is not junk. It is people, some of whom are your customers.
This matters when you start excluding. Google's own guidance is to exclude Unknown only if you are certain you want a genuinely narrow audience, because removing it can remove a large number of people you actually wanted.
It also caps what your analysis can claim. If a meaningful slice of your traffic is unidentified, then the pattern you are reading across the identified slice describes part of your audience, not all of it. Worth holding in mind before you describe your customer confidently on the strength of it.
When the Demographic Is the Strategy
The report is not only defensive. Sometimes the demographic is the business model.
We built an account where income targeting was deliberate: the offer was priced for the top income brackets, so we used Google's income bands as part of the targeting rather than merely observing them. That is a legitimate use, and Google's income data is more robust than most alternatives available for that job.
The distinction is whether the demographic constraint comes from the business or from the report. A premium service that only works for high-income households has a reason that exists independently of any data. Excluding the 41 to 50% band because it converted poorly across sixty clicks does not.
One honest caveat we have given clients: you cannot fully control for whether the person searching has the budget for what you sell. Income targeting improves the odds. It does not make the filter perfect, and treating it as though it does leads to disappointment when the leads still need qualifying.
The Twenty Minute Version
- Open Audiences, keywords and content, then Demographics. Look at age, gender, parental status and household income.
- Confirm you are in observation, not targeting, unless you deliberately chose otherwise.
- Write down the clicks and conversions per segment before you look at any rate. The sample size is the first number that matters.
- Apply the tiers. Under 50 clicks, look only. 50 to 200, directional and wait 60 to 90 days. Over 200 clicks and 5 or more conversions, you may act.
- Investigate every zero. Check whether the segment is excluded, sitting at minus one hundred percent, or genuinely quiet.
- Check the size of Unknown. If it is large, temper every conclusion you draw from the rest.
- Do not exclude a demographic on a poor conversion rate at the bottom tiers. That is the single most common way this report is misused.
- Do exclude when the business requires it. Age-restricted services and similar are facts, not optimizations.
The demographic report is one of the few places in Google Ads where the platform volunteers something about your customers that you did not already know. It is also one of the easiest places to do damage, because every number in it is smaller than it looks.
We pull demographics on every audit and we attach a tier to every line before we write a single recommendation. Most of what we find is worth watching rather than acting on, and saying so honestly is the point. Hand it over and we will run the pass properly. Or open the report yourself and read only the click counts. That column alone will tell you how much of what you are looking at you are actually entitled to believe.
