We use cookies. You have options. Cookies help us keep the site running smoothly and inform some of our advertising, but if you’d like to make adjustments, you can visit our Cookie Notice page for more information.
We’d like to use cookies on your device. Cookies help us keep the site running smoothly and inform some of our advertising, but how we use them is entirely up to you. Accept our recommended settings or customise them to your wishes.

What You Need to Know About Google Demographics for Search Ads, Including 20 Months of Data

Publically announced last Tuesday at Google’s Performance Summit but long available for marketers whitelisted into the Beta, AdWords Demographics for Search Ads targeting allows advertisers to adjust bids and targeting based on searchers’ age and gender.

Using performance data for one Merkle advertiser that’s had age and gender targets added to some campaigns in its AdWords account with 0% bid modifiers since October 2014, I’ll walk through some of the major points that marketers should understand as these targeting capabilities roll out to more advertisers.

Google Cannot Identify Age and Gender for All Users, but They’re Getting Better

The first and perhaps most important thing to understand about these targeting capabilities is that they do not target the entirety of searchers that are a particular gender or age, as Google is unable to attribute all searchers to age and gender buckets.

For example, here we have the share of total search traffic attributed to each age bucket as well as traffic share from users attributed to ‘undetermined’ for the campaigns studied:

As you can see, the share of traffic attributed to ‘undetermined’ declined from 69% in October 2014 to 47% in May 2016. Thus, while Google is attributing many more users to age buckets than it was two years ago, only about half of paid search traffic for this advertiser can be attributed to users tied to age buckets.

Taking a look at a similar chart for the gender demographic, we find that the share of traffic that cannot be attributed to a gender declined from 62% in October 2014 to 41% in May 2016.

This means it’s slightly easier for Google to assign users to a gender than an age bucket, which makes sense given that Google is using users’ declared information in Google accounts as well as other signals Google collects to infer these assignments.

While both gender and age are included in the personal information that’s necessary when signing up for a Google account, the non-declared information used to infer this demographic information about users, such as search history, is likely more predictive of gender than age.

It’s worth noting that Google didn’t always collect gender information from users signing up for a new account, and only started doing so in 2012 after the launch of Google+.

Regardless, both age and gender demographics can only be assigned to about ~50%-60% of searchers. This means that campaigns using age and gender targets with the 'target and bid option' selected for specific demographics (meant to limit ad reach to only those users included in the demographics targeted) aren’t actually targeting all of the users in those demographics targeted, but just the share that Google is able to attribute to a gender or age bucket.

Another important point comes up within the gender data, where we find that Google attributed roughly the same amount of traffic to males and females through the first several months of tracking. However, since September 2015 females have accounted for at least 40% of traffic every month while males have accounted for 17% or less traffic over the same time frame.

As this advertiser still sells very nearly the same exact products now as it did in 2014 and is bidding on basically the same keywords, it’s unlikely that the demographics of search traffic have actually shifted this much over the past 20 months.

This calls into question how accurate Google is when it does assign a user to a gender or age, or at least how accurate they used to be, though isn’t proof in and of itself that these buckets are incorrect. Just something to keep in mind when making use of these targeting options.

It’s also important to go into demographic targeting with an open mind in terms of which ages and gender might perform best for a particular brand.

Performance Should Drive Strategic Decisions

One point I can’t stress enough: don’t assume which demographic groups perform best for you in search.

While you may know the age groups and gender that drive most of your brand’s orders and sales, the only way to know which users perform best in search is to add these demographic groups to your campaigns.

For example, while I discussed earlier that our long time demographic advertiser sees significantly more traffic attributed to females, conversion rate is almost identical for males and females. Thus, male traffic is just as valuable on a per-click basis as female traffic, even if it accounts for a much smaller share of traffic.

As such, advertisers shouldn’t go into age and gender targeting thinking they’ll just target all campaigns to the demographic buckets that perform best for their brand overall and exclude all the others. Aside from the questionability of how well these demographic groups are getting populated, you may be surprised by how these gender and age buckets perform compared to one another.

The only way to know is to add these targets and accumulate data, and our example advertiser now has twenty months of data on how different genders and ages perform for different campaigns to inform optimization and targeting decisions.

What else might advertisers want to know going into using age and gender targets?

Here are a few additional points:

  • Age and gender demographics are not currently available for Google Shopping campaigns. This limits how much they can play into ecommerce advertisers’ strategies, as Google Shopping accounted for 70% of all non-brand clicks for Merkle retail clients in Q1.

  • Average position does not populate for age and gender targets, so advertisers will not be able to use this metric in measuring the impact of bidding adjustments.
  • Users deemed under the age of 18 are not attributed to an age bucket and clicks from these searchers are rolled up to undetermined.


Even if Google’s ability to assign users to gender and age buckets is somewhat limited for now (and perhaps a bit questionable), it’s great that more advertisers will now be able to use Demographics for Search Ads age and gender targets, which have long existed in Beta.

I highly recommend taking advantage of these targets if you aren’t already to at least begin collecting data on how gender and age groups perform once Google begins rolling them out in earnest, though the timeline for when they will be available to all advertisers is currently uncertain.

Further, with Google’s formal announcement marketers can finally begin publishing best practice tips and tricks for making the most of these targeting options.

The fun has just begun.