While Bing and Yahoo have been offering this feature for quite some time now, it wasn’t until very recently that Google finally made it possible for all advertisers to use demographic targeting over the Search Network. This meant that, until October of this year, there was simply no way to make sure your Search ads were targeting people within the right age and gender bucket. This was a serious shortcoming for Adwords advertisers, as nearly all businesses are age sensitive, and a lot of businesses are gender sensitive too.
Think of the case scenario where a man is planning to purchase leather jackets over the internet. A lot of people don’t bother to type “men’s leather jackets” in Google, they would simply type “leather jackets”. Up until recently, there was no way to prevent these men from seeing ads that read “women’s leather jackets”, which inevitably resulted in missed opportunities for Adwords advertisers.
A good example of age sensitive products would be mortgages, since people under 30 years old often don’t have sufficient means to afford a loan, and people over 50 years old typically are unlikely to repay their loan by the time they retire. So by allowing all age groups to request a mortgage on your website, you might get lots of leads from people in their twenties or in their fifties, maybe even at an interesting Cost Per Lead, but you’re unlikely to take those leads very far in the purchase funnel.
But just how accurate is Google demographic targeting? Well, judging from the amount of users that Adwords lists among the “Unknown” age and gender bucket, it seems like Google still need to do quite a bit of work to get it right…
How does Google know your age and gender?
For those who might have forgotten it, Google kindly asked you to fill in your age and gender details when you opened your Gmail account (unless you opened it prior to 2012, before Google started collecting gender data). On top of that, Google uses sophisticated algorithms to determine the likelihood of your gender based on your browsing history. This means Google can’t know for sure your age and gender, but they’re getting better at it judging from the steady decrease in the Adwords’ “Unknown” demographic bucket over the past few years.
From what I could gather on a number of Adwords accounts, demographic targeting accuracy would be over 50% as of 2016, with less than half of Google users listed under the “Unknown” category. That’s still a lot of people whose age and gender can’t be targeted effectively through Adwords campaigns, but hopefully Google will be able to fill that gap within the next few years…
Why did it take Google so long?
Demographic targeting has been available over the Google Content Network for years now, so how come it took so long for Google to include it in their Search Network? After all, what made Google’s success in the first place is how they were able to focus on improving our experience while searching for products and services, at a time when search engines looked more like a jungle than a well organized market place. So why hold back on a feature that would dramatically improve people’s search experience?
Let’s not forget that around 75% of Google’s total revenue is generated by Adword’s Search network (versus 20% for the Content Network). Any wrong step in the monetization of their Search network would result in dramatic losses for the company. And it’s actually quite easy to misconfigure your Google account by relying too much on demographic data. Many less experienced advertisers could decide to filter out all the users listed in the Adwords’ “Unknown” demographic bucket, and thus lose about half of their potential customers. Or they could also decide to only target people within a certain age range without having performed enough data analysis over their existing customer base… all of which would ultimately damage Google’s revenue stream.
Age and gender targeting are already a big step ahead for Adwords Search marketers, but there are many other areas of interest that Google haven’t included in their Search targeting yet, probably due to lack of reliable user data. Let’s hope that in the future they’ll be able to include in their Search network targeting some of the data that’s already available in Content network targeting such as user interests and in-market segments.