Google's introduction of its Demand Gen campaigns late last year may have garnered less fanfare compared to the Performance Max launch in 2021, but it marks a very significant update to the platform. It underscores the clear direction in which Google Ads is evolving, highlighting its ongoing commitment to first-party data and artificial intelligence.
Demand Gen campaigns run across YouTube, the Discover Feed and Gmail. They support landscape and vertical videos, images, and for the first time in Google Ads, also carousels.
Demand Gen campaigns, true to their name, are designed to generate demand, strategically positioned at the top or middle of the purchase funnel. Bridging the gap between view-based video campaigns and Performance Max, they nurture potential customers who are likely to complete their purchase through a click on either a Performance Max or a search ad.
These campaigns offer three primary goals: conversions, conversion value, and clicks. Unlike Display or Video campaigns, options for view or impression goals are absent, emphasising their distinct focus. Advertisers have the flexibility to either let the system optimise for maximum results or to set specific targets like cost per conversion or return on ad spend (ROAS). Notably, there is no option for a target cost per click (CPC).
Like other modern campaign types, Demand Gen mixes and matches several assets, namely videos, images, headlines and descriptions, to find the best combination for the best user at the best time.
But unlike other campaign types, including Performance Max, Demand Gen only runs on Google-owned properties. It doesn’t run on partner websites or on the Google Display Network.
This is what sets Demand Gen campaigns truly apart from the rest of the Google offering.
However, there are also other aspects that make Demand Gen a novelty in the Google Ads ecosystem:
The scheduled blockage of third-party cookies by Google Chrome by the end of 2024 will significantly alter audience targeting across the open internet. Without third-party cookies, ad platforms like Google Ads will be unable to recognise a user consistently across different websites.
For instance, a user who visits site A and then moves to site B will appear as two distinct users. This limitation profoundly impacts both retargeting and interest-based targeting strategies.
While the advertising industry is actively researching privacy-centric, long-term alternatives to third-party cookies, currently the most viable solution is the use of first-party audiences. To be more precise, the use of first-party audiences within a first-party environment.
First-party audiences refer to any user data that we directly gather from our users, as opposed to data that we buy from third-party providers like Google or Meta.
For instance, a list of users who have made a purchase on our website constitutes a first-party audience. This data is typically collected through advertising tags or CRM systems, directly on our website.
Conversely, a list of users interested in winter sports would be considered a third-party audience. This is because such a list isn't compiled from direct interactions on our site.
Instead, it's aggregated by ad tech providers like Google, who analyse users' browsing activities across various websites to deduce their interests.
Now, what do we do with those audiences?
Imagine you want to re-engage users who added items to their cart but didn't complete the purchase. One approach is using retargeting ads.
Purchasing these ads on the open internet, like display ads on the Google Display Network, presents a challenge in a world without third-party cookies. The issue is that the website where the ad appears cannot recognise your users.
This limitation persists even when you're targeting a "first-party audience." As a result, despite targeting your users, you can not display personalised ads due to the inability of the third-party publisher's website to match its user data with your customer list.
This scenario illustrates the use of a first-party audience in a third-party environment, essentially targeting your customers on someone else's online property.
Now consider the same retargeting strategy, but this time, your ads are displayed only on YouTube or Gmail, rather than the open internet. Even though YouTube and Gmail are “third-party” platforms to your business, they integrate seamlessly with Google Ads by allowing you to upload your customer email lists. Google then matches these emails with the Google accounts of individuals logged into YouTube or Gmail, enabling precise targeting of your own users on these platforms.
This approach, while theoretically possible with other publishers' websites, is less feasible in practice. No single publisher has the vast user base comparable to Google's platforms, making it challenging to effectively match customer emails.
Furthermore, attempting this with multiple publishers would be incredibly labour-intensive and inefficient. Platforms like YouTube, Gmail, and Discover boast billions of logged-in users, providing advertisers with the necessary scale to effectively target first-party audiences.
Given that this strategy encompasses not just one, but three distinct platforms, the term “environment” becomes a better fit. Google, in this scenario, offers advertisers a "first-party environment" to connect with their own first-party users.
While this approach might seem groundbreaking, and in some ways it actually is within the context of Google Ads, it's important to recognise that it’s not novel. Meta and other social media companies have successfully employed similar strategies for over a decade.
With Demand Gen, Google is finally taking from the social media playbook!
What if you want to go beyond retargeting and want to find new users, effectively generating demand?
This is where Lookalike audiences come in!
A staple in social media advertising for years, Lookalike audiences are finally making their debut into Google Ads.
Historically, Google offered similar features like "similar audiences", "audience expansion" and "optimised targeting." However, with the introduction of Demand Gen, the term “Lookalike” is deliberately used, resonating more with advertisers familiar with social media advertising.
This renaming reflects Google's ongoing strategy to integrate tactics from the social media playbook.
Advertisers can base their lookalike audiences on their pixel-based or customer match segments. They can choose three levels of expansion, 2.5%, 5% or 10%.
This strategy of integrating typical social media features isn't confined to Google Ads. It's also evident in the evolution of Google's own products. For instance, the rise of YouTube Shorts in recent years clearly mirrors TikTok's format.
As social media evolves from pure social networking to entertainment media, and as YouTube enhances its social capabilities with features like posts and Shorts, we're witnessing a convergence of these two worlds. This integration is naturally extending to advertising products as well, although I see it only with Demand Gen for now.
A first-party environment and Lookalike audiences are not the only features that Demand Gen takes from social ads. For the first time ever in Google Ads, advertisers can now create and run Carousel Ads.
Carousel Ad from Tages-Anzeiger, a popular Swiss newspaper.
Advertisers can add up to 10 carousel cards. Each card has its own headline and Final URL.
In my experience so far, Carousel Ads tend to underperform compared to Videos and Images. Nevertheless, it’s worth testing them for yourself as they provide room for creativity.
The campaign creation flow for Demand Gen has been overhauled, mirroring the user experience of Meta ads. This redesign allows for seamless navigation between campaign settings, ad group settings, and ads using the same left-hand menu, eliminating the need for a "back" button. For those acquainted with Meta Ads, this new interface will feel strikingly familiar.
Although Demand Gen campaigns have been around for less than a year, I've already observed some patterns in their performance. Google suggests benchmarking Demand Gen against social ads. I agree with that, but I also find it useful to compare them with other Google campaigns, such as Video and Display.
Generally, I've noticed that high-quality videos uploaded in Demand Gen campaigns often attract the majority of traffic, overshadowing images and carousels in the same campaign. From my experience, carousels barely run. However, I would still recommend having them in your campaigns, as your product might be more suitable for this specific format.
CPM and CPC tend to be closer to that of Video for Action campaigns on YouTube rather than that of image-based Meta Ads campaigns, which are usually cheaper. Conversion rate tends to be better on Demand Gen than on video-only campaigns.
In my experience, Demand Gen is really a hybrid between traditional video campaigns and social ads. The perfect middle man between top funnel strategies and Performance Max or search.
Demand Gen campaigns are still new, but they already offer a glimpse into the future of Google Ads. First-party strategies, a streamlined campaign creation flow and audience expansion options will probably propagate to other campaign types very soon.