In 2019, Google released Discovery Ads to show ads in Google's discovery spaces, such as the Discover feed on mobile devices, YouTube and Gmail. These campaigns brought an innovative way to reach consumers through visually rich, highly personalized experiences. Four years and an AI boost later, we now have its replacement: Google Demand Gen with the promise of radically transforming the way brands interact with their target audience.
These AI-powered campaigns are reminiscent of performance max campaigns also quite new.
🤨 We break it down for you.
What is Discovery Ads?
Before 2019, Google already had search ads, which (still today) have the advantage of being highly relevant because they are displayed to users according to their expressed needs. A user types in "buy blue bike", and we show them an ad for this bike brand that has blue bikes. Simple and effective, everyone's happy. There were also display ads on the display network (2M third-party partner websites and applications), which display traditional advertising inserts.
And then Google brought out Discovery campaigns, to integrate more naturally into the user experience: the emphasis is on images and storytelling, the locations are specific, more qualitative (because they're owned by Google and therefore less random) and a little less intrusive like the Discover feed (on mobile), YouTube's homepage and Gmail's Promotions and Social Media tabs. Targeting is primarily based on the user's search and browsing activity.
The aim of Discovery (and Display for that matter) is to broaden the audience for a product. It's no longer just those who just typed in "buy blue bike", but also those who researched bikes some time ago, visited a bike e-commerce site, or saw a YouTube video, or visited websites on other similar subjects such as an electric scooter site. Basically, Discovery Ads rely not only on search activity, but also on intent signals based on user behavior on the various Google platforms (search, YouTube viewing, etc.).
What's new with Demand Gen?
Demand Gen is more than just a name change for Discovery Ads. The idea is to simplify the interaction between customer and advertiser:
Source: Tobias Maurer, Google Group Product Manager
Clearly, the new features are up to scratch:
✅ Broadcast slots: YouTube Shorts, In-Stream
✅ Experiential possibilities: Image + video in the same campaign, new creative process (boosted by AI), A/B test
✅ Bidding options: Maximizing clicks
✅ Reports: Conversion tunnels, Attribution through data, Impact on brand perception and awareness, Impact on purchasing behavior
✅ Audiences: Lookalike, Targeting by device
While every element is welcome, the most important points are the new locations, Lookalike targeting (as with Meta) and the new creative possibilities and experience.
The benefits of these new locations
Demand Gen expands horizons by integrating new locations, including YouTube Shorts and In-Stream. YouTube Shorts, Google's answer to TikTok, has become a very popular content space, enabling advertisers to reach a young, engaged audience. As for In-Stream, it offers the possibility of integrating ads into videos as they play, increasing the visibility and memorability of ads.
💡 These are therefore particularly desirable locations to reach over 3 billion monthly users AND to better cover the middle of the sales tunnel.
Parameterized audiences vs. Lookalike targeting
Inspired by the success of Meta's "lookalike" audiences, Google now also offers lookalike targeting. This feature enables advertisers to target users with similar behaviors and interests to their best customers. Utilizing Google's massive database, this feature offers unrivalled targeting precision, maximizing return on investment.
Instead of us defining the type of person to show our ads to, Google analyzes the characteristics and behaviors of users in your original audience (for example, people who have purchased on your site or filled out a form) and finds other users with similar profiles and behaviors.
👉🏻 Basically, Lookalike targeting aims to find new customers by reaching people who are similar to your current customers.
New creation and configuration
🤯 No keyword selection required.
🖼️🎥 Combining images and video in the same campaign offers unprecedented flexibility. This hybrid format captures attention and increases engagement. What's more, with the new AI-assisted creative process, campaigns are optimized in real time to deliver the best performance. Integrated A/B testing also allows advertisers to experiment with different ad variations to determine the most effective. This means that creative is becoming increasingly important.
🙌🏻 A product feed can be added (as is already the case with Shopping Ads and Dynamic Display Ads). A big win for e-tailers. This means that Demand Gen Ads can automatically display the most relevant products based on the user's buying or browsing behavior. It's a mini-revolution that will save you time and make your ads even more relevant.
Getting the most out of Demand Gen? campaigns
1️⃣ Start with work with and diversify your creations and your copywriting. As I said earlier, graphic assets will become even more important than before. You can also use A/B testing to try out different hooks. diversify = explain that we're now broadcasting on YouTube Shorts formats, so we want story-style videos like on Facebook in addition to traditional YT formats. For info: 1 landscape video max + 2 story videos per Adset 1 wide image max + 2 others per Adset
2️⃣ Set up good tracking of your conversions (if you haven't already) to make the most of the lookalike.
3️⃣ We optimize by fine-tuning data analysis in Analytics (particularly with UTMs, especially since iOS 17 now removes gclid), and we arbitrate between different placements (of which there are now many), different audiences (lookalike or "normal" audiences, and even different lookalike audiences from different conversion actions depending on the case), and by further optimizing creatives.