Generative AI for Google Demand Gen campaigns: first impressions
We’re now three weeks on from the launch of Google’s new generative AI tools for Demand Gen campaigns and have been able to put them to the test. So, where do we stand on this new addition to our digital marketing armoury?
Well, one thing’s for sure: Google's commitment to Demand Gen is unwavering. As is the ongoing momentum of generative AI.
In this article we’ll share our stance on the immediate effectiveness of this new Google Ads feature and its ongoing potential. We’ll also share our take on who we think would benefit most from generative AI for Demand Gen campaigns, for what purposes, and how you can get the best results out of the tool.
What is Google Demand Gen?
If you’re already well-versed on Demand Gen campaigns, feel free to jump to the next section. For those who aren’t fully up to speed, Demand Gen campaigns were introduced to Google Ads in 2023, enabling advertisers to utilise highly engaging creative assets in a range of ad formats, across action-centric platforms such as YouTube and Discover feeds.
This campaign type aims to discourage passive scrolling and promotes brand storytelling. This not only enables organisations to hone their brand image, but it empowers them to create demand for their products and services through powerful, creative-first advertising.
Generative AI for Demand Gen
On April 16th, Google announced the introduction of a generative AI feature for Demand Gen campaigns.
The Google AI-powered tool allows advertisers to generate “stunning, high-quality image assets”, optimised for PPC campaigns. In a similar way to Chat GPT (or tools such as DALL-E for those with specific experience with image generation), the user creates and refines their prompt and lets AI take care of the rest.
Another key element to this new feature is the “Generate more images like this” functionality, which allows advertisers to immediately recreate similar versions of their best-performing ads. This feature is not to be sniffed at – when it comes to testing and ongoing optimisation, being able to create new variants in seconds without going back to the Design team is a huge time saver and can have a big impact on campaign optimisation.
How does it work?
Google’s announcement reiterated that the AI generated imagery is still very much driven by the user, stating “your knowledge and expertise are crucial to help Google AI generate images tailored to your business or client's needs”.
Because of this, the user also gets full control of which images are selected and integrated into Google Ads Demand Gen campaigns.
Google provided an example prompt for ads targeted towards users researching camping trips to Iceland, on behalf of an outdoor lifestyle business that sells camping gear:
“vibrantly colored tents illuminated from the inside under the aurora borealis with snow capped mountains in the background”
It’s important to note that Google AI will never generate the same image twice, meaning that every asset created is completely unique to its creator.
The good
By just being able to save design time and create new imagery as you go - especially if you’re just testing new variations of high-performing existing campaigns - is already a massive bonus.
Even if you opt to not use the generated creatives, it’s a good option to have. At the very least, it’s a great tool for coming up with new ideas.
When it comes to the pros and cons of generative AI for Demand Gen campaigns, it’s also important to think about who the user is. For example, small businesses with limited design skills will benefit from being able to generate imagery without having to outsource resource or opt for an alternative campaign types. Large businesses with stringent brand guidelines and more of a spotlight on their advertising may opt to avoid AI generated imagery altogether.
Generally, we’ve found the functionality beneficial for the following reasons:
Optimised for Google – images are generated in the exact specifications and dimensions required for relevant Google placements.
Efficiency – there’s no need to leave Google Ads to find and export files, modify assets, reupload images, etc.
Testing – being able to generate several asset variations makes multivariate testing much more feasible.
Creativity – not having to commit to a creative brief enables advertisers to explore and properly visualise unlimited design ideas.
Performance-driven - being able to make immediate modifications within the Google Ads interface, with campaign data at hand, encourages a data-driven approach.
As a result of the points above, businesses generate much higher rates of engagement across a wider range of placements, which in turn has seen an increase in conversions.
That’s really the point. After all, Google launched generative AI with the following justification:
“To keep audiences engaged and deliver results, advertisers need to diversify their creative strategy with multi-format ads. In fact, those who run both video and image ads with Demand Gen campaigns saw 6% more conversions per dollar than those running image-only ads.”
From what we’ve seen so far, it’s hard to argue otherwise.
The not so good
If you do have experience using other image generating tools such as DALL-E or Midjourney, Google’s AI will feel a bit off the mark from a capability and output perspective. There are restrictions in place around including faces and branding, which limits the creative potential.
That said, Google’s AI image generator is limited for a reason. Firstly, AI generated faces do tend to be on the whacky side, and the inclusion of branding can lead to obvious issues around copyright. One significant benefit to the restrictions is that you won’t see your ads being disapproved, a common headache that can take days to resolve.
As alluded to earlier in the article, the tool doesn’t account for brand guidelines. So, it may also be difficult for advertisers to build and maintain brand image.
How to get the best out of the functionality
Here are our top tips for using Google’s generative AI for Demand Gen effectively:
Assess your requirements – Do you have design resource at hand? Are you limited by brand guidelines? Do Google’s restrictions prohibit you from effectively representing your services? If the answer to any of the above questions is “yes”, this probably isn’t the solution for you.
Utilise the “Generate more images like this” tool – this is another way of saying “let’s do more of the stuff that works”. Just make sure to apply this feature to the high-performing ads only, and test new variations, rather than replacing the old ones.
Learn as you go – Don’t just test ads, test your input too. We’ve found small tweaks to prompts make a big difference. Despite the Aurora Borealis example provided by Google, as a general rule of thumb we’ve found that shorter prompts work better for simple creatives. That said, the more niche the product or service you’re advertising, the more creative you will need to get with your prompts.
Conclusion
If there’s one thing we’re sure of, it’s that generative AI for Demand Gen campaigns is only going to improve. Given that Google have created this functionality, we also know that Demand Gen campaigns are here for the long haul.
Whilst limitations are obvious currently, this is a tool that is undoubtedly beneficial for a multitude of reasons. Learning to reap the rewards of Demand Gen campaigns now will only provide a competitive advantage going forward.
In a similar sense to Performance Max campaigns, it’s also important to not throw caution to the wind. Don’t discount the power of manual activity or become reliant on AI, rather take a mixed approach. Only use AI generated assets that genuinely enhance your campaign and represent your products or services. Don’t replace strategy. Continue to review performance and optimise campaigns as regularly as you would do without the use of this feature.
As always, we hope this article puts you on the path to achieving your digital marketing goals. If you’d like further advice or to discuss how Ducie Digital can help you to meet your goals by using pay-per-click advertising, get in touch today.