At the end of May we were fortunate to attend Google Marketing Live 2024 in Dublin. Over the two day event we got to hear more about the upcoming product launches for the year and was also a great chance to chat with product managers and give them feedback on the upcoming roadmaps.
It was interesting to contrast this year with the 2023 event, where the focus was on surfacing AI in the Google Ads platform after OpenAI brought ChatGPT to the public at the close of 2022.
The messaging was slightly different (AI heavy, of course), as by now many of us have had a chance to test out the initial round AI powered products. Overall the theme was building on the ground laid back in 2023, as Google sees what products advertisers are taking up, and which they are not.
We also saw a few nods to Google I/O and Project Astra, Google’s next generation AI assistant. Whilst not directly related to advertisers right now, search is deeply baked into the product - is this the future, or another Google Glass?
Whilst we covered a lot of ground, here are some of my highlights:
Ads In AI Overviews
One of the big ticket announcements was plans to insert ad units within AI Overviews, Google’s recently launched update to the SERP (currently available in the US)
As it stands, those who are served AI Overviews may see ads above or below:
We now know the next step is ads within the AI Overview itself. The example provided was a user searching for how to remove wrinkles from their clothes. The AI Overview detailed the options available and added a promoted listing for a wrinkle removing clothing spray:
The timing of this announcement felt quite early, with AI Overviews only just launched. In what seems to be becoming a trend when it comes to Google’s AI product launches, it's not been a particularly smooth one (having previously pulled Gemini’s image generation capabilities back in February)
We’ve already seen feedback from SEOs and users on the types of answers being provided by AI Overviews, from dogs playing in the NHL or the health benefits of eating rocks.
Google have responded, including some more details on how this happened:
One area we identified was our ability to interpret nonsensical queries and satirical content. Let’s take a look at an example: “How many rocks should I eat?” Prior to these screenshots going viral, practically no one asked Google that question. You can see that yourself on Google Trends.
There isn't much web content that seriously contemplates that question, either. This is what is often called a “data void” or “information gap,” where there’s a limited amount of high quality content about a topic. However, in this case, there is satirical content on this topic … that also happened to be republished on a geological software provider’s website. So when someone put that question into Search, an AI Overview appeared that faithfully linked to one of the only websites that tackled the question.
The fact remains it’s jarring to see such low quality responses being returned, although work is clearly being done to improve this.
Adding ads into this mix feel like it only complicates the situation, but is certainly one for advertisers to watch out for (and any reporting options that may or may not appear)
2024 Search Trends
Early on in the first keynote we heard a few updates on wider search trends:
15% of searches being made are new to Google
This number should ring a bell - this percentage has remained the same since 2013, despite overall growth in query volume
Faster growth across Exactly-What-I-Want queries than short tail queries
Moving from short queries with fewer words, to longer tail queries with more words (modifiers, descriptors etc..)
Complex Search, where the ideal response would involve data aggregated from several different sources
There was no detail on growth in users searching this way - but does tie nicely into what Google is doing with AI Overviews.
Next was Search Beyond Words
This can be wrapped up under multimodal search, which came up a few times and covered in more detail below.
The last was Search Beyond Answers, which felt like a build on the second Complex Search trend.
This felt more like an aspirational goal, presenting users with custom results that best answer open-ended queries, going beyond just an AI Overview and reimagining what the SERP is.
It was quite surprising to see that the growth in Exactly-What-I-Want queries did not move the needle on new query percentage (completely flat at 15% for over a decade now), but perhaps not too surprising given the scale of queries being made each and every day.
Multimodal Search
This emerged as a theme over the two days across both keynote and product sessions, with quite a few mentions of products like Google Lens or Circle to Search on Android. This aligns with Google seeing themselves as a “matchmaker” and not being limited to traditional text search - if you see some dinnerware you’d love to own, just take a photo and Google will help you find it.
We heard Google Lens currently serves 12 billion average monthly queries, and will shortly feature ad slots, an easy fit with shopping ads where advertisers are providing imagery to match queries with in Google Merchant Centre.
Plotting out data provided by Google at previous I/O events, after a spike across 2022 Google Lens is seeing steady (not explosive) growth over the years:
Source: https://arinsider.co/
As a point of comparison, Statista estimates that Google serves around 255 billion text queries per month, which would put visual search at around 5% the volume of traditional text queries.
Google seem to be betting on Circle to Search (powered by Lens) taking off, with both them and Samsung promoting it, including Samsung’s recent London Tube activation.
Source: Samsung
This would seem to be part of Google partnering with Samsung and presenting a more united Android front against Apple in the mobile market.
Will need to see if this strikes a chord with the wider public, and if less tech-savvy users see the value in using Circle to Search or their camera above traditional text search.
There are slight echoes of the voice search hype back in 2018, when interest in voice assistants was at a peak. This explosive growth plateaued, and while voice search in still very much in the mix, it wasn’t a moment that fundamentally changed the landscape such as the move to mobile search.
Performance Max Reporting
One smaller (but very well received by the crowd) announcement was improved performance max reporting, giving us asset level conversion insight.
This goes somewhat against the grain, as advertisers we’re used to getting fewer reporting options from Google Ads. However increased transparency on what is working and what isn’t is always welcome!
Generative AI & Advertiser Control
We heard a lot about additional control coming to the Generative AI products, especially on the image generation side.
This is a good move as it addresses a big challenge, most advertisers will have a bank of existing creative and will need to stick with brand look and feel. This wasn’t realistic just with prompt engineering, and coupled with the mixed quality of the output meant low uptake.
One particularly promising product announcement was Generate More Like This, which will allow generative AI to take a existing image as a seed, and create images with a similar look:
In a similar vein, Reference Images will allow us to visually define the sort of output we’re looking for - again helping to make sure any assets generated fit in with the brand look and feel.
The example shared below turned out well, but this art style is quite simple and forgiving, helping hide any imperfections:
These two options lower the barrier to entry when it comes to generative AI imagery in the platform, and more incentive to quickly give it a go. Rather than sitting down and trying to create a suitable prompt without any guarantee of getting anything usable at the end, it becomes much simpler to ask Google Ads for “more like this” from your current creative and see what you get, or provide a reference image if going the prompt route.
The final barrier is of course quality - imperfections are inherent to AI generated art, and a non-starter for some brands because of this.
Closing Out
Whilst we didn’t get as many big ticket announcements as in 2023 (i.e. generative AI tools such as conversational campaign creation or performance max image generation), there were plenty of product iterations peppered in with new announcements.
Much like 2023, one takeaway continues to be that AI is not a marketer, but a tool. The technology can do a lot and is a fantastic supplement when amplifying, rather than replacing, human ability.