The future of Google Search & Ads in the age of Gen AI

Matz Lukmani:

Hi, my name is Matz Lukmani and I'm a Product Lead at Google, focusing on Performance Max and attribution. In our session today, I'll focus on the future of Google search and ads with Gen AI. So let's get started.

At Google, AI is nothing new. We started this journey almost a decade back with some of our back then latest technology, features like Transformers,

TensorFlow, BERT, AlphaFold to what now is the latest with Bard and PaLM 2, which are essentially large language learning models. We've introduced AI based features in a subtle way with updates like Smart Compose in Google docs and live traffic updates and maps. You can think of Transformers, for example, one of our premier technologies as the grand parent of all the models being used today. In fact, when you see capital T in an acronym from something in the AI space, it's a safe bet that T stands for transformers.

These vastly improved language models combined with better data and better computing house power have paved the way for the leaps we are making now. But the best is yet to come.

While we make some of these changes, there are certain principles we've developed and learned, which I would love to share with you. To start us off, for example, at Google, we have up to 15 products that are reaching more than half a billion users. The top six of them reach more than 2 billion users. This is a big responsibility, and a similar responsibility is with you when you're reaching your users while building AI based solutions.

It's really critical to keep in mind the impact it has on your users, as well as business partners, and bring changes in a responsible fashion. One of the examples, which is really close to my heart is with Google Maps. Within Google Maps, we've used analytical and predictive AI to optimize the world around us.

If you used Google Maps, you may have noticed at times when you're plotting between point A and point B, you see paths which are slightly bit more eco-friendly, but may take a little longer to get there. We've seen a very strong usage of these eco-friendly paths, which has helped us save up to a million tons of CO2 emission per year. That's equal to removing 200,000 cars off the road. That's the level of responsible impact that we are very proud of having in this world.

Similarly, this applies to all businesses using AI to drive better outcomes.  Now, the technologies we mentioned for maps, large learning models, also apply to Google search just the same way as the technology is used for products like AlphaGo.

Now in Google search, what you should expect are some of these following changes that I'll walk you through. For example, search is becoming more visual, and you may have noticed that already in your UIs. For many years, Google has mastered the art of looking at textual data across the world and organizing it and making it meaningful and accessible for our users.

The next wave of change will evolve around visual searches and blending the dimensions of text and visual experiences. For example, you may have already noticed multi-search experiences, which allow you to take a picture of a physical entity and search for information about it online. These features have slowly become the most popular features used by our eCommerce partners.

Similarly, we've introduced something called the Immersive Views in Google Maps, which allows you to get a bird's eye view of your entire journey before you start, from whether it's on foot, bike, or in the car, and have started using certain very subtle features like AI-powered weather predictions. These features are bound to stay and are here to be improved over time.

Now, while I talk about maps and visual experiences within search, it's also important to acknowledge that we will continue to supercharge search using generative AI functionalities to help you unlock entirely new types of questions you never thought search could answer and transform the way you can search for information and organize it for your daily purposes.

These new experiences will make search smarter and searching simpler while making sure ads are clearly marked, labeled, and blended within these experiences.  Now, we spoke about search and the way our user products are changing, but also from a marketing standpoint, it's really critical to acknowledge the rise of the use of AI based technologies.

For example, a study done by the McKinsey group. They noticed AI based solutions usage has increased up to 2.5x in just the last five years. 70 percent of businesses that are using AI are seeing increased revenue and 30 percent of adopters are seeing a reduction in cost. So it's not just a revenue driver. It's also a cost reducer.

Now, while AI is a technology and is really reliant on automation at its core, your input is a critical piece of the component structure. For example, a study done by BCG group saw better advanced technologies driving up to 20 percent advanced performance or marketers. But advertisers who added additional human input, whether it's better creatives or better input data from your first party data sources to optimize performance to constantly managing and optimizing the output of these campaigns. So up to 15 percent better performance hence your input and AI technologies together on an average drive up to 35 percent better performance versus those managing businesses in a manual fashion.

Hence, it's critical to acknowledge that you and the technology that you use are in essentially a flywheel where AI learns from you and you learn from AI to improve your business performance at Google. That flywheel is really well developed in some of our products. For example, you supply creatives goals and conversion data, which AI then mixes and matches to create creative experiences that lead to better conversions and as AI costs a wider net beyond just traditional media, it also delivers returns as well as insight that can help you fine tune again, your creatives, your goals and conversion data.

Hence the flywheel is, is really critical to constantly learning from AI as well as powering it to drive better outcomes. Now, at Google, we've tried to build the best of our AI solutions for advertising into our newest and latest marketing campaign type called Performance Max.

The idea of Google's Performance Max, when we launched it up to two years back, was to bring the best of Google ROI in a single campaign, allowing you to go across all of Google's services for advertising under the shell of one campaign type. Over the year, we've seen better and better results coming out of advertisers using Performance Max campaigns.

For example, last year, based on a study done at Google, we noticed advertisers that used Performance Max campaign growth up to 13 percent more conversions at a similar cost. A similar study this year noticed advertisers were seeing up to 18 percent better results. This improvement is based on AI based technologies that have been developed and launched in our product this year, as well as new inventory that PMAX can operate and unlock for your business.

Now, performance improves with using usage of AI. But a critical input is your creative when it comes to marketing. For example, a study done by Nielsen proved that advertisers’ overall performance impacts have more than half its bearing on creatives that they use. It's not the media. It's not the brand. It's the creative that drives up to half of the performance ROI that you see from your campaigns.

Hence your creative inputs are critical while you need to improve on adding better creatives and driving better outcomes. We are also committed to making sure we use your creatives in a way that it gets published correctly on all service. For example, we launched a feature called Flip Video which allows us to take your horizontal videos, and using vision AI technologies, render it for vertical and square format on all our services. At Google, looking at the performance from users of Performance Max campaigns, we noticed advertisers that upload their own custom video instead of using auto generated videos, so up to 12 percent incremental conversions coming from their campaigns.

Hence, your creative is really critical in driving outcomes, especially in a world where you're using AI powered solutions. Now, we are committed to bringing better solutions that can help you improve the creative appeal of your assets. And I'll walk you through some features we are launching that can help you envision that even further.

The first example is Product Studio. If you are a retailer, you may be using Google's Product Studio to help upload and create versions of your product assets with the new generated AI features that are planned to launch in this  Google's Product Studio. You'll be able to edit backgrounds and improve the resolution of the images, your product images that you use for advertising.

It'll also allow you to have a conversation with the UI in terms of fine tuning the background. Relevant to the ad campaign you're running, hence you can reuse some of your stock images and make it more appealing and more timely. Additionally, we will go beyond just visual and video data to also allow you to create better headline text and body content for your ads.

For example, our plan is to bring in conversational experiences within Google Ads where you can have a conversation about the right variation of headline text and description within your ad copy and ad groups.  Now our aim is to make the process as seamless as possible. And with features like improved asset creation flow, which is planned again to launch in the US this year and to wider Europe markets next year, we aim to allow you to improve your asset diversity by uploading a stock image or an existing image and going through an AI powered prompt to create variations of these assets for your business.

Now, while we are allowing and building new features that allow you to switch and create better creatives, we also are focusing on making as many of these changes automatic. For example, by the launch of automatically created assets, we automatically build relevant ads based on your campaign data and serve them to the right users at the right times, matching the query that's been searched online. These features are prevalent in products like Performance Max, which are already powered by using the best of Google's AI.

Now, features are great and creators are really important. What's super critical is you picking the right part now, especially when it comes to enterprise clients. I want to take the example of Versace, which is one of our largest fashion retail brands working in the retail space. Versace tried Performance Max campaign and upgraded some of their shopping campaigns to Performance Max, and they plan to scale this using SA360 as a foundational platform, partnering closely with their agency of record. Here in this case, Versace was able to build Performance Max campaigns and optimize it to floodlight conversions. That's a source of conversions used for SA360.

By using performance Max campaigns, Versace was not only just able to drive better performance in terms of 20% increased transactions, 5 percent increase ROAS and 18 percent increase conversion rates, but were able to do all of this while maintaining a strong grip on their brand image and the customer reach. This is a really good example of an enterprise client using AI based solutions and partnering closely with their trusted agency.

Now, your solutions are great and creatives are really critical. Partners are super useful, but the best input you can give to an AI based solution is training data, in this case for marketing campaigns, it's really critical to power your Performance Max campaigns or any AI based campaigns using your first party data, which is much more durable than any third party data set.

One of the features which we are proud to build out, and we've seen good results from advertisers using this is campaign types built specifically to drive new customer acquisition. And in some cases, high value, new customer acquisition using your customer match data, which is essentially your first party data around your new customers.

I want to give an example of one of the clients that has used this feature to drive good results, hoping to inspire some of you to do something similar. One of these, this example comes from a Dutch bike company, Swapfiets, which has a subscription-based business. The biggest challenge was to continue to grow its new customer base by combining new customer acquisition goals with its customer match data and smart bidding in Performance Max, Swapfiets saw a 36 percent increase in new customer transactions as a result of this test.

This new customer acquisition campaign for PMAX is now part of Swapfiets’ always on campaigns across all of their markets. Hence, it's really, really critical for you as an advertiser to use your first party data to power your advertising, especially when it's using AI powered solutions.

Now we talked a lot about using input data creatives, picking the right partners to drive AI in the right direction, it's also really critical to acknowledge that AI can help you extrapolate new data and insights, especially in the world where your third party data sets is forever shrinking. For example, with enhanced conversions and Google, Google AI can take first party conversions, which is tracked by your platforms and help model that data to improve your overall conversion measurement in a privacy resulting fashion.

These technologies are going to be prevalent in the future and will become foundational for any AI based solutions to flourish.

Now, while we focus on bringing in better quality data, it's also really critical to use the insights that you can get from your aggregated first party data. And drive actions to improve performance with the launch of PMAX campaigns in the last year and a half, we've slowly launched up to 10 different insights within the insights module of Google ads, where you can see aggregated data that you can use to drive insights and actions.

Some of them are fairly simple, like the search terms insight that lets you see which search terms and keywords are driving actions performance within your campaigns. Others are much more intuitive as an asset audience report or performance shift insights report that help you understand which assets are performing best, which Google audiences they're resonating most with, or what could be the cause of a change in performance shift in your campaigns.

We’ve consolidated all of these 10 insights and actions tied to these insights in a quick playbook. Please take a, take some time to read through these insights and insights playbook actions to drive better performance with your BMAX campaigns.

Now, we talked a lot about insights and creatives and partners while you do all of this and see good proof of action, it’s critical to scale some of these across your entire markets and businesses. With the launch of Duet AI in the core studios, we planned for you to be able to do just that.

Our aim is to have fast and simple conversations with your data going beyond just questions and answers to drive actions informed by your unique businesses, give you the ability to create entire reports or advanced visualizations with only a few sentences of instructions, greatly saving time and minimizing the need for technical expertise.

We also aim to automatically create brand new Google Sheets presentations from your reports with automatically generated summaries, and then use natural language to create formulas for calculated fields that you can then use to transform information from different data sources, as well as use formulas that can be transformed in a visual fashion.

Finally, we aim to easily generate Looker ML code using natural language. All of these features are being packaged within Duet AI and Looker Studio integrations. And with the launch of Looker Studio pro in experimental format, this year, you'll be able to utilize this at scale.

At this point, I would love to take a quick pause and mention Supermetrics is one of our key partners and are premier when it comes to scaling these actionable reports and data inputs for their clients. So if you're curious about getting insights to go and act at scale for your business, please reach out to them and look at some of these partnerships as a potential for next year.

Now, to summarize my entire presentation, let's go through some of the key elements. Firstly, at Google, 80 percent of advertisers are using AI powered ad solutions like Google Performance Max and Smart Bidding.

And overall, we've seen similar trends, hence now is the best time to embrace the power of AI. It's important to be responsible while making and leveraging these technologies, and also super critical to pick the right partner that can help you scale these technologies across your entire business.  Some of the advances made in generative AI can help you improve your creative diversity and in campaigns, which has a direct impact on better performance or AI powered solutions.

The best input is your first party data that can help train and fine tune your models. Finally, you look at data, but with the eye of generating insights that can drive actions, especially in the future where you'll have less and less raw data to act on. And finally, when you're ready with your proof of concept, invest in platforms that can help you scale this across your entire enterprise.

With that said, I'll end with my last note, which is this could be a fairly complex and uncomfortable time for a lot of advertisers. And like you, I also sometimes feel that I'm competing against AI. At those points, it's important for you to acknowledge that AI has the potential to level the playing field for all businesses of all sizes, not just for now, but also most likely for the foreseeable future.

Hence, you're not really competing against the AI, you're competing with marketers using AI and the one that uses their inputs best and tunes the model in the right direction drives the best output. With that said, I'll end my session. I hope you found this useful. Please reach out if you have more questions and hope you enjoy the rest of the seminar today. Take care.