AI Technology and the Impact to Paid Search
Google has dominated paid search since the early 2010s, but Microsoft has shaken things up with ChatGPT (and Google has countered with Bard). These AI tools will change how users find information, and impact the level of engagement advertisers can expect from users. Let’s look at what these technologies are, where they’re going, and what advertising spaces they’ll disrupt.
ChatGPT is a product of Microsoft’s $10 billion investment with OpenAI, an American artificial intelligence research lab. Google followed with Bard, powered by LaMDA, Google’s proprietary conversation technology.
From Clippy to ChatGPT
(Remember “Clippy,” the paperclip chatbot in Microsoft Office?) Chatbots are several decades old, but their scripts have been improved with the advent of machine learning, allowing the tools to make more “informed” decisions as their response databases grow. Even with these advances, though, these programs do have limitations. Google confirmed in their press release for the launch of Bard that “while it is exciting technology, they’re not without their faults” and that “Bard is an experiment and will not always get it right.” And the ChatGPT FAQ clearly acknowledges that it may provide false or outdated information.
How Does This Impact Paid Search?
AI in paid search isn’t new: for years it’s been used to help with bid strategies and, more recently, in responsive and dynamic search ads. But then, AI helped advertisers; now, it’s become a consumer-facing tool that can completely reshape the user experience of search.
How a user searches for information
Currently, when a user searches for information on a search engine, they land on a results page and get a list of pages deemed relevant. From there, a user must find answers for themselves. But with generative AI, a user simply asks a question and gets an answer.
What makes this interaction unique is that the detailed response immediately translates to a more informed user (if, of course, the response is accurate — in testing to date, we did find that the information we got was of adequate quality). The user can move much more quickly than before from “upper funnel awareness” to “brand consideration” (if, of course, a brand’s content is part of the database of information). We anticipate that this will increasingly be how patients, in particular, research.
We expect that as these tools are accepted and adopted, KPI expectations will need to shift — in particular, cost per click and site visits. Since users will be getting information directly, we expect to see a change in the way users visit a site. Some users will still find the website via traditional search engines, while others may visit via a chatbot.
That said, users who continue to look for more information will be more engaged. Once they land on the website, it is likely they will engage even more than before. Therefore, we expect each user to be worth more per click – effectively increasing overall cost per click. While it’s too early to know the full extent of how performance metrics will change, we’re closely watching campaign performance across the board.
Accuracy & Outdated Information
AI doesn’t get everything right. We can’t stress this enough. AI can’t contextualize and comprehend information like humans – although they can do a fairly good imitation. It’s important to realize that this gives opportunity for misinformation to be spread. It’s also important to realize that new information — for example, information about newly launched brands — will take time before it’s included in AI datasets and, therefore, in its answers.
Ad Placement + Rules and Testing
Paid advertising will be available in the chatbot landscape. Microsoft Ads is already beginning to test ads in ChatGPT and has opened it to all advertisers. For pharma advertisers to serve in the chat, a campaign disclaimer must be created to ensure that brands have the necessary ISI to remain compliant in the new AI experience.
We anticipate Google will also make ad placements available within Bard. Bard may look like ChatGPT or it could even be paired with the existing SERP to augment user search results.
Obviously, there are still many unknowns. We continue to monitor changes, discuss updates with our search partners, and listen to the marketplace for any additional insights.
Paid search and the digital landscape are ever-changing. While it’s too early to know the full impact, AI is certainly disrupting the way users and advertisers approach paid search. Overall, we believe the impact will be a net positive for consumers and advertisers.
As of April 2023, advertisers are in an “observation and testing” stage as this begins to become available for paid advertising opportunities. At this time, we do not recommend testing pharmaceutical paid advertising until Microsoft and Google can guarantee that our legal advertising regulations can be met. Our team will remain in the observation stage for now and continue opting out of serving within the AI platforms, but we are excited for the new opportunities that will become available for our clients soon. As always, we are continuously monitoring account performance from both a pre- and post- click perspective to ensure that we can quickly communicate any variance to historical performance with our clients.