By Scott Hendler, Paid Search Manager
——
As marketers, it is crucial to stay vigilant to the ever changing landscape, which has never been more true with the roll out of AI. It’s no secret that AI is going to change the game for Search, but how can you be prepared and readily adapt? One of the greatest impacts it will have and has already made, is on search engines, where results are matching more based on the intent of your search, rather than being dependent on the keywords itself.
This change can be traced all the way back to 2013, when Google rolled out its “Hummingbird” update. This was Google’s first step in updating its search functionality from strictly keyword based to leaning into the users’ intent behind the search. Using their AIs Natural Language Processing model, user location and intent, they are able to provide more accurate search results, making Google Search more conversational.
There are countless ways AI is going to impact Search on the consumer side & the marketers/brand side, but let’s stick to the biggest shake ups we can expect.
For some background, the AI technology landscape generally consists of four categories:
- Analytical – Machine learning, Computer vision, Agent-based modeling
- Conversational – Natural Language Processing (NLP), Conversational Systems
- Generative – Pre Trained Transformer (GPT), Generative Adversarial Network (GAN)
- Hardware – Virtual/augmented reality, Edge Computing, Internet of things (IoT)
From the user searching perspective, Conversational & Generative AI are likely to make the greatest impact to the way people search and the results provided.
Analytical AI Impacts
With the integration of AI into search engines, Machine learning models are delivering much more personalized search engine results. Machine learning models are better able to understand the users search intent and context to deliver the most relevant results at the top of the page. These models leverage behavior, preference, and historical interactions to provide the most accurate results, leaning away from strictly keyword match based, leading to improved relevance of the search results.
Conversational AI Impacts:
AI’s Natural Language Processing (NLP) has turned searching more into a conversational query, rather than robotic, overhauling how users search now. Conversational AI is making search more intuitive, efficient, and personalized leveraging the following enhancements:
- Semantic Query Understanding: Gives search engines the ability to understand the meaning, context and intent behind the keywords, rather than simply matching to the keyword.
- Voice Search: Although voice search has been around for quite some time now, NLP enhancements have led to users leveraging it more often, naturally leading to more conversational search queries.
- Answering Question: NLP powered search engines are able to understand and answer questions directly at the top of the SERP (search engine results page), rather than having to click through various articles to find what the user is looking for.
With the release of Virtual assistants (Siri, Alexa, Google Assistant), users may not even realize it, but they are becoming increasingly familiar and comfortable with conversational search applications. This is likely to extend into search engines, with users becoming more comfortable leveraging the voice search feature, which inherently turns the search more conversational.
To account for the changes in the way people are searching, marketers should have already begun optimizing campaigns for voice search. Keyword lists should be updated to reflect more conversational queries, including long-tail keywords, more conversational phrases and question-based queries.
Generative AI Impacts
Generative AI has been the most popular in society as of late, with OpenAI’s release of ChatGPT & Google’s Gemini. With the ability of these technologies to provide large amounts of content, including articles, product descriptions, along with other very specific results, it is conceivable that GPTs will replace search engines altogether in the future. The ability that these technologies have around content summarization, results in a more seamless and efficient user experience as users receive the information requested immediately and without having to read through multiple articles or links on a standard search engine.
In May of 2024, Google released its newest Gemini model customized to enhance search. One of its features is the “AI Overview” which has enhanced the ability of Google to provide a quick overview of the topic searched, brief summaries, and direct links to learn more.
The AI Overview is powered by Large Language Models, which use deep learning to understand the query and generate a response. In simple terms, it has the ability to understand what you are asking and generate the answer right on the spot, negating the need to scroll the Google SERP. The LLMs are trained on extensive amounts of data, being self sufficient, they are able to continuously learn as they collect more data, increasing its ability to generate detailed answers at the top of the page, including product recommendations. This requires an entirely new way of thinking for setting up your marketing campaigns for success. Since it’s an emerging change, there are no tools currently available to test if your brand is being featured in the AI Overview when someone is making a relevant query.
In the current state of the AI overview, you can manually perform searches yourself related to a product your brand sells and evaluate if you are being featured in the AI Overview. If Google does not feature your product in the AI overview, it’s possible the consumer will never reach your landing page. The LLM results operate completely differently from how historical search results are generated, which means marketers are going to have to make a major pivot, particularly in the SEO department. The LLM model pulls the content from a variety of sources, text, images, videos, reviews, site content, and user-generated content, making it a necessity to ensure all of these sources are up to date, with relevant information. A recent study performed by Harvard analyzed the impact of integrating strategic text sequences that align with often used phrases, such as affordable or “low-priced” to a product page. Throughout the test (using a particular coffee maker), the baseline page was almost never featured, but when the strategic text sequence was introduced, it became one of the most frequently recommended products. The need for marketers to ensure all sources are properly updated with these strategic text sequences is crucial to success in the changing landscape of the Google SERP.
Hardware AI Impacts
Consists of specialized hardware such as graphics processing units (GPUs), and tensor processing units which make a significant impact on the efficiency of search engine results. This advanced hardware leads to:
- Accelerated Processing: designed for the use of AI algorithms, including NLP and deep learning models, accelerating the processing time of search queries and learning models, leading to faster search results and improved user experience.
- Real-Time Processing: Enables search engines to process search queries and data in real time creating the instant search suggestions, the autocomplete feature and suggested search corrections.
How to Prepare & Adapt as Marketers
Now that we have a basic understanding of how AI is impacting the search landscape, how can we, as marketers, keep up with the pace? AI will force marketers to adapt in real time or fall behind the competition. It has been predicted that by 2026 organic search engine volume will drop by 50% as a result of chatbot applications. Even as of now, 60% of all searches result in zero clicks, as AI-generated answers reduce the need for users to click through the traditional search results. We need to lean into the ever changing landscape and begin to learn and test these new AI powered tools. Marketers should also consider familiarizing themselves and develop generative AI skills before the competition to remain current and ahead of everyone else.
To start developing these skills, you should start by gaining a fundamental understanding in the various areas of machine learning, deep learning, and generative modeling techniques. By understanding the basics of machine learning (supervised, unsupervised, and reinforced learning) you can begin to form the foundation for understanding more advanced concepts like Deep Learning fundamentals and Generative Models and be able to use this technology to your advantage. At a minimum you should start experimenting with AI applications, like ChatGPT, to start understanding how it works, what you can ask, and what you can expect as a response.
Test and Learn with AI-Powered Targeting
There is virtually no reason why you should still use manual bid strategies and update CPC bids for your keywords. With the current tools at our disposal, you should be using some type of smart-bidding available. Google’s AI smart-bidding, such as “maximize conversions” or” Target ROAS” bid strategies are far more efficient in adjusting bids in real time to increase efficiencies, where a human making the manual bid adjustments would not be able to keep up. Not only does leveraging Google’s AI smart bidding lead to greater efficiency, it saves the user time which can be put to better use evaluating strategy and tactics for the brand as a whole.
Test the Current AI Tools available in Platform
Campaign types such as Performance Max, which uses AI driven bidding models, provides a simple way to start familiarizing yourself with the technology and beginning to learn how to best use the features it provides. Pmax combines multiple of the AI categories listed above, including Generative AI, where we have the ability to let Performance Max automatically generate different variations of the static and video assets for better optimization capabilities.
Test & Learn with Generative AI
- Content Generation: Generative AI gives marketers the ability to generate a vast amount of different creative assets in a very short time. You can start to test these different asset types and determine which works the best.
- Ad Copy & Keyword Generation/Optimization: Train a generative model, like a GPT on data collected from existing keywords & tune the model to improve the keyword set and generate new keywords and copy. Use the trained model to generate new keywords based on a seed list and target themes. These models are able to produce synonyms and related terms that could be relevant to your brand, filling in your keyword blind spots.
Analytical AI for Marketers
The days of manually pulling and analyzing data yourself are numbered. Analytical AI tools can process large amounts of data rather quickly and provide clear concise results indicating what optimizations should be made. While these AI analytical tools can be extremely helpful in analyzing vast amounts of data quickly, it is still crucial to perform data checks and QA the findings to ensure they are accurate. Similar to when you ask a Chatbot to write an email for you, there are minor edits that are typically needed as the AI is not perfect just yet.
- Predictive Analytics, which is a data science tool that uses historical data, statistical algorithms, and machine learning techniques to predict future trends or outcomes based on patterns found in the historical data. These models can take in vast amounts of data and provide forecasted performance for your campaigns based on the historical data and market trends. You can then use these models to better allocate funding throughout the calendar year and make the necessary changes, whether you should pull back or push more aggressively at any point through the campaign’s flight.
- Optimize Existing Campaigns: These models give us the ability to make data driven decisions, optimize campaign performance better than ever, and help drive results above your businesses goals.
As AI transforms the digital & search landscape, agencies must embrace the opportunities presented to create more personalized, relevant, and engaging content for their target audience. By staying current with the tools readily available and leaning into the AI revolution, marketers can better navigate the ever changing landscape and use these new tools to drive stronger results for the brands they represent. And while it’s easy to become reliant on the different AI applications available, remember that the technology has yet to be perfected. The ones who get it right will be the ones who are able to marry AI with the human element – working together to embrace and revolutionize this new marketing era.
