By Otavio Dias on 4 Aug 2023
“The Law of Uneven AI Distribution will determine how much, and how quickly, you and your company benefit from AI.”
The Law of Uneven AI Distribution states the value you gain from AI and how quickly and consistently that value is realised, is directly proportional to your understanding of, access to and acceptance of the technology.
At Wolfgang, we never fight or deny change in our industry. We look for the opportunity. Whether it’s AI, the phase out of first-party cookies or any of the countless changes the ‘Tech Giants’ decide to roll out, we own our responsibility to our clients to explore the opportunities for business growth that lie underneath the turbulence. We encourage clients to come with us and enjoy the first-mover advantage.
Can we predict exactly how and to what extent ChatGPT and other AI technology is going to change the digital landscape? No. And if anyone says they can, they’re lying or deluded.
But, to ensure we as an agency and our clients come out on the right side of the Law of Uneven AI Distribution, we’ve been developing our understanding and comprehension of AI, attending industry webinars and conferences, running internal education sessions and testing and evaluating potential use cases across all channels.
The more we understand the underlying ideas behind the technology, the better we can understand its strengths and weaknesses and the better we can identify creative and powerful use cases to beat the competition.
What is AI and ChatGPT?
GPT-3.5 and GPT-4 are types of large language models (LLMs).
While they might appear ‘intelligent’, LLMs do not KNOW anything. The way they work is they are trained on an enormous set of training data. When faced with a query or a prompt, LLMs respond by predicting words that go together based on its training data. So you can see where their propensity for high-profile mistakes comes from (isn’t that right, Google?).
ChatGPT is the chatbot that is built on top of GPT-3.5 and GPT-4. The free application has made interacting with the LLMs accessible to the wider public.
At the end of the day though, when you cut through all the hype, AI is just smarter technology - a natural evolution that you, as a digital marketer are well used to.
How Can We Use ChatGPT to Get Ahead in our Digital Marketing?
This has been one of the most frequent questions we’ve received from clients this year. The use cases are obviously endless so in this article, we will focus only on how we’ve been using AI specifically in relation to doing better Paid Search.
5 Ways We’ve Used AI To Deliver Better and Faster Results for Our Paid Search Clients
Ad Copy Optimisation
Responsive Search Ads, Performance Max and Responsive Display Ads all require the marketer to feed in multiple headlines and descriptions for the target search topic. Google's machine learning algorithms then test various combinations of headlines and descriptions to determine the best-performing ones for different search queries and user contexts. This not only speeds up the process of adding multiple relevant headlines and descriptions but also enables us to conduct more A/B ad copy tests that yield better results.
For example, when we conducted an A/B test comparing our ad copy against ad copy augmented with insights derived from AI over a period of 30 days. The results showed a staggering 164% increase in conversions, a 64% reduction in cost per acquisition, and a 31% higher conversion rate for the ad copy enriched with AI insights.
With the growing use of AI in our workflow, our team has conceived and developed our own AI tool named "Wolfgang GPT - PPC Robot Helper." This tool not only provides valuable insights for ad copy generation but also excels in generating highly effective exact-match keywords and identifying negative keywords. As a result, our efficiency in leveraging AI is further accelerated.
By feeding AI details about the search topic, we can derive a comprehensive list of relevant, high-traffic keywords, facilitating more effective keyword research. By clustering related keywords into distinct themes and quality, the process becomes more efficient and leads to multiple relevant suggestions for each cluster.
Negative Term Analysis
The extent of the search term report may vary based on the account's structure, potentially requiring over 1 hour to analyse all the search terms for the exclusion from each ad group. However, by uploading the search term report to AI along with the business industry sector, a comprehensive table is generated. This table categorises the search terms based on their distinct quality and determines whether they should be excluded or included. As a result, the time required for analysis is reduced to approximately 30 minutes.
Google Shopping Product Title Optimisation
When managing a client’s shopping feed, we not only have a platform that allows us to optimise the product attributions, but it’s also an AI-integrated dashboard that provides suggestions for the titles and descriptions that are hugely valuable for e-commerce retailers.
Modifying Google Ads Scripts
A quick and practical example is modifying an existing script that excludes a display placement by domain (.viral, .xyz, .pl) or a script that excludes irrelevant URLs based on certain words contained within; (e.g. sites obviously targeting children). Given the vast number of websites on the Google Display Network (over 2 million), excluding placements can be a daunting task for a human to perform alone. With the help of AI, we have turned a 30-minute daily task into a <1 minute daily task and can ensure that your budget is used even more effectively.
When AI isn’t the Solution
AI is a formidable tool for streamlining processes, but it doesn't serve every purpose. Certain areas still require the human touch and judgement.
We tailor every campaign to our clients’ specific objectives and strategies. This process requires a deep understanding of their business, USPs, target audience, as well as the ability to make strategic decisions based on this information.
As such, this type of decision-making involves creativity and critical thinking, which is not replicated by AI. Additionally, as a Google Premier Partner, we enjoy access to the very latest betas and test and apply these new innovations to ensure that our clients are always at the forefront of the latest trends and innovations.
In conclusion, it's crucial to bear in mind that the knowledge cutoff for ChatGPT was September 2021. Given the breakneck speed of digital marketing advancements, two-year-old 'knowledge' simply will not suffice in our dynamic field.
So, let us learn, adapt, and find innovative ways to leverage AI, while remembering the human touch that makes our strategies truly effective.