It’s astonishing that only a year has passed since ChatGPT burst onto the scene, captivating the world with its remarkably adept chatbot capable of generating human-like text on command.
As users transitioned from amusing experiments like having it craft jokes or write in the style of Shakespeare, more practical applications quickly emerged.
Iryna Imago/Shutterstock
Teachers, for instance, discovered innovative ways to utilize large language models to convey complex subjects to their students. In Hollywood, writers renegotiated contracts as generative AI could craft new plot twists and storylines based on their ideas. Even software companies recognized the potential, of incorporating genAI into their tools to simplify mundane tasks.
For PPC (Pay-Per-Click) marketers, ChatGPT unlocked transformative possibilities. Let’s look back at ChatGPT’s inaugural year and explore what we can anticipate from this groundbreaking technology in the future.
Reflecting on ChatGPT’s Debut
ChatGPT was introduced on November 30, 2022, offering free access to what seemed like an impossibly intelligent AI chatbot. Powered by OpenAI’s GPT-3.5 language model, it demonstrated an unprecedented ability to comprehend context, follow logical reasoning, and generate coherent, convincing text. Almost overnight, people began using it to compose emails, blog posts, code, and ad copy. The excitement grew as users realized that ChatGPT was not a narrow AI tool but the foundation for the next wave of computing, with far-reaching implications across numerous industries.
Various industries have undergone significant transformations since introducing ChatGPT and other genAI models a year ago. In our industry, generative AI has streamlined the content production process, from blog posts and social media updates to landing pages, keywords, and ad copy. Its capacity to rapidly produce high-quality text has proven invaluable to marketers, saving time on project ideation and execution.
Google took a step further in May by introducing a generative AI campaign creator. Such announcements posed a crucial question on the minds of digital marketers for years: How will our roles evolve as AI continues to perform tasks that were once done manually?
To address this question, let’s first explore how ChatGPT has influenced search marketers.
Early Applications of GPT-3.5 in PPC
Naturally, PPC experts were eager to harness the capabilities of ChatGPT, even in its initial form. They found numerous uses, including generating keyword ideas from seed keywords or landing page URLs, organizing extensive keyword lists into coherent ad groups, crafting compelling ad headlines and descriptions, offering suggestions for landing page optimization, and evaluating the relevance of search queries to improve ad targeting.
To streamline some of these tasks on a larger scale, I began using GPT extensions for Google Chrome. Additionally, I extensively explored the OpenAI API Playground, where various settings affecting performance could be fine-tuned. For instance, the temperature setting could be adjusted to make the system more or less creative, a critical consideration in regulated industries where excessively creative ad text ideas might violate regulations.
These early applications showcased the immense potential of ChatGPT in PPC account management and campaign optimization. However, they were just a glimpse of what would soon become possible as AI continued to advance.
GPT-4: A More Advanced yet Costly Writer
In March 2023, OpenAI introduced GPT-4, which brought significant improvements. This model was trained on a more extensive set of input parameters, 25 times larger than Meta’s Llama-2 model and five times larger than Google’s PaLM-2 model. Consequently, it produced higher-quality writing and enhanced mathematical abilities, a crucial factor in PPC, where optimization decisions rely on accurate mathematical and statistical analysis.
While earlier versions of ChatGPT handled basic PPC tasks like keyword generation adequately, GPT-4 proved exceptionally valuable for generating long-form text, such as email campaigns, blog posts, and landing pages. However, it came with the drawback of being slower and more expensive than previous models. Nonetheless, as a wider range of models became available, it opened the door for optimization.
Users could evaluate each model’s strengths and weaknesses and determine which language model (LLM) and version best suited their specific tasks. GPT-3.5, for instance, proved just as adept as GPT-4 in suggesting keywords, making it a more cost-effective and quicker choice for keyword-related tasks. On the other hand, GPT-4 excelled in generating written content, justifying the additional cost and latency when crafting landing page copy or client reports.
ChatGPT Introduces Beta Features
In July 2023, OpenAI rolled out new beta features for ChatGPT, including the Code Interpreter, later renamed Advanced Data Analysis. These features introduced entirely new possibilities for PPC.
The Advanced Data Analysis feature empowered GPT-4 to write and execute Python code. By uploading campaign performance files directly into ChatGPT prompts, advertisers could leverage it to perform statistical analyses and create graphs, thus gaining fresh insights and optimization opportunities.
Another significant upgrade for ChatGPT was the support for plugins. These special add-ons greatly expanded its capabilities, enabling ChatGPT to connect directly with external datasets, documents, and live websites rather than relying solely on its indexed knowledge. This meant that advertisers could feed ChatGPT raw .csv files of PPC campaign statistics for advanced analysis, upload entire catalogs of product descriptions to enhance copywriting relevance, or have ChatGPT scrape and summarize landing pages in real-time during optimization efforts.
Generative AI as an Assistant, Not a Replacement
While it’s easy to envision generative AI completely automating digital marketing tasks, in practice, these tools work best as assistants that enhance human capabilities rather than replacing them. As I explored back in 2017, well before the advent of generative AI, humans play three essential roles in digital marketing:
1. The Teacher: This role involves curating data and fine-tuning models like ChatGPT to align with an organization’s objectives and industry context.
2. The Doctor: The doctor prescribes solutions, such as selecting the right LLM for specific business challenges.
3. The Pilot: The pilot carefully monitors performance and intervenes when generative AI deviates from the intended course.
Even as AI has transitioned from mere pattern recognition to content generation, I believe these roles remain applicable. The teacher, doctor, and pilot work together in harmony. For example, an account doctor can allocate the time previously spent on copywriting to data analysis and consultation. They can also prompt ChatGPT to summarize crucial account insights in client-ready reports. The AI generates content in this partnership while the human provides creative direction and quality control.
Therefore, generative AI should not be viewed as a threat but as a collaborative assistant that amplifies human strengths. Marketing teams embracing this human-AI partnership will gain a competitive edge, although it is vital to recognize that the future remains human, with AI serving as a helpful tool.
What’s Next: OpenAI Dev Day Announcements
At OpenAI’s DevDay on November 6, 2023, the company unveiled significant upgrades that offer a glimpse into the future potential for PPC professionals.
Most notably, OpenAI launched GPT-4 Turbo, which supports much longer prompts of up to 128,000 characters. This substantial expansion of the context window enables ChatGPT to handle more advanced PPC analysis and optimization tasks. Users can now submit the equivalent of approximately 300 pages of text, providing in-context learning and
