The launch of Chat GPT by OpenAI in November 2022 marked a radical change in the internet search engine market. In this post, we’ll discuss how it has revolutionized the search market and how to measure traffic to our website from AI-based search engines, such as Chat GPT or Gemini.
The Evolution of the Search Market Since the Launch of Chat GPT
Before the launch of Chat GPT, Google’s leadership in international market share was indisputable. This chart shows its global market search in 2022, with Google capturing 93% of searches.
In November 2022, OpenAI launched the first free version of Chat GPT, revolutionizing the way people search on the internet worldwide.
This launch kickstarted a wave of similar search engines—some based on Chat GPT and others using proprietary technology.
Although Chat GPT’s release made OpenAI a household name, the company had been working on AI since 2015, testing various learning models known as Large Language Models (LLMs), which underpin Chat GPT.
Since then, its growth has been phenomenal, reaching 1 million users in just five days and achieving 100 million users in only two months.
Chat GPT’s market debut was very much in the spirit of Silicon Valley—a startup launching an open product that could potentially disrupt its entire market, albeit without fully addressing ethical dilemmas related to how it presents information.
Google responded swiftly by launching its AI-based search engine, Bard, in March 2023, initially accessible only by invitation. Google had to approach Bard’s release more cautiously, given the potential impact on its ad-based business model and its ongoing legal battles in the US and Europe.
Following its initial release, Bard evolved into Gemini, which is now publicly available with a subscription model similar to Chat GPT’s.
How AI-Based Search Engines Drive Traffic to Websites
For some queries, AI-based search engines can provide complete answers within their interface. For others, they include links to various websites, which, if followed by users, generate traffic for those sites. Let’s look at a few examples:
If we search for inspiration for a New Year’s Eve party dress on Chat GPT, using this prompt:
“Recommend a color for a party dress for New Year’s Eve 2024”
The search engine provides several color options and recommendations for each, in this case without linking to any sites:
With a charming final recommendation:
From this initial query, if we then ask, “Websites to buy black party dresses” it provides results with links that can direct traffic to those websites:
How to Measure AI Traffic with GA4
If we follow Chat GPT’s links, we arrive at various websites with URLs tagged using “utm” parameters:
…?utm_source=chatgpt.com
These tags, and others used by similar search engines, allow us to identify this traffic.
Currently, GA4 doesn’t have a dedicated channel for this type of traffic; instead, it is categorized under “referral.” Within referral data, we can see traffic with those “utm” tags.
Here is an example from one of the projects we work on at Relevant Group, showing traffic from GPT:
By identifying all traffic sources, we can create a comprehensive report on AI-based search engine traffic. This chart shows the evolution of AI traffic throughout 2024 for one of our clients.
At present, total traffic from AI sources is low but shows an upward trend.
The most interesting aspect of identifying this traffic is the potential to use these insights to improve our online visibility:
For example, by expanding this analysis to include the URLs driving this traffic to our website, we can estimate the types of searches users are performing on AI-based search engines related to our business. This enables us to develop strategies to optimize our content for these search engines.
And this is just one example of the vast possibilities offered by analyzing this data.
Are You Measuring Your AI Traffic? If not, at Relevant Group we can help you improve your project’s measurement.
