Now that it has become clear with the Bing Search & OpenAI integration and the upcoming Bard public release from Google, one important question to ponder is the impact of this development on Search engine marketing (both paid search and SEO).
In this article, I will share my thoughts on this topic and also the potential impact on publishers.
Update May 2023: after this article was published in Feb 2023, I wrote the sequel in May here.
1) Potential impact on Paid search revenue
The key question is, what will happen to paid search revenue if people start to use Chat more to find the answers they are looking for instead of the normal search engine interface?
Many people speculate that in the short term, two things might happen:
- The total volume of the search query will go down because it takes fewer search queries for people to find exactly what they are looking for. The chat machine will give us the “exact” answers.
- The volume of queries using the traditional search engine interface will decrease since people seem to like the chat interface more.
However, given that the current Bing Chat results in many factual errors (even during the big announcement last week), it is not a foregone conclusion that people will use Chat bot for all types of searches and that results in a significant drop in the total query volume to the search engines. (Side note: this is my reaction to the wrong factual information from Bing Search).
We haven’t been able to test Google Bard yet as it is not released to the public. Given Sundar Pichai (Google’s CEO) public tweet “3/ We’ll combine their feedback with our own internal testing to make sure Bard’s responses meet our high bar for quality, safety, and groundedness and we will make it more widely available in coming weeks. It’s early, we will launch, iterate and make it better.” we have to assume that Google is well aware of many factually wrong answers given by large language models (LLMs) and try to minimize them and iterate.
People may learn very quickly that they need to use the traditional search interface when they want to look at the search results and judge the validity of the information.
So in the short term, one possible impact on paid search revenue can be:
- As the volume of search queries to the traditional search engine interface goes down, the total impressions/inventory goes down.
- With a smaller inventory and since it is an auction, advertisers may need to bid more for their ads to appear
- How much they will increase their bids depend on the ROI that the ads can bring. So there is a limit, depending on each vertical (industry)
- Given the macroeconomic situation in many advanced countries, we know that efficiency is something that advertisers will pay a lot of attention to in 2023, so they may not have the luxury of increasing the bid much.
- Proportionately, if the volume of queries goes down more than the increase in the bid price, then the paid search revenue might come down.
So this is bad for Google search advertising business?
Microsoft CEO Satya Nadella made it clear multiple times that Microsoft’s ambition is to go after Google’s market share in the search business. With Bing & OpenAI integration, Satya hopes that more people will be attracted to Bing’s search, Bing chat and start to use them more. So on balance, this potentially can bring in more paid search revenue for Microsoft.
In this scenario, many people are saying that the situation is worse for Google paid search business because the total volume of queries might come down, and Google might lose market share. It’s a double whammy.
There is another possibility in the medium term. That is Google/Bing finds an intuitive way to incorporate sponsored data/messages into the Chat interface. And that represents a net revenue stream for the search engines. This is not easy to do because it may impact the quality of the chat answers. But it is not impossible. (Well, we didn’t have to wait long. Here is the first iteration of the ad on Bing Chat that I saw this week Feb 16)
So what is my conclusion about the impact of chat bot like experience (OpenAI integration with Bing, Bard integration to Google search) on paid search revenue?
It is too early to tell. I know some of you (or many) will be disappointed to see this conclusion from me. But given the above context and rationale, I don’t think it is clear at all which direction the situation can go. And don’t forget that the entire process is reflexive, so we may use search engines and their chat functions more, not less.
2) Potential impact on Search engine optimization (SEO)
To me, the potential impact on SEO is a bit clearer. The SEO game is changing rapidly because
- With ChatGPT, Bing Chat, and Bard in the near future, the cost of generating content will decrease significantly.
- The speed of generating content is reducing quickly too.
- As reported by Search Engine Journal, Google sets the record straight that they value high-quality content, regardless of whether humans or machines generate it. In other words, as long as Google machine thinks that your content is of high quality, matching the user search intent, your content will show up, even if AI writes it.
- The above means that SEO practitioners will need to up their game significantly because everyone now has better tools.
In a world where AI-generated content is flooding the market, unique perspectives and unique content will stand out more. So in a way, the focus on the user, on high-quality content, has Not changed for SEO. So perhaps from a reflexivity angle, while the cost and speed of generating mediocre content is going down quickly, the cost and effort of generating really high-quality content are getting even higher 🙂
But that is not all. Searching using a chat interface means people no longer click on websites. This has a direct impact on the potential traffic from organic search.
So well, it means SEO practitioners need to work harder (against the competition), for a potentially smaller return (i.e., less traffic).
Another possibility is that SEO practitioners adapt and learn quickly and find ways to work with how people prefer to use a Chat interface for certain queries and keep organic traffic coming.
3) Potential impact on publishers
This seismic shift in generative AI and its integration with search engines means that:
- Many publishers might see a drop in organic search traffic to their site, especially around non-navigational search queries.
- More importantly, as Bing Chat/chatGPT and other LLMs can synthesize the content across the entire web so well, users may find it less compelling to visit individual websites to find what they are looking for.
- Of course, this is contingent upon different companies’ ability to improve LLM’s tendency to make things up.
Having less traffic to their website means publishers need to try even harder to monetize existing traffic. Will this mean an acceleration towards more and more content being behind a paywall, accelerated adoption of the subscription model, etc.?
Another possible outcome is to produce more non-text content like videos, podcasts, etc.
Also, as AI tools are becoming more and more advanced and prevalent, the cost of generating text articles or images will decrease quickly. These tools can improve the efficiency of content producers. Hence, this may have a net positive impact on publishers.
It does feel a bit premature to have a firm conclusion at this early stage.
That’s all from me. What do you think?
P.S: If you are keen to read about my educated guesses on how OpenAI & ChatGPT’s Integration with Office 365 can look like, here it is.