The integration of large language models into search engines will radically disrupt all elements of web-based marketing. In fact, tools such as Chat-GPT could end the practice of “inbound marketing” as we know it altogether.
Inbound marketing has been a staple of web-based marketing since the idea was first expounded in 2005 by HubSpot co-founder Brian Halligan. Put simply, and in contrast to traditional “outbound marketing”, inbound marketing relies on creating lots of unique content that gets a website found via organic search. Search Engine Optimization (SEO), which is the science of how to make sure a website ranks highly in organic search results thus became a core competency of all successful inbound marketing teams.
Inbound marketing’s appeal comes from two facts:
First, people who are searching for something generally have intent. Therefore, it isn’t necessarily the company that spends a lot of money creating awareness and intent, but the company that is best at SEO that gets traffic to their website, hence customer engagement.
Second, as the price of online ads (particularly display ads) has escalated, SEO and organic results generally have much higher ROI.
As an example, I am aware of a software company that receives organic search traffic worth $10 million annually yet pays less than $1 million annually for the total cost of running their web presence.
The question then is why ChatGPT, bing and other large language models could threaten inbound marketing. There are two reasons:
1. Search engines will no longer be an “entry point” to the web. Currently, when we use search engines, they return a list of web addresses. The user then has to click on those address links, visit the related website and consume the content presented.
In the case of a large language model, answers are presented in-line and conversationally. There is no need to leave the experience and thus no need to visit a particular website. There may in fact be no awareness whatsoever of the source of the information.
As an example, if I ask Google search to define the term MRR (monthly recurring revenue), it returns many, many links to websites that define the term, and I need to click one of those links and maybe do some navigation to find the answer I’m looking for.
However, if I ask Chat-GPT the same question, I get the following answer (current as of March, 2023):
MRR stands for Monthly Recurring Revenue. It is an essential metric for subscription-based businesses or managed service providers (MSP) to track. MRR can help forecast gaps in revenue goals and set milestones for revenue targets. It can also aid in creating a budget by providing a reasonable expectation of monthly income.
Clearly there is no need to go any further. This is a massive time saver and much better consumer experience. And, all those companies with the highly optimized SEO strategies have lost out on my visit to their websites.
2. Search Engine Optimization will become irrelevant. As mentioned above, most websites are specifically tuned so that they show up in organic search results and as high as possible. SEO is an “arms race” of sorts. At this time, we don’t know how the large language models decide what content to display, how this can be influenced, or even if the content they return is factually correct or the most relevant available. Can these models be manipulated (or “gamed”) the way SEO does today? I suspect it would be difficult and will get even more difficult as the models become more and more sophisticated.
What is the upshot of all this for companies that rely on inbound marketing today?
Frankly, it is unclear at this stage what, if anything, can be done to fully mitigate the impact of large language models on inbound marketing, especially the incredible cost-efficiency of SEO-driven approaches.
However, there are several activities which companies should start doing now that may be able to mitigate the impact.
First, organizations should begin immediately collecting as much first-party data as possible. If large language models disintermediate companies from their customers and potential customers, it will be extremely difficult and costly to find channels like the current search engines which are so efficient and scalable when it comes to delivering prospects to your website. Having a strong data approach will underpin the future of marketing and the development of your own AI-based approach.
Second, begin thinking about how to shift focus to activities that are higher in the funnel. This could be outbound marketing, social/earned media, digital marketing.
Finally, and most importantly, is to get ahead of this. Stay connected to those in your network that are dealing with the same challenge, follow the space and stay current with the latest developments/research, and make it your job – or someone in your team’s job – to be the expert on AI, its impact on marketing and search. Queen’s University in Ontario has recently launched a Master’s of Management program in AI management, and there are other similar programs popping up around the world, so there’s no excuse for not having a good understanding of AI.
ความคิดเห็น