Way back in the stone age during the earlier part of my career I worked in digital marketing and large part of that job was an emerging mysterious discipline known as SEO. The job was focussed trying to understand a new search engine called Google and how to rank number 1. Many businesses built their whole commercial engine on the basis of ranking on Google for certain keywords. If you were someone that knew how to optimise websites and get them ranking for keywords that drove business you were in high demand.
The challenge was no one really knew how it worked. I had to spend hours every day trying to figure it out by consuming the latest insights from related forums and blog posts. I studied both black hat and white hat tactics. Blackhat tactics were cheat codes for getting ranked. People in the early days would hide keywords as background text. For example, if the website had a white background, people would add keywords as white font. Kinda like how people are keyword stuffing CVs today, so AI detectors choose their CV.
AI sits between the business and the buyer. This means when someone is asking ChatGPT or Claude what to buy or which vendor to consider. The AI will research a set of options and bring back a shortlist. The person may make the final call, but businesses who are not on this shortlist are invisible and will never be considered. It’s the same as not being on the first page on Google search when search was one of the main drivers of business.
OpenAI is already asking merchants for structured product feeds so ChatGPT can understand what they sell and keep prices current. Google is building protocols that allow agents to move through discovery, buying and payment.
I work in payments, so of course my mind went straight to the digital transaction. Google has built a payments protocol that lets people give an agent a budget and rules for what it can buy. .
An AI system has to work out what you sell, where you operate, how you compare and whether there is proof behind the big claims you make. It may compare your website with case studies, reviews and what other companies say about you. It needs to figure out what is legit and what is just vaporware. Because if the recommendations that come back in the chat are low quality, it’s like Google returning search results that were low quality. It’s bad for business.
We will still need to deeply understand the human side of marketing. In fact this will grow in importance as the logical science side of marketing will also be machine driven. Understanding customer psychology, culture, story and all the strange reasons why people make buying is how many business will differentiate themselves from bland machine driven competitors. But someone will also need to understand how AI systems work.
I don’t think every marketer needs to become an AI techie. But you will need to be fluent in AI.
I’d imagine the machine side of marketing starts in a similar messy way to SEO, spread across many functions like product marketing, data and some early adopter that notices the problem first. Knowing marketing, we will probably invent a terrible acronym and act like it has existed for years.
I’m going to both wholeheartedly agree with you while also disagreeing with you. Marketers who’ve gone all in on digital are going to have a bad year. In fact, I would argue that last year was the bad year siren for those who were paying close attention to their data and have been charting the trends over time, this year will just be a rapid revolution based on more of the same. But those who have multi-channel routes to market (including analog) will hold steadier. Integrated marketing has never been more critical and marketers skills beyond just the Promotional P should get a dusting off as they come of the shelf and play with Pricing (powerful, given the state of the world’s finances), Place (even more powerful) and hopefully some good old Product innovation to boot. The market has also shifted. The channels to market are just a component of that.