This morning's post on the business blog was planned, two weeks ago, around the keyword "ai for marketing". It probed live then and it probed live again this morning. Ten suggestions back from Google, all ten genuine. By every test I had a month ago, that is a green light.
Then I read the ten.
Course. Course free. Certification. HubSpot Academy. Strategy. HubSpot. And advertising. And sales. Free.
Five of those are somebody trying to learn the subject, two of them from one named vendor. The person typing "ai for marketing" into Google in South Africa this week is, more often than not, a marketer who wants a certificate on their profile before the next round of job applications. They are not a business owner looking for someone to do it for them. They are not even a business owner.
That is a real audience. It is a large audience. It is just not mine, and no amount of ranking for it would put a lead in the pipeline.
What the market did, not what I did
I want to write this one as an observation about the market rather than as a story about a mistake, because it is not a mistake. Nothing was published wrong. The rule that caught it was written on the last day of August after a compliance keyword turned out to belong to Indian corporate lawyers, and this is the fourth time in twelve days it has changed a decision. It is doing exactly what it was built to do.
What interests me is what the ten suggestions say about where AI in marketing actually is right now, at least on this side of the search box.
The demand is for credentials, not outcomes. When a category is new and the tools are cheap, the first wave of search interest is not "who can do this for me". It is "how do I become the person who can do this". That is what a certification cluster looks like from the outside. It is a labour market signal, not a buying signal.
I find that slightly reassuring and slightly alarming in equal measure. Reassuring because it means the owners I actually work with are not being flooded with AI marketing pitches yet, or if they are, they are not searching for them. Alarming because the next wave is obvious: a lot of newly certified people are about to go looking for businesses to practise on.
What I did instead
The neighbouring term is "ai marketing tools". Also ten out of ten live. But read the ten: for small business, free, for social media, list, for content creation, for B2B, for small business free. That is somebody shopping. Twice in the list they say who they are, and it is my reader.
So the post shipped under "ai marketing tools" as the primary and kept "ai for marketing" in the opening paragraph and one FAQ heading, because the term is alive and the post genuinely answers it, just not for the person most likely to type it.
The post itself is about which marketing jobs are safe to hand to a tool when you are a team of one or two, and which are not. Four to hand over: drafting from a source, reshaping one piece into many, checking against rules, scheduling. Three to keep: deciding what to say, any number that has not been checked, and replying to a person. The second of those three is Wednesday's near miss, and the first is what this post is about. Choosing not to chase a live keyword is a positioning decision. A tool can measure the ten suggestions. It cannot tell you that your business should walk past them.
The part I keep coming back to
A month ago my keyword tests only asked whether a term was alive. Then whether it was alive in this country. Now whether the people behind it are the people I want. Each of those is a cheaper question than the one before, and each one has thrown out a keyword that passed the earlier tests.
I do not think the sequence stops here. There is presumably a fourth question I have not thought of yet, and the keyword that fails it is already sitting in the October plan looking perfectly healthy.
But that is a different post, and I would rather write it after the market shows me what the question is than guess at it now.