Essay
For most of the last two decades, the company that could produce more content, in more formats, more consistently, usually won the channel. Volume was a genuine competitive advantage, because volume was expensive: it required writers, designers, editors, a production calendar and a budget most competitors did not have. Scarcity of supply is what made output valuable.
That scarcity is gone. A team of any size can now generate more articles, images, videos, product descriptions and social posts in a day than a large agency could have produced in a quarter. The machines do not get tired, do not need a brief twice and do not ask for a deadline extension. Supply has effectively become infinite, and the moment supply becomes infinite, whatever used to be scarce about it stops being an asset.
Infinite supply does not mean infinite value
This is the part many companies are getting wrong right now. Faced with tools that can produce content at almost no marginal cost, the instinct has been to produce more of it. More blog posts, more variants, more languages, more channels, on the theory that some of it will land. That theory used to be defensible when producing a unit of content was itself a filter, because only things worth the effort got made. The filter is gone, and without it the strategy collapses into noise competing with noise.
A market flooded with competent, well-formatted, machine-produced material does not reward the company that adds one more competent, well-formatted, machine-produced piece to the pile. It rewards whoever manages to be the exception in a feed where everything else looks the same. Abundance does not raise the average outcome for everyone. It raises the bar for what counts as worth noticing at all.
When everyone can produce everything, producing something is no longer the achievement.
Relevance is decided by emotion, context and choice
People do not read a feed the way an analyst reads a report, weighing every item on its merits. They scroll past almost everything and stop for almost nothing, and the decision to stop happens before conscious evaluation, driven by whether something connects to a feeling, a moment or a problem they are actually carrying that day. A piece of content can be technically accurate, well designed and perfectly on-brand, and still be invisible, because it arrived with nothing that mattered to the specific person at the specific moment they encountered it.
Context does more work than most content strategies admit. The same message lands completely differently depending on what else is happening in someone’s life, their industry, or their week. Machine-generated volume optimizes for coverage: producing a plausible version of content for every audience, every keyword, every occasion. It rarely optimizes for the much harder question of whether this particular audience, at this particular moment, has any reason to care. Coverage and relevance are not the same objective, and companies chasing the first are often actively working against the second.
- Would we notice if half of what we publish simply disappeared?
- Is this piece answering a question our audience is actually asking this week, or a question our content calendar assumed they would ask?
- Could a competitor’s machine have produced this exact piece for their exact audience with a different logo?
- What are we saying that requires us specifically to have said it?
That last question is the sharpest one, and it is worth applying to an entire content operation, not just a single piece. Very little of what fills most companies' publishing calendars would fail that test in an obviously honest way, which says more about the current state of content strategy than any benchmark report could.
What companies should actually stop producing
The practical move is not publishing less out of caution. It is being ruthless about which categories of output deserve human attention at all, and which should stop, or be handed entirely to the machine, because nobody was reading them for the writing in the first place. Compliance-adjacent updates, routine product listings, basic explainer content answering a question already answered a thousand times elsewhere: none of that was ever a relevance play, and pretending otherwise wastes the attention of people who could be doing work that actually differentiates the company.
What should not be automated away is the smaller amount of content that carries a genuine point of view, a specific piece of judgement, or an emotional register that comes from someone in the company actually caring about the outcome. That is the material a machine can imitate the shape of but not the substance of, because the substance depends on a human having actually thought something, felt something or risked something in saying it.
The business implication is organizational, not creative
The companies that will do well in an infinite-supply environment are not necessarily the most creative ones. They are the ones willing to shrink their output, kill content categories that were never earning attention, and reallocate the freed capacity toward the much smaller number of things that actually require a person to have made a judgement call. That is an uncomfortable decision for any marketing organization built around a publishing calendar and volume targets, because it means being measured on relevance rather than throughput, and relevance is harder to defend in a monthly report.
It is also the only sustainable position available. Competing on volume against systems that produce volume without cost or fatigue is a race no company can win by definition. Competing on relevance, which still requires understanding a specific person’s specific moment well enough to say something that matters to them, is a race machines can assist with but not settle on their own.