What I Got Wrong About Publishing at Volume

Does publishing more actually work?

It worked for me, but not for the reasons I expected and not without costs I did not price in at the start. I have published more than 350 articles on AI answer engines, schema and E-E-A-T, and the returns came from a few specific things rather than from the volume itself.

I want to write this one honestly, because the case for publishing at volume is usually made by people selling the idea. The case against it is usually made by people who tried it badly. Neither is much use if you are deciding what to do this quarter.

So here is what I got wrong, what I would keep, and what I would not do again.

What did I underestimate about volume?

The maintenance. Every article you publish is a small permanent obligation, and I planned for the cost of writing without planning for the cost of owning. At a few dozen articles that is invisible. At several hundred it becomes the largest recurring demand on your time.

The obligations are unglamorous and they compound. Facts go stale. Products get renamed. A tool changes its pricing and now four of your articles are quietly wrong. Links rot, internal links point at pages you later restructured, and none of it announces itself. You find out when a reader does.

If I were starting again, I would decide the maintenance budget before the publishing target. Something like one day a month per hundred articles, held as seriously as the publishing schedule. I did not do that, and I spent a long period doing maintenance reactively, which is the most expensive way to do it.

Where does quality actually break at scale?

At the research stage, not the writing stage. Writing degrades gracefully. A tired article still reads acceptably. Research does not degrade gracefully. It fails silently, because an unverified claim looks exactly like a verified one on the page, and you cannot see the difference by rereading your own draft.

This is the thing I would tell anyone considering volume. The bottleneck you think you have is production. The bottleneck you actually have is verification, and the temptation to skip it grows in direct proportion to how many pieces you have committed to publishing. That pressure is structural rather than a failure of discipline.

My response was to make verification the part that cannot be compressed. Everything else in my process can shrink under time pressure. Research breadth, commentary, length, all negotiable. Whether a claim traces to a primary source is not, and I would rather publish fewer pieces than relax that.

What is the real cost of a near-duplicate article?

Higher than I thought, and it is not mainly a ranking problem. It is that you now have two pages that half-answer a question, neither of which is the best page on the internet for it, and a reader who lands on the weaker one leaves with a worse impression of you than if you had published nothing.

At volume, near-duplicates are not a risk, they are an inevitability unless you actively prevent them. The topic space in any niche is finite, the obvious angles get used early, and after a few hundred pieces you are genuinely at risk of rewriting your own work without noticing. Memory is not a defence at that scale.

The fix that worked was mechanical rather than editorial. Check every proposed title against everything already published, by comparison rather than by recollection, and treat a near match as a prompt to find a sharper angle. It is a boring step and it has saved me from republishing my own ideas more times than I would like to admit.

How do you keep a personal voice across hundreds of articles?

By writing from a position rather than to a template. What makes an article mine is that it contains an argument I would defend in a room, including the parts where I say something did not work. A template produces consistency and no reason for anybody to choose your version.

Volume does put real pressure on this. The efficient move is always to reuse a structure, and structures are fine. The dangerous version is reusing opinions, which produces a body of work that agrees with itself everywhere and says nothing anywhere. I have caught myself doing it, and the tell is an article I could have written without thinking.

The test I apply now is simple. Is there one sentence in this piece that only I would write, because it comes from something I have actually seen? If not, the piece is a summary of the internet and it does not deserve to be published under my name.

What would I automate, and what would I never automate?

I would automate every part that is mechanical and checkable. Duplicate checking, validation, field formatting, publishing, logging, link verification. I run automations in production for client work too, including an Airtable and WhaleSync pipeline for Ajust and a HubSpot workflow for Kismet Health through Zapier, and the pattern is the same in both contexts.

What I would not automate is the decision about whether something is true, and the decision about whether a piece is worth publishing at all. Those are the two places where an error is not a bug you fix later, it is a thing a reader read and believed. The economics of getting them wrong are completely different from the economics of a formatting mistake.

The useful line is between checkable and judgeable. A machine can confirm that a heading structure is valid or that a link resolves. It cannot decide whether you have understood a source correctly, and pretending otherwise is how people end up with a large archive they cannot stand behind.

How do you know whether volume is working for you?

Look at whether the archive compounds. If older pieces keep earning impressions, keep getting linked, and keep sending you enquiries, the volume is building something. If only the newest pieces get attention and everything older goes flat, you are running a content treadmill and the volume is a cost.

Search Console answers this in about ten minutes. Compare performance for pages published in the last month against pages published six months ago. A healthy archive has a long tail that keeps working. An unhealthy one has a spike and a cliff, and no amount of additional publishing fixes the underlying shape.

The related question is whether frequency itself is doing anything, which is worth separating from whether your content is good. I looked at that separately when asking whether publishing more often gets you cited by AI search, and the honest answer is more nuanced than either side of that argument usually admits.

Who should not do this?

Anyone who cannot commit to the maintenance, and anyone whose business does not need a large surface area. If you sell to fifty companies in the world, a hundred articles is the wrong shape of effort, and twenty excellent pages plus direct relationships will beat it every time.

I would also steer away anyone who is publishing to feel productive. Volume is unusually good at producing the sensation of progress, because there is always a next piece and it always feels like work. That is exactly why it is worth being strict about what the archive is actually returning.

My own case works because search and answer engines are how my clients find me, so a large, accurate, well-linked archive is the product of the marketing rather than a by-product of it. That is a specific situation. It is not a general recommendation, and I have seen people copy the tactic without the reason.

What would I do differently from the start?

Three things. Set the maintenance budget alongside the publishing target. Make verification the incompressible step from day one rather than after a scare. And build the duplicate check before the archive is big enough to need it, because retrofitting it is far more work.

I would also structure the internal linking as I went rather than fixing it later. Orphan pages accumulate quietly and then require a project to resolve, which is why I ended up writing about auditing orphan pages on a large blog. Ten seconds per article at publishing time is cheaper than a week of remediation.

What I would not change is the decision to publish under my own name with my own opinions in it. It makes every piece slower and it is the only reason any of it is distinguishable from the rest of the internet.

What should you do next?

Before you commit to a publishing target, write down the maintenance number that comes with it and decide whether you will actually pay it. If the answer is no, publish less and maintain it properly. A small accurate archive beats a large one you are quietly embarrassed by.

Then decide which step in your process is incompressible. Every process bends under deadline pressure, and the only question is which part you have decided will not. Make that decision deliberately now rather than accidentally in month four.

And pace yourself honestly, because this is a long game and the failure mode is not usually a bad article, it is stopping entirely. I have written about how I avoid burnout running a solo practice, and it applies here more than anywhere. If you are weighing up whether volume is right for your business, reach out and I will give you an honest answer rather than an encouraging one.