How many articles before SEO starts working?

People ask this expecting a number. Thirty. Fifty. A hundred.

There is no number, and the reason is not evasion. Article count is not a thing Google measures. Nothing in how ranking works counts your posts and unlocks something at a threshold. A site with 12 pages can outrank a site with 400, and does, constantly.

But the question is not stupid — it is pointing at something real, just with the wrong instrument. What people are actually asking is: *how much work before this starts paying?* That has an answer, and it is about coverage and depth rather than volume.

Why the count itself does nothing

Ranking happens per query. Each page competes on its own for the searches it addresses. A hundred pages is a hundred separate competitions, not one big score.

So publishing more helps only in so far as it does one of three things:

1. Enters more competitions. Ten pages target ten sets of queries. A hundred target more. This is the honest, boring reason volume correlates with traffic.
2. Covers a subject more completely, which appears to help a site be treated as a credible source on that subject.
3. Creates internal links, which help pages be found and help establish what your site is about.

If publishing does none of those — twenty posts on unrelated topics that nobody searches for, linked to nothing — the count rises and nothing happens. That is the commonest version of "I published 50 articles and got no traffic".

What actually determines when it starts working

Four things, roughly in order of impact.

Coverage of one subject, not many

Fifteen articles that map a single subject thoroughly outperform fifty scattered across five unrelated ones. This is what "topical authority" means when the jargon is stripped out — you have covered a territory well enough to be a reasonable source on it.

Practically: pick one subject your business can genuinely own and cover it properly before starting a second. Topic clusters and pillar pages sets out the structure.

Whether the queries are winnable at all

Twenty articles targeting queries dominated by large established sites produce nothing. Twenty targeting specific questions nobody has answered well produce results quickly.

This is why volume advice is so unreliable: the same twenty articles can be a waste or a success depending entirely on target selection. Judging that is covered in judging whether you can realistically win a keyword.

The site's starting position

A site with existing links and history ranks new pages faster than a site with none. Same article, same quality, different outcome — the new site simply has less to go on.

Time, which volume cannot buy

Pages need to be crawled, indexed and observed before they settle. Publishing fifty articles in one week does not compress that; it just means fifty pages are simultaneously waiting. How long SEO takes covers the sequence.

A better question, and how to answer it

Replace *"how many articles?"* with "have I covered one subject well enough that someone researching it would find me useful more than once?"

That is answerable without a number. Take the subject you want to be known for and write down every question a customer asks about it — from sales calls, emails, the phone. Most businesses reach 20 to 40 without effort.

Then: how many have you answered properly? If it is three, you have not covered the subject, whatever your post count says. If it is thirty, you have, and the remaining variable is time and competition rather than volume.

Turning the question list into a publishing order

Having thirty questions is not a plan. The order you write them in changes how long it takes to see anything, so it is worth ten minutes.

Sort them into three groups.

Questions with a clear, specific answer that few pages address well. These go first. They are narrow, the competition is thin or evasive, and they are the ones that will actually rank within a few months. They also produce the earliest evidence that the approach works, which matters more than it sounds — a plan with no feedback for a year gets abandoned.

Questions everyone answers. These go second. You will not win them early, but a subject looks uncovered without them, and they are what the earlier pages link to.

The broad question the whole subject sits under. This is the pillar, and it is written early but expected to rank late. Its job at first is to exist so that everything else has somewhere to link.

A workable order for a first fifteen: the pillar, then eight narrow questions, then the six obvious ones, revisiting the pillar once there is something to point at.

What that avoids is the standard failure — starting with the broadest, most contested question because it feels like the most important one, ranking for nothing, and concluding after four months that content does not work.

Does cadence matter?

Less than people think, and not in the way they assume.

There is no evidence that a particular publishing rhythm is rewarded in itself. What consistency actually buys is more mundane: a site that publishes something every fortnight for a year has thirty pages, and a site that publishes six in one burst and stops has six. The rhythm matters because it produces the work, not because it is observed.

Two practical points. Publishing everything at once means every page begins its waiting period simultaneously, so you learn nothing until they all resolve — spreading it out gives you readings along the way. And a cadence you can sustain beats an ambitious one you abandon in month two, which is the more common outcome by a wide margin.

What to expect, honestly

Illustrative rather than measured — these are the shapes to expect, not figures from a study.

| Stage | What it looks like | Roughly | |---|---|---| | Nothing at all | Indexed, no impressions | First weeks | | First flickers | A handful of impressions on very specific queries; positions deep in the results | 1–3 months | | Something real | Consistent impressions, a few pages on the second page, occasional clicks | 3–6 months | | Working | Steady clicks, some first-page positions, enquiries mentioning what they read | 6–12 months |

A new site in a competitive subject sits at the slow end. An established site in a niche subject can move much faster.

The most common mistake is quitting between "first flickers" and "something real", because the numbers are small and look like failure. They are the normal appearance of progress at that stage, and the cost of the work is already spent.

Signs you are on track before traffic arrives

Traffic is a lagging indicator. These move earlier:

• Impressions rising even while clicks stay near zero — you are being shown, just not yet high enough
• The number of distinct queries you appear for growing, even at poor positions
• Pages you never promoted picking up impressions
• Average position improving on queries you already appeared for
• Someone mentioning an article unprompted

If none of these are moving after three or four months of genuine publishing, something is wrong with target selection or with the site itself, and the answer is diagnosis rather than more articles. Start with the diagnostic checklist.

Publish more, or improve what exists?

At some point the better return moves from adding to improving. A useful rule:

If you have covered a subject and pages are stuck between positions 8 and 20, improving them usually beats writing new ones. They are already close, already indexed, already understood. A rewrite that genuinely answers the question better moves them further than a new page starting from nothing.

If you have not covered the subject, keep publishing. Improving three articles into excellence does not compensate for having only three.

Content refreshes: what to update and what to delete covers the mechanics of the first case.

A note on publishing volume for its own sake

Google's guidance on helpful content is explicit that producing large amounts of content mainly to attract search traffic is not the goal, and that the standard is content made for people. Volume strategies that ignore this have a poor record, and the sites that pursue them tend to look identical to one another.

The practical version: if you would not send an article to a customer who asked that question, it should not be published, and it will not help your count.

When more articles are the wrong answer entirely

Worth saying because we sell an alternative and would rather say it plainly than have you discover it late.

If you need customers within a couple of months, no article count solves that. The work described here compounds and pays later; it does not pay soon. A business in that position should be doing something with a shorter cycle — direct outreach, advertising, or paid placement — and starting the articles in parallel rather than instead.

The comparison, including where our own product is the wrong buy, is in paid versus organic. Current placement prices are on the rankings board.

Common questions

Is there a minimum to be taken seriously? No. Small sites rank. Coverage of a subject matters; a count does not.

Should I publish daily? Only if daily articles are as good as weekly ones would be. They usually are not.

Do old articles keep working? Good ones do, for years, which is what makes this worth doing at all.

What if I run out of topics? You have almost certainly not covered the ones you have. Re-read your last twenty customer emails.