A Calm Approach to Shipping Content at Scale

Producing a lot of content used to be limited by how fast people could write. That constraint is mostly gone. Software, including AI writing tools, can now generate pages faster than anyone can read them. For a studio that runs several sites, the temptation is obvious: publish more, faster, everywhere.

We've learned that this is the wrong instinct. Volume without judgment produces pages nobody needs, damages reader trust and, as search engines have made explicit, can hurt a site's visibility. What follows is the calmer approach we try to practice: templates for structure, people for judgment and logs for honesty.

Why "more" stopped being a strategy

In March 2024, Google updated its spam policies to address what it called "scaled content abuse": generating many pages primarily to manipulate search rankings rather than to help users, whether those pages are produced by automation, by people or by a combination. Google's guidance is explicit that the method of production isn't the issue; the purpose and quality are. Its documentation on creating "helpful, reliable, people-first content" asks questions such as whether the content provides original information or analysis, and whether a reader would leave feeling they learned enough to achieve their goal.

Advertising networks apply similar standards. Google AdSense's program policies, for example, call out sites with low-value or replicated content. A page that repeats the same paragraphs or pads a thin answer to look substantial fails both tests.

We've seen this firsthand. Content that was produced quickly to fill a publishing calendar, with repeated structures and filler paragraphs, didn't serve readers and had to be rewritten. That cost more time than doing it carefully the first time.

Templates for structure, not for substance

Templates are genuinely useful. A consistent article structure helps readers know where to find things and helps writers start faster. What a template should never do is supply the content itself.

Good template elements:

• A clear title that matches what people actually search for or ask.
• A short opening that states who the article is for and what it answers.
• Headings that reflect the real sub-questions of the topic.
• A place for sources and a "last reviewed" date.
• Links to related articles on the same site.

Bad template elements:

• Stock paragraphs that appear in every article.
• Fixed sections that must be filled whether or not they apply.
• Word-count targets that encourage padding.
• Generic conclusions that could end any piece.

If a section doesn't apply to a particular topic, it should be cut, not filled.

Humans for judgment

Whatever tools assist with drafting or research, a person has to make the decisions that determine quality:

1. Should this article exist? Is there a real question readers have, and do we have something useful to add that isn't already answered well elsewhere?
2. Is it accurate? Every factual claim needs a source a reader could check. Statistics, quotations and citations must be real. If we can't verify something, it comes out.
3. Is it in the site's voice? Each site has a short voice note. An article that could appear unchanged on any site in the fleet isn't finished. We explain the reasoning in Fleet Sites and Shared Standards Without Blandness.
4. Would we be glad a reader found it? This is the simplest test and the hardest to fake.

Logs for honesty

When content is produced over months by different people and tools, it's easy to lose track of what was checked and what wasn't. A simple publication log keeps the process honest:

• What was published, where and when.
• Who reviewed it, and what was checked (facts, links, voice).
• What sources were used.
• What changed in later updates, and why.

Logs also make problems easier to fix. When we found filler on some of our own pages, a record of when and how those pages were produced made it straightforward to find every affected article and rewrite it rather than guessing.

Automated checks that catch the obvious

Some quality problems are mechanical and can be caught by simple scripts before anything goes live:

• Repeated paragraphs within an article or across a site.
• Thin pages whose real content, after removing navigation and boilerplate, is very short.
• Unfinished text, such as stub sections or empty brackets left over from drafting.
• Broken internal links and pages that return errors.
• Pages missing from the sitemap, or sitemaps listing pages that don't exist.

Automated checks can't tell you whether an article is good. They can reliably tell you when it's obviously bad, which frees human reviewers to focus on judgment.

What this looked like when we fixed our own pages

When we audited our sites, the problems we found weren't subtle once someone actually read the pages: articles where one paragraph appeared many times, generic "notes" sections bolted onto unrelated topics, and posts that were technically long but said very little. The fix wasn't a clever tool. It was the process above, applied one article at a time: decide whether the piece should exist, rewrite it from scratch for its real reader, check every claim against a source, keep the original URL so existing links still work, and run the automated checks again before publishing. It was slower than the original batch, and it produced pages we're glad to have our name on.

A sustainable pace

Calm doesn't mean slow for its own sake. It means publishing at the rate you can actually review. For a small team, that might be a few genuinely useful articles per site per month rather than dozens of thin ones. Over a year, that steady pace adds up to a library readers return to.

It also leaves room for maintenance: updating articles when facts change, consolidating overlapping pieces, and retiring pages that no longer serve anyone. We describe the measurement side of that in Measuring Fleet Health Beyond Vanity Traffic.

The short version

Use tools and templates to remove friction, never to replace thinking. Put a person in charge of whether each piece should exist and whether it's true. Keep a log so you can see what you did. And publish only as fast as you can stand behind every page. Our editorial policy sets out the standards we hold our own articles to, and Why We Prefer Explainers Over Outrage Posts covers the kind of content we choose to make.