Ask how long a blog post should be and you will get a number: 1,500 words, or 2,000, or "long-form ranks better". Ask how often to publish and you will get a schedule: three times a week, or daily, because "Google rewards freshness".
These numbers are not from Google. It is worth being precise about that, because they drive an enormous amount of wasted effort.
What the documentation says
Google's guidance on creating helpful, reliable, people-first content includes a list of self-assessment questions — questions to ask about your own pages. Among the ones framed as warnings about search-first rather than people-first content is this:
Are you writing to a particular word count because you've heard or read that Google has a preferred word count? (No, we don't.)
That parenthesis is Google's, not ours. The question is listed as a symptom of writing for a search engine rather than a reader.
The same document asks whether you are "producing lots of content on different topics in hopes that some of it might perform well in search results" — which is publishing frequency, framed as a problem rather than a target.
There is no image quota in the guidance either. There is no minimum heading count, no keyword density, no ideal paragraph length.
Why the folklore persists
The numbers did not appear from nowhere. They come from correlation studies: someone measures the top-ranking pages for a set of queries, finds they average some length, and publishes the average.
The finding is usually real. The inference is backwards. Pages that thoroughly answer a complex question tend to be long because the question needed that much answering. Length is a side effect of completeness. Padding a thin article to 2,000 words reproduces the symptom and not the cause, and produces exactly the kind of page the helpful content guidance describes: content that leaves a reader feeling they need to search again.
The scaled content abuse policy is the sharper end of the same problem. It concerns generating many pages primarily to manipulate rankings rather than to help people — and it applies regardless of how the pages are produced. "We published a lot" is not a defence. Volume produced without a reason for each page to exist is the behaviour the policy names.
So how long should it be?
As long as it takes to answer the question completely, and no longer.
That sounds like a dodge. In practice it is an operational test, and a strict one:
- Write down the specific question the page answers.
- Answer it, including what a reader would reasonably ask next.
- Remove everything that is there for length rather than for the reader — the throat-clearing introduction, the definitions your audience already knows, the "in today's fast-paced digital landscape" paragraph.
- Whatever remains is the right length.
When we generate drafts, articles tend to land somewhere between 1,500 and 2,500 words. That is an outcome of answering a question properly, not a target the drafting was aimed at, and a post that genuinely finishes at 900 words is finished.
And how often?
Here is the honest answer: as often as you can publish something worth reading, and not one post more.
For most small teams, twice a week is already demanding — not for the writing, but for the reviewing, the fact-checking, and the original insight each post needs to justify itself. One genuinely useful post a week beats three thin ones, and it beats them on the exact criteria Google's guidance describes.
There is a real argument for consistency, and it is not the one usually made. A steady schedule helps because it forces a habit of noticing what your customers actually ask, not because a crawler prefers Tuesdays. Publishing on a schedule you cannot sustain at quality is worse than publishing less often.
If it helps to have a concrete starting point rather than a principle: two posts a week, on fixed days, with a genuine reason for each one to exist. Adjust down the moment quality slips. Adjust up only when you have a backlog of things worth saying.
What about images?
Use an image where it explains something a paragraph cannot — a process, a comparison, a screenshot of the thing you are describing. That is the entire rule.
Stock photographs of people pointing at whiteboards do not help a reader and do not help a search engine. A diagram that makes a workflow legible does both. If you cannot say what an image is for, it is decoration, and decoration has a cost: page weight, and a reader's attention spent on nothing.
Where you do use images, the accessibility work is the SEO work — real alt text describing the content, sensible file names, dimensions set so the layout does not shift while loading.
The uncomfortable part
Word counts, posting schedules and image quotas are popular because they are measurable and achievable. You can hit 2,000 words on a topic you know nothing about. You can publish daily without having anything to say. They convert a hard problem — do we have something worth publishing? — into an easy one.
The guidance Google actually publishes keeps returning to the hard problem. Does this content demonstrate first-hand expertise? Would someone reading it feel they had learned enough? Would they bookmark it or recommend it?
Those questions have no target number, which is exactly why they are the ones worth asking.