How to Write a Blog Post That Ranks: 10-Step Process

Most business blogs fail at the topic, not the writing. A well-written post about something you cannot rank for, or something nobody searches, is wasted effort no matter how good the prose is.

This is the process we run for every client post. It is deliberately unglamorous, and we are giving away the whole thing, because the work is in the doing rather than the knowing.

Step 1: Can you actually rank for this?

Before writing a word, decide whether the topic is winnable. Three questions.

Does anyone search it? If a term gets almost no searches, a perfect post will still bring nobody. Check it in Keyword Planner or simply try Google autocomplete: type your topic and see whether Google offers to complete it. If it does not, people are not asking it.

Who is already ranking? Search the term and look at page one. If it is dominated by national brands, aggregators and sites a decade old, a young site is not displacing them this year.

Can you genuinely write something better? Not longer. Better. If the top three results already answer the question well and you have nothing to add, pick a different topic. Real prices, a real process, or first-hand experience are what "better" usually means.

If the answer to any of these is no, move on. Choosing the right topic is most of the job.

Step 2: Read the SERP before you write

Google has already decided what a query means. Your job is to notice, not to argue.

Open the search results and fill in a table like this before drafting.

Two things this reveals immediately. If page one is all guides and you planned a service page, you will not rank with a service page. And if half the results are from one competitor, they have built topical depth there and you need a narrower angle.

Step 3: Build the outline from real questions

Do not invent headings. Collect them.

• People Also Ask. Expand a few boxes and note the questions. These are literally what people ask.
• Google autocomplete. Type your topic and add "how", "why", "cost", "vs" and let it finish.
• Your own enquiries. The questions clients ask on calls are the highest-value source and nobody else has them.
• The comments and reviews on competitors, where people say what the existing content missed.

Then arrange them so the post answers in the order a person actually thinks. A reusable shape that works for most informational posts:

The outline template

1. The direct answer, in the first 60 words
2. Why this matters, briefly, only if it is not obvious
3. The main body, one H2 per real question from your research
4. A table or checklist, so the post is worth returning to
5. A worked example, with real numbers
6. What does not work, which is often the most useful section
7. FAQ, five questions with short answers
8. What to do next, one clear step

Not every post needs all eight. Most good ones use six.

Step 4: The first 60 words decide everything

Answer the question immediately. No throat-clearing, no history of the industry, no "in today's competitive market".

Weak: "In today's fast-paced digital world, businesses across India are increasingly recognising the importance of a strong online presence. In this comprehensive guide, we will explore everything you need to know about local SEO."

Strong: "Local SEO is the work that gets your business into Google's three-result map pack. For most Indian small businesses it means fixing your Google Business Profile, keeping your address consistent everywhere, and collecting reviews steadily. It takes about a weekend to set up and a few months to show."

The second version tells a reader whether to keep reading. The first wastes their time and yours.

This matters more now than it used to, because assistants lift self-contained passages when they answer questions. A direct opening is what makes a page quotable, which we covered in detail in our guide to checking whether AI search cites your business.

Step 5: Structure the body to be skimmed

Almost nobody reads a post end to end. They scan, stop where it looks relevant, and read that part.

• One idea per section. If an H2 covers two things, split it.
• Short paragraphs, two to four sentences. Long blocks get skipped entirely.
• Headings that are statements or questions, not labels. "How much does it cost" beats "Pricing".
• Tables for anything comparative. They are the most-read element on most pages.
• Bold the sentence that matters in a long section, not random keywords.
• Numbers over adjectives. "Four to eight months" beats "takes time".

Step 6: On-page mechanics, without the stuffing

• Title tag under 60 characters, with the main term near the front
• One H1, matching the promise of the title
• URL slug short and readable, words separated by hyphens
• Image alt text describing the image, not stuffed with keywords
• Internal links to two or three genuinely related pages
• Meta description under 160 characters, written to earn a click rather than to rank

On keyword density: there is no target number, and there has not been for years. Write naturally, use the term where it belongs, and use variations. If you find yourself counting, you have already lost the plot.

Step 7: Internal linking is a system, not an afterthought

A single post rarely ranks for a competitive term. A group of related posts, all linking to one another and to a central page, has a much better chance.

The shape is pillar and spoke. One broad page covers the whole topic. Several narrower posts each cover one part in depth. Every spoke links up to the pillar, the pillar links down to every spoke, and related spokes link across to each other.

Practically, when you publish a new post: add two or three links from it to existing pages, then go back and add one link from an older relevant post to the new one. That second half is the step everyone skips, and it is why most blogs are a pile of orphans rather than a cluster.

Step 8: The edit pass

Write the draft, then leave it. Come back and cut.

The reliable rule is to remove about 20 per cent. What goes first:

• Every sentence that restates the previous sentence
• Hedging: "arguably", "it could be said that", "in many cases"
• Adjectives doing a number's job — replace "significantly cheaper" with the figure
• The introduction paragraph you wrote before you knew what the post was about
• Anything you included because a competitor had it, not because it helps

If a paragraph could be deleted without the reader noticing, delete it.

Step 9: Using AI to draft without sounding like it

Being direct, since most people now use it. AI is a decent research assistant and outliner, and a poor final drafter. The tells are recognisable and readers notice.

What works:

• Use it to gather questions, structure an outline, or challenge your draft
• Give it your real numbers, prices and examples — it cannot invent those and should not try
• Rewrite the output in your own words rather than editing around it

What produces obviously generated text:

• Publishing a first draft
• Letting it write your opening paragraph
• Keeping its vocabulary: delve, leverage, robust, seamless, landscape, game-changer
• Rule-of-three sentences repeated as a tic
• Confident claims with no source, which is where invented statistics come from

The failure mode that actually costs you is not style, it is accuracy. Anything the model asserts as fact needs checking before it goes out under your name.

Step 10: Publish, then watch

Publishing is the middle of the process, not the end.

On the day. Add Article and FAQPage schema. Submit the URL in Search Console. Add the internal links from older posts.

At 30 days. Check it is indexed. Look at Search Console for the queries it is picking up — often not the ones you targeted, which is useful information.

At 60 days. If it is ranking positions 8 to 20 for something, that is your cheapest win available. Improve that section and add internal links pointing at it.

At 90 days. Decide honestly whether it worked. If it has no impressions at all, the topic was wrong, not the writing.

When to update instead of writing new

Updating an existing post usually beats writing a new one when:

• The post already ranks somewhere on page two
• The information has dated, especially prices and platform features
• Several thin posts cover the same topic and should become one strong page

Write something new when the query is genuinely different from anything you have, not when you want a fresh publish date. Republishing the same content with a new date fools nobody.

The pre-publish checklist

Frequently asked questions

How long should a blog post be?

Long enough to answer the question and no longer. Check the top-ranking pages for your term and aim to be comparably thorough. Length is a consequence of covering the topic properly, never a target to hit, and padding is easy for both readers and Google to spot.

How many blog posts should I publish per month?

Two genuinely useful posts beat eight thin ones. For most small businesses two to four a month is sustainable and enough, provided they form a cluster around related topics rather than being scattered across unrelated subjects.

Does keyword density still matter for SEO?

No. There is no target percentage and has not been for years. Use your term naturally where it belongs, include variations, and write for the reader. If you are counting occurrences, you are optimising for a system that stopped existing a long time ago.

Should I write the blog post or hire someone?

Write it yourself if you have the time and the subject knowledge, because your first-hand detail is the part nobody can copy. Hire out the research, structure and editing if consistency is the problem, which for most business owners it is.

How long before a new blog post ranks?

Expect indexing within days to a few weeks, first impressions on long-tail terms in three to six weeks, and a settled position around month three or four. Competitive terms take considerably longer, and a post with no impressions at 90 days usually means the topic was wrong.

Where to start

Take the next post you were planning to write and run steps one and two on it before drafting. Check whether the term is winnable and read the SERP. That half hour will save you from writing something that was never going to rank.

If keeping this up every month is the problem rather than knowing how, that is what our content marketing service does, and the free audit will show you which topics you already have a foothold in.