When a page doesn't rank, the most common advice a business owner hears is "make it longer." So they add 800 more words. They add a history section nobody asked for, a few paragraphs of definitions, and a long list of "tips." The page gets heavier. It rarely gets better.
Length is not the goal. Completeness is. A complete page answers the questions a searcher actually has, in the order they have them, and then stops. Sometimes that takes 900 words. Sometimes it takes 3,000. The difference is not padding. It is subtopics.
In this guide, let me explain how I find the subtopics a query expects, how I decide which ones to include and which to leave out, and how I add the one thing competitors are missing. You will get a simple grid you can build in a spreadsheet, a three-bucket rule for cutting, and a checklist to use before you publish.
What a "complete" page means to Google (it isn't word count)
Let me clear up the word count myth first. Google's SEO Starter Guide lists content length in its myths section. In Google's own words:
“The length of the content alone doesn't matter for ranking purposes.”
— Google Search Central, SEO Starter Guide
Google's guide to creating helpful, reliable, people-first content goes further. It lists "writing to a particular word count" as a sign of search engine-first content. But in the same document, it asks a question that every writer should take seriously: "Does the content provide a substantial, complete, or comprehensive description of the topic?"
So Google wants completeness, not length. Those two ideas only look the same from a distance. Here is how I tell them apart:
| Padding | Real subtopic coverage |
|---|---|
| Repeats the main point in new words | Answers a separate question the reader has |
| Adds history or definitions nobody searched for | Adds steps, conditions, costs, risks or choices the reader needs |
| Could be cut without the reader noticing | Would leave the reader with an unanswered question if cut |
| Written to hit a number from a tool | Written because the search results and real questions show demand |
| Usually sits at the top, delaying the answer | Sits where the reader would naturally ask it |
If you are new to the idea of matching what searchers want, our beginner lesson on search intent is worth five minutes before you continue.
Why missing subtopics cost you rankings in 2026
There are two reasons subtopics matter more now than they did a few years ago.
First, relevance is judged across the whole topic. Google understands related words and ideas, not just your main keyword. A page about "how to descale a coffee machine" that never mentions vinegar, citric acid, how often to do it or what to do with a machine that has a built-in descaling program looks less complete than one that covers all four. My guide to semantic SEO explains the meaning side of this in more depth.
Second, AI search breaks questions into pieces. When Google announced AI Mode in May 2025, it explained a technique called "query fan-out." According to the Google blog post on AI Mode, the system breaks a question down:
“into subtopics and issuing a multitude of queries simultaneously on your behalf”
— Google, AI in Search: Going beyond information to intelligence
Think about what that means. If the AI looks for an answer to each subtopic separately, a page that covers more of those subtopics clearly has more chances to be found and used. A page that covers only the headline question has fewer. I treat this as direction, not a proven formula, because Google doesn't publish how it picks sources. But the logic is hard to ignore.
Reading the top ten results for the subtopics they share
The fastest way to learn what a query expects is to look at what already ranks. Ahrefs' article on keyword relevance makes the case well. It notes that Google tends to favor what it already ranks, which is why the top results often look so similar. Its advice is simple and, I think, correct:
“Make your content relevant before you make it unique.”
— Ahrefs, Keyword Relevance: What It Is, and How to Demonstrate It to Google
In other words: first make sure you cover what the searcher expects. Then add your own angle. Here is how I would do the first part.
- Search your main keyword in a private browser window, set to the country you target.
- Skip anything that isn't a direct competitor: videos, forums and big marketplaces can teach you about intent, but their headings won't help much.
- Open the top 5 to 8 pages that match your page type. If you are writing a guide, look at guides. If you are writing a service page, look at service pages.
- Copy every H2 and H3 from each page into a spreadsheet, one row per heading, with the page URL in the next column. A browser extension that lists headings makes this quick, or you can use a crawler.
- Rewrite each heading as a plain question. "Descaling frequency" becomes "How often should I descale?" This makes it easy to spot that two pages are covering the same thing under different names.
- Group the questions and count how many of the pages cover each one.
Don't just read headings, though. Some pages cover a subtopic inside a paragraph with no heading at all. Skim the body of the top three pages too.
The subtopic grid
The output of those steps is what I call a subtopic grid. Each row is a subtopic. Each column is a competitor. You tick the boxes. Here is a short, hypothetical example for "how to descale a coffee machine":
| Subtopic (as a question) | Page A | Page B | Page C | Page D | Page E | Count |
|---|---|---|---|---|---|---|
| What is scale and why does it build up? | Yes | Yes | Yes | Yes | - | 4/5 |
| How often should I descale? | Yes | Yes | Yes | Yes | Yes | 5/5 |
| Step-by-step with a commercial descaler | Yes | Yes | Yes | Yes | Yes | 5/5 |
| Can I use vinegar or citric acid? | Yes | Yes | - | Yes | Yes | 4/5 |
| Steps for pod machines vs espresso machines | - | Yes | - | Yes | - | 2/5 |
| Signs your machine needs descaling | Yes | - | Yes | - | - | 2/5 |
| Does descaling void the warranty? | - | - | - | - | Yes | 1/5 |
| History of espresso machines | - | - | Yes | - | - | 1/5 |
Already, you can see the shape of a complete page. We will decide what to keep in a moment.
Mining People Also Ask and autocomplete without drowning in questions
Competitor headings show you what writers think matters. People Also Ask and autocomplete show you what searchers actually type. You need both.
People Also Ask
Google's visual elements gallery calls this feature the related questions group and describes it as:
“a cluster of questions that are related to what the user initially searched for.”
— Google Search Central, Visual elements gallery of Google Search
Here is how I use it. Click on two or three questions in the box. Each click loads more questions. Stop after about 15 to 20, because the questions start drifting away from your topic. Copy the ones that fit your page into the grid as new rows. For a deeper look at the feature itself, see my guide on People Also Ask.
Autocomplete
Type your main keyword into Google followed by a space, then try adding letters (a, b, c...) or words like "how," "can," "vs," "without" and "for." Google's help page on autocomplete explains that predictions reflect real searches, and that they take into account the language and location of the query and trending interest. That makes them a good, free signal of what people want to know.
A word of caution: Google also says predictions are not assertions of fact. They tell you what people ask, not what the right answer is. I go deeper on this method in my guide to mining Google autocomplete for keywords.
Three other places I check
- Your own Search Console data. For an existing page, the Performance report filtered to that page shows queries it already appears for. Queries with impressions but low clicks often point to a subtopic the page touches but doesn't answer well.
- Your sales or support inbox. The questions customers ask before they buy are subtopics competitors often miss.
- Forums and Reddit threads that rank for the query. They show the follow-up questions and frustrations real people have.
Must-cover, worth-adding or cut: the three-bucket rule
Here is where most writers go wrong. They collect 40 subtopics and try to cover all of them. That is exactly how pages become bloated. A good outline is as much about what you leave out as what you put in.
I sort every row of the grid into one of three buckets:
| Bucket | Rule of thumb | What to do |
|---|---|---|
| Must-cover | Covered by most of the top pages (say, 3 or more of 5), or asked repeatedly in PAA | Give it its own section. Missing it makes the page look incomplete. |
| Worth adding | Covered by one or two pages, but clearly useful to your reader; or a question from your own customers | Cover it briefly, or use it as your point of difference. |
| Cut or move | Off-intent, a different search on its own, or something only one writer added to pad the page | Leave it out, or give it a separate page and link to it. |
Two questions help me decide the hard cases:
- "Would the reader leave and search again if this were missing?" If yes, it belongs on the page.
- "Does this subtopic have its own search intent?" If people search for it separately and expect a full answer, it probably deserves its own page, not a squeezed section.
Back to the coffee example. "History of espresso machines" goes in the cut bucket: one page included it and no searcher asked for it. "Pod vs espresso machine steps" goes in worth-adding, because the steps really are different. "Does descaling void the warranty?" is interesting. Only one page covers it, but it is a real worry before someone tries vinegar. I would keep it as a short answer.
Sometimes the cut bucket shows you that you have two pages fighting for one topic, or one page trying to do two jobs. That is a keyword mapping problem, not a writing problem.
Where each subtopic should live on the page
Once you know what to cover, decide how to cover it. Not every subtopic deserves an H2. Here is the rough guide I use:
- H2: a must-cover subtopic that needs more than a paragraph or two. These are the main stops on the reader's journey.
- H3: a variation or detail inside an H2, such as "For pod machines" under the main steps.
- A table: anything that compares options, lists settings or shows how often to do something by condition. Tables cover many small subtopics without many words.
- A short answer block: quick questions, such as the warranty question, grouped near the end.
- A diagram or photo: a process that is easier to see than read. Just make sure it is an image Google can actually see, which I explain in my guide to what Google can index in SVGs, icons and background images.
- A link to another page: a subtopic with its own intent. Use clear, descriptive anchor text, as I explain in my guide to the internal anchor text audit.
Order matters too. Put sections in the order the reader would ask them. For a how-to, that usually means: quick answer, what you need, the steps, variations, problems, then quick questions. If you want a refresher on structure, see my guide on heading tags and page structure.
Write your headings the way people ask. "How often should you descale a coffee machine?" is clearer than "Frequency considerations." It also matches the words in PAA and autocomplete without forcing anything.
The subtopic only you can cover: adding information gain
If every page covers the same subtopics, what makes yours the one to rank? This is where the second half of Ahrefs' advice comes in: relevant first, then unique.
Google's helpful content guide asks: "Does the content provide insightful analysis or interesting information that is beyond the obvious?" and warns against "mainly summarizing what others have to say without adding much value." A page built only from a competitor grid is, by definition, a summary of what others already said.
So after the grid, I add at least one subtopic that no competitor covers well. Good sources for this:
- Your own data: prices, timings, failure rates or results from your work.
- Your own process: the checklist your team actually uses.
- A decision aid: a table or flow that helps the reader choose, when competitors only describe.
- The edge case: what to do when the standard advice doesn't work.
- An honest trade-off: when NOT to do the thing the page is about.
I cover this idea in detail in my article on information gain. In the coffee example, a strong addition would be a small table of how often to descale based on local water hardness, since most competitor pages give one generic answer.
A worked example: rebuilding a thin guide around its subtopics
Let me put it all together with a hypothetical example. Say a kitchen appliance repair company has a 700-word guide called "How to descale a coffee machine." It sits at position 14 for its main keyword.
Here is how the rebuild could look:
- Grid. Five competitor pages give 16 distinct subtopics. PAA adds 6 more questions. Autocomplete adds 4. After merging duplicates, there are 19 unique subtopics.
- Current coverage. The existing page covers 5 of the 19.
- Buckets. 8 are must-cover (3 or more of 5 pages, or repeated in PAA). 6 are worth adding. 5 are cut, including history, coffee bean storage and a "best machines" list that has its own intent.
- Gap. Of the 8 must-cover subtopics, the page is missing 4: frequency, vinegar safety, machine types and signs of scale.
- Information gain. The team adds a water-hardness table and a short "what we see when people skip descaling" section from their real repair jobs.
- Result on paper. The page grows to about 1,600 words, covering 14 of the 19 subtopics, all of them on purpose.
Notice the page roughly doubled, but every new section answers a question someone actually has. That is the difference between completeness and padding. Whether it moves from position 14 depends on many other things, such as links and competition, so I would track it for a few weeks rather than expect overnight change.
Also notice what we did not do: we didn't chase a content tool's target word count or term list. I explain why in my article on whether content optimization scores matter.
Traps that turn subtopic coverage into bloat
A few traps I'd warn you about:
- Covering everything a little. Twenty thin sections are worse than eight solid ones. If you can't say something useful, leave the subtopic out.
- Copying competitor structure exactly. Use their headings to find questions, not to build a clone.
- Merging two intents. "How to descale" and "best descaler to buy" are different searches. One page trying to do both usually does neither well.
- Forcing keywords into every heading. Subtopics bring related words in naturally. Stuffing the main keyword everywhere, including city names on local pages, is its own problem. See my guide on city-name stuffing on local pages.
- Scaling the method with no human input. Running this grid through AI to produce hundreds of near-identical pages is exactly what Google's spam policies describe as scaled content abuse: many pages made mainly to rank, with little value added.
Questions writers ask about subtopic coverage
How many subtopics should one page cover?
There is no fixed number. Cover every must-cover subtopic, add the useful ones, and stop. For a simple how-to, that might be 6 to 10. For a broad guide, it might be 15 or more.
Should every subtopic get its own heading?
No. Big subtopics get H2s, details get H3s, and small questions can sit in a table, a list or a short answer block. Headings should help the reader scan, not just hold keywords.
What if competitors cover a subtopic I think is useless?
Check whether searchers actually ask about it in PAA or autocomplete. If they don't, and it doesn't fit the intent, leave it out. Matching competitors is a starting point, not a rule.
When should a subtopic become its own page?
When people search for it separately and expect a full answer, or when covering it properly would double your page. Give it a short summary on the main page and link to the new one.
Can AI tools find subtopics for me?
They can speed up the grid, especially grouping similar headings. But check their suggestions against real search results, because AI can invent questions nobody asks.
Does covering more subtopics guarantee better rankings?
No. It makes the page more useful and more relevant, which helps. But links, site reputation and how well you write still matter.
A subtopic checklist to run before you publish
| Check | Question to ask |
|---|---|
| Intent | Does the page type match what ranks (guide, list, service page)? |
| Must-cover | Is every subtopic covered by most top pages included? |
| Real questions | Have I answered the relevant PAA and autocomplete questions? |
| Cuts | Did I remove anything off-intent or purely there for length? |
| Separate intents | Did I link out, rather than cram in, subtopics that need their own page? |
| Order | Do sections follow the order a reader would ask them? |
| Format | Did I use tables, lists or images where they say more with fewer words? |
| Information gain | Is there at least one useful thing no competitor offers? |
| Headings | Would each heading make sense to a reader scanning the page? |
If you write briefs for other writers, put the subtopic grid and the three buckets right into the brief. My guide on how to write an SEO content brief shows where they fit.
Where to start with your thinnest page this week
Pick one page that ranks between positions 8 and 20 for an important keyword. Build the grid for it, sort the buckets, and compare. In my view, that one exercise teaches you more about your content than any score or word count target.
In brief: make the page complete for the reader, cut what doesn't help, and add one thing only you can say. That is how a page becomes longer without becoming padded.
If you would rather have experts do the research and the writing, our content writing service builds every article from this kind of subtopic research. You can get in touch for a free quote and tell us which pages you want to improve first.
Further reading on subtopic coverage
- Do Content Optimization Scores Matter?
- Does Bold Text Help SEO? Formatting for Skimmers
- City-Name Stuffing on Local Pages: How Much Is Too Much?
Sources for this subtopic coverage guide
- Google Search Central: Creating helpful, reliable, people-first content
- Google Search Central: SEO Starter Guide
- Ahrefs: Keyword Relevance
- Google, The Keyword: AI Mode announcement (May 20, 2025)
- Google Search Help: How Google autocomplete predictions work
- Google Search Central: Visual elements gallery of Google Search
- Google Search Central: Spam policies for Google web search
