Use suggest keywords by grouping them by answer shape — definition, how-to, tool, comparison, improvement — and covering each shape with one page. One article per term makes your own pages compete. Suggestions are not ordered by search volume, so decide what to write by four checks: overlap with existing pages, what you can add, volume, and the shape of the top-ranking pages.
What Suggest Keywords Tell You — and What They Don't
Suggest keywords are the predictions (autocomplete) Google shows as you type in the search box. Google says they reflect real searches, adjusted for language, location, and trending topics. To use them well, separate what they tell you from what they don't.
| Question | From suggestions | Note |
|---|---|---|
| The phrasings people actually search | Yes | Google says predictions are based on real searches |
| What people want alongside the term | Yes | "what is", "how to raise" and so on show the kind of answer wanted |
| How often each is searched, or what the order means | No | Google says predictions are not simply the most common queries |
| Whether you can rank | No | Judge by looking at the pages that rank now |
| Whether it overlaps your existing pages | No | Check your page list and Search Console |
In short, suggestions tell you what people want to know, not which terms matter. Importance is decided by the grouping and the four checks that follow.
Pulling Them for Real: 47 Requests for "ドメインパワー" (Domain Power)
On October 15, 2026, we sent 47 requests to Google's suggestions using "ドメインパワー" (domain power) as the seed: the term itself, plus the term followed by each of the 46 hiragana characters. Appending one character at a time is how most bulk tools work, including Direbase's keyword suggest.
- 35 of the 47 requests returned nothing. After removing duplicates, 19 terms remained.
- The same day, "seo 診断" (with a space) returned zero predictions, while "seo診断" (no space) returned five. How you write the seed changes what you get.
The 19 terms
ドメインパワー とは / ドメインパワーとは / ドメインパワー 意味 ない / ドメインパワー 確認 / ドメインパワー 確認方法 / ドメインパワー 調べる / ドメインパワー 調べ方 / ドメインパワー 計測 / ドメインパワー 無料 / ドメインパワー 計測 無料 / ドメインパワーチェック 無料 / ドメインパワーチェックツール / ドメインパワー計測ツール / ドメインパワー ahrefs / ドメインパワー moz / ドメインパワー ラッコ / ドメインパワー 上げ方 / ドメインパワー 目安 / はてなブログ ドメインパワー
How many you get varies widely by seed, and a short list does not mean low demand. The grouping that follows works the same whether you have dozens of terms or hundreds.
Group by Answer Shape
The usual way to group suggestions is the four search intent types (Know, Go, Do, Buy). When deciding what to write, though, it helps to go one level finer: what shape of answer would satisfy the searcher. Here are the 19 terms grouped that way.
| Answer shape | Terms | Fitting page | Where we answer it |
|---|---|---|---|
| Definition | とは (what is) / 意味 (meaning) | An explanation of the term | Guide article (what domain power is) |
| Doubt / objection | 意味 ない (pointless?) | A piece that answers with evidence | Lab article (fact-check) |
| How-to | 確認 / 確認方法 / 調べる / 調べ方 / 計測 (check, how to check, measure) | A page that walks through the steps | Tool page (you can check right there) |
| Looking for a tool | 無料 / チェックツール / 計測ツール (free, checker, tool) | The tool itself | Tool page |
| Comparison / brand | ahrefs / moz / ラッコ | How it differs from other vendors' metrics | Tool page FAQ and the guide's comparison section |
| Benchmark | 目安 (what's a good score) | How to read the number | Tool page benchmark table and the guide |
| Improvement | 上げ方 (how to raise it) | A prioritized list of actions | Guide article (how-to-raise section) |
| With context | はてなブログ (on Hatena Blog) | An answer narrowed to one platform | Not written (see the decision below) |
Nineteen terms collapse into eight shapes. "Check", "how to check", "look up", "how to look up" and "measure" differ only in wording; the answer they want is the same. Count shapes, not terms.
Answer Each Shape With One Page
On this site, domain power is answered by three pages: the tool page, a guide article on the term, and a fact-check article. The last column of the table above is that split. In Search Console for the 28 days from September 17 to October 14, 2026, "how to look up", "checker tool", "how to raise" and "what's a good score" all drew their impressions to the tool page, at average positions of 10 to 17.
It shows one page can surface for many phrasings without a page per term. Make a page per term instead, and your own pages compete for the same query, leaving each one stuck in the middle.
Google's spam policies list, as an example of scaled content abuse, creating many pages that contain search keywords but offer little or no value to readers. Turning each suggestion into its own article title moves toward exactly that.
Four Checks Before Writing a Term Up
For each answer shape, run these four checks in order. If any answer is no, don't create a new article — add to an existing page or skip it.
1. Is the answer shape different from your existing pages?
If it is the same shape, don't make a new page — add a section or FAQ to the existing one. Separate pages end up competing with each other for the same term.
2. Can you add information or judgment only you can?
If you can only repeat what the top pages say, there is no reason to write it. Pick terms where you can add tested results, numbers, or decision criteria.
3. Did you check search volume another way?
Suggestions don't show volume. Use a search volume tool or impressions in your own Search Console to confirm the term is worth targeting.
4. In what shape do the top pages answer?
Search it and look at the top 10. Publishing an explainer for a term where tools fill the results does not match what searchers want.
The last row, "domain power on Hatena Blog", is not written up here because it fails the second check: we have no tested results from running a Hatena Blog, so there is nothing to add over the top pages. For someone who does have that experience, it is a promising term.
Common Mistakes
One article per term
"How to check", "how to look up" and "check" want the same answer. Splitting them makes your own pages compete, and edges toward what Google's spam policies describe as creating many pages that contain search keywords but mean little to readers.
Cramming everything into one article
The opposite failure: showing the same page to someone looking for a tool and someone doubting the metric serves neither. Separate what has a different answer shape.
Throwing away doubt/objection suggestions
Terms like "pointless?" are exactly where readers get stuck. Addressing them head-on is what sets an article apart.
Trying only one spelling
As measured below, a single space changes the results. Try the seed two or three ways — with and without spaces, in different scripts.
Researching while signed in
Google says predictions can reflect your search history when signed in. Research signed out or in a private window so you don't collect suggestions skewed toward you.
You Can Pull Them by Hand — a Tool Only Shortens the Pulling
Typing "ドメインパワー あ", "ドメインパワー い" and so on into the search box collects the same terms without any tool. It just takes 47 entries and de-duplication.
Direbase's keyword suggest does the pulling and de-duplication from one entry, and paid plans show search volume alongside (Basic plan and up). But a tool only shortens the pulling. Grouping by answer shape and running the four checks is still your call, as described here.
Related Tools
Sources (checked October 15, 2026)
- Google Search Help — How Google autocomplete predictions work
- Google Search Central — Spam policies for Google web search
The suggestions shown were pulled on October 15, 2026. Predictions change with time, location, and sign-in state.
FAQ
Are suggest keywords ordered by search volume?▾
If a term returns few suggestions, does that mean there is no demand?▾
Should I write one article per suggestion?▾
Is it fine to research suggestions while signed in?▾
What should I do with negative suggestions like "... pointless"?▾
Once You Have the Terms, Check the Page That Answers Them
Pull suggestions in bulk, group them by answer shape, then check that the page you assign them to has its title, headings, and structured data in order.
Use keyword suggest
Written by
今井政和SEO Director / Frontend Developer
SEO Director with 20+ years of web industry experience. Creator of Direbase and the official WordPress plugin "ORECTIC SEO CHECK." Author of a book on web strategy inspired by Edo-era merchant principles.
@imai_directorRelated Articles
How to Build an SEO Article Outline | Extract and Compare the Top Pages' Headings
Extract the top pages' headings, count how many cover each topic, and decide what to match and where to differ — with a real comparison of one of our own articles.
Google May 2026 Core Update | What Google Officially Stated, Backed by CodeQuest.work's Real Data
Google rolled out the May 2026 core update on May 21. The official statement was only three lines. We cross-check 'no specific recovery action' and 'a ranking drop isn't necessarily a problem' against CodeQuest.work's measured data on structured data and AIO citations.
Can Low Domain Power Still Rank? | Beating a High-DA Site with One Page
Even against a high-Domain-Power competitor, a single page can overturn rankings for a specific query through search-intent fit and depth. From an observed reversal against a far stronger legacy site, we dissect what worked and what it does not prove — the difference between a query-level win and site-wide superiority.
What Is Search Intent? The 4 Types, How to Identify It & Apply It in SEO
Search intent is the goal behind a user's query. Learn the four types (Know/Go/Do/Buy), how to identify intent from the SERP, how to diagnose content mismatch, and what search intent means in the AIO era.