One Question, Nine Phrasings: Do You Need Nine Pages?

One page. Nine rows on a keyword report that all ask "what does SEO cost in Singapore" are one question, and one question is one page.

We can show you that from our own Search Console rather than from a principle.

Nine of our own top queries are one question

We pulled our top 60 queries by impressions on 12 October 2026. Nine of those 60 rows ask the same price question in nine wordings, and between them they drew 5,769 impressions and no clicks at all.

Query	Impressions	Average position
seo pricing singapore	1,432	44.0
seo price singapore	1,114	48.4
seo singapore price	1,091	48.0
seo pricing	516	59.1
seo services pricing singapore	514	54.1
seo cost singapore	434	42.8
seo rates	269	39.1
seo cost	216	48.3
seo packages singapore	183	39.9

A second group is larger. Sixteen more of the same 60 rows ask who to hire: "seo agency singapore", "best seo agency singapore", "seo company singapore", "singapore seo agency", "top seo agency singapore", and eleven further arrangements of the same four words. Those 16 drew 6,493 impressions and one click.

Put the two groups together and 25 of our top 60 queries are two questions, asked 25 ways, worth 12,262 impressions and one click between them.

Why the split happens anyway

Because a keyword report pays by the row, and nobody is paid for deleting rows.

An agency that hands you nine rows looks like it found nine opportunities. An agency that hands you one row and says "this is one page" looks like it did less work. If the retainer is priced per article, as ours is, the incentive runs the same way: nine pages bill more than one. We sell programmes of 10 and 20 articles a month, so read our advice here against that fact.

The second reason is more sympathetic. Splitting feels safer. If you are not sure which phrase Google will favour, writing a page for each looks like covering the ground. It is the opposite. Two pages aimed at one question split the links, the mentions and the internal anchors that would have made one page strong, and Google then picks between them.

The rule: does the answer change?

Write the first paragraph of the answer for each phrasing. If you would write the same paragraph twice, it is one page. If the paragraph changes, it is two.

That is the whole test, and it is quicker than any tool. Run it on our own nine rows. "What does SEO cost in Singapore", "what are SEO prices in Singapore", "what are SEO rates": same paragraph, same numbers, same caveats. One page. Now try "seo packages singapore" against "seo cost singapore". The first wants a list of what is bundled at each tier; the second wants a number and what moves it. Those are different paragraphs, which is why that row is the one row of the nine we would consider splitting out, and only if the packages answer runs long enough to stand alone.

Three signals that the answer really has changed:

• The number is different. A commercial cleaning price and a home cleaning price are two pages. A "cleaning price" and a "cleaning cost" are one.
• The buyer is different. An employer searching for a recruiter and a candidate searching for a job are two pages, even when the words barely differ.
• The next step is different. If one phrasing ends in a quote request and the other in a download or a phone call, the pages do different jobs.

If none of those three has changed, the phrasings belong in one page, in its headings and its FAQ, where they can be answered without being re-answered.

Where we broke our own rule

Three times, by our own measurement, in the last 28 days.

Our cannibalisation check found three queries where two of our pages both draw impressions for the same question. On "seo for specialist clinics", two of our pages split 73 impressions, one sitting at position 8.4 and the other at 27.7. On "ai seo results singapore", two pages split 52 impressions at positions 19.3 and 28.1. On "healthcare seo consultant", two pages split 22 impressions at 3.9 and 11.6.

None of those pairs was written to compete. Each was written months apart, by someone who had a good reason at the time, and the overlap only became visible when we measured it. That is the ordinary way this happens on a site that publishes often, including ours.

There is a second number from the same morning that makes the point more plainly. When we asked our own system what to write next, it matched 226 queries to pages we already have and found 7 genuine gaps. Thirty-two reasons to improve a page for every one reason to write a new one. If your site has been running for a year or more, your ratio probably looks similar, and your agency's content plan probably does not.

What Google says it does with near-duplicate pages

It picks one and consolidates the rest, and it would rather you told it which.

Google's guidance on consolidating duplicate URLs says that without a stated preference, "Google will identify which version of the URL is objectively the best version to show to users in Search". It calls rel="canonical" "the preferred solution" for choosing within your own site, advises against noindex for that job, says permanent redirects should be used "only when deprecating a duplicate page", and tells you to link internally to the canonical URL rather than to the duplicates.

Read that as an owner and the sequence is clear. Google is going to choose. You can choose first, keep the signals on one URL, and point your own internal links at it. Our wider piece on how many pages a website actually needs covers the other half of this question, which is how long the list of real questions is in the first place.

What to do this week

Four steps, in order. An afternoon for a small site.

1. Export your Search Console queries for the last three months and sort by impressions. Read the top 50 aloud. Grouping is obvious when you hear it, and invisible when you scan a spreadsheet.
2. Mark the groups. Rows that would get the same first paragraph go in one group. Our nine price rows would have taken two minutes.
3. For each group, find which of your pages Google already shows. If it is two, keep the one with the clicks, canonical the other to it, and change the internal anchors to point at the keeper.
4. Only write a new page for a group with no page at all. For everything else the job is to improve a page you already own, which is also the cheaper job.

One warning on step 3. A page with no clicks is not automatically the loser; a page at position 8 with no clicks may simply have a title that no one taps, which is a different problem and one we have written about in impressions without clicks. Check the title before you canonicalise anything away.

What merging costs you

It costs you the keyword report. Nine rows become one line in a plan, and if you report to a board or a parent company, a shorter plan is harder to defend than a longer one. That is a real cost and we will not pretend otherwise.

It also costs you a small, genuine chance. Occasionally a second page does catch a phrasing the first one misses, and you only find out by having both. Our view is that the chance is small and the split is reliably expensive, but it is not zero.

And the honest one: our own nine price rows sit between position 39 and 59 with no clicks, and we have three cannibalised queries on the books. We are not writing this from the winner's side of the table. We are writing it because merging is the fix we are applying to ourselves, and our SEO, AEO and GEO pricing page is the page that has to carry all nine of those phrasings.

One next step

Send us the top 50 queries from your Search Console and we will group them the way this piece describes, and tell you which groups already have a page, which have two, and which have none. It starts as a free visibility check through our contact page, and if you want the merging and the writing done for you, our SEO, AEO and GEO programme starts from S$500 a month.

What we will not promise is that consolidating lifts a ranking. It removes a known problem; the rest is competition.

FAQ

Should I create a separate page for every keyword?

No. Create a page for every distinct answer. If two phrasings would get the same first paragraph, the same number and the same next step, they are one question and one page, with the variants living in the headings and the FAQ. Nine of our own top 60 queries are one question.

How do I tell whether two keywords need two pages?

Write the opening paragraph of the answer for each. If the number changes, the buyer changes, or the next step changes, write two pages. If none of those changes, write one. The test takes a minute and does not need a tool.

What is keyword cannibalisation, in plain terms?

Two pages on your own site drawing impressions for the same question, so the links and mentions that would have made one page strong are split between them. Our own check on 12 October 2026 found three such queries, including two pages splitting 73 impressions at positions 8.4 and 27.7.

Should I delete the weaker page?

Usually not. Google says rel="canonical" is the preferred way to choose between near-duplicates on one site, and that permanent redirects are for deprecating a page outright. Keep the content, point the canonical and your internal links at the keeper, and fold anything unique into it first.

If one page answers nine phrasings, is that page keyword stuffing?

No, as long as the phrasings appear where they are genuinely answering something: a heading that asks the question the way a buyer asks it, with the answer underneath. Repeating a phrase in body text without answering anything is the practice worth avoiding, and it does not help.

Figures describe SingRank's own Search Console data as reported on 12 October 2026, and our own cannibalisation check over the preceding 28 days. Your site's ratios will differ.