Up 457%: How to Tell a Real Trend From a Dead End

Somewhere in your marketing report there is a slide headed "trending keywords", with percentages next to them. It is the most persuasive slide in the deck and the easiest one to get wrong, because a percentage tells you a direction and nothing about the size, the buyer or the money.

We ran our own recommendation engine against singrank.com on 21 September 2026, over a 28-day window. The top item in its "write this next" queue was a search for another rank-tracking tool's login page, up 457% week on week, with 233 impressions and zero clicks.

We are not writing that article. This is the test we used to decide, and it works on anybody's report.

What our own tool actually recommended

Eight queries came back in the "write something new" bucket. Two of them were somebody else's brand: one a rival tool's login page, one another company's name with "seo" after it. Neither could ever end in an enquiry for us, no matter how well we ranked.

The tool was honest about its own limits. Every row in that bucket carried the label "CANDIDATE, NOT A VERDICT", with a note that the match is keyword-only and should be verified before anyone writes. That label is doing real work. Plenty of reports drop it and present the same list as a plan.

The trending bucket showed ten queries, and all ten had zero clicks in the window. Nine of the ten sat below average position 30. The rises were large: 457%, 144%, 133%, 133%, 131%, 129%, 116%, 115%, 86% and 40%.

Why a good tool still surfaces a bad idea

Because momentum is loud and intent is quiet. Our engine scores each query as impressions multiplied by how much room there is to climb, multiplied by momentum, multiplied by an intent weight. It already discounts brand-research queries heavily, weighting them at 0.05 against 1.0 for transactional intent.

The rival's login page still came first. A big enough impression count multiplied by a big enough percentage rise will overwhelm a small weight. That is arithmetic, not a bug, and every scoring model in this category has the same property.

So the tool did its job and produced a ranked list. Deciding what a query means for your business is still a person's job, and it takes about four questions.

Question 1: Is this query about you, or about somebody else?

Read the words. If the query contains another company's brand name, or a product's login, pricing, support, alternatives or reviews page, the person typing it is doing business with somebody else.

Two of our eight write-new candidates failed this immediately. You can rank for a rival's login page. People arriving will be annoyed that you are not the login page, and they will leave.

There is one honest exception. A comparison page aimed at people already shopping for a category can be legitimate, and it converts. The line is whether you can genuinely answer the question. Being another company's login screen is not a question we can answer.

Question 2: What is the base under the percentage?

Ask for the two absolute numbers, not the percentage. A rise from 4 impressions to 22 is 450%, and it is 18 impressions.

Our own 457% sat on 233 impressions in 28 days, which is roughly eight a day. That is real but modest, and it would be worth exactly nothing if those eight people a day are looking for a login screen.

Any report that shows you a percentage without the before-and-after count is asking you to feel something rather than decide something. It is a reasonable thing to ask for and it takes ten seconds to produce.

Question 3: Where are you actually ranking?

Position and visits are not the same measurement, and neither are impressions and eyeballs.

Google's Search Console documentation states that "an impression is counted whenever an item appears in the current page of results, whether or not the item is scrolled into view". So an impression is a chance to be seen, not proof of being seen. The same page explains that the position shown is "the topmost position occupied by a link to your property or page in search results, averaged across all queries in which your property appeared", which means an average position of 40 can hide a page that is sometimes ninth and usually nowhere.

Nine of our ten trending queries sat below position 30. A rise in impressions at position 50 is a rise in the number of times nobody saw us. It is a signal that the query is growing, which is useful, and it is not a signal that our page is working. Our guide to rank tracking in Singapore covers how to read the report itself.

Question 4: Would answering it bring someone who can buy?

Say out loud what the person wants and what happens next. If the honest answer is "they read the article and go back to Google", the query is a traffic idea, not a business one.

This is where a lot of plans quietly fail. Informational demand is always larger than commercial demand, so an unweighted list will always be led by questions that are popular and not saleable. Our engine's own method note says it ranks commercial demand first on purpose, "because informational queries carry far more volume and would otherwise always win the queue".

If you cannot see the path from the article to an enquiry in one step, put it below the queries where you can.

What page one with zero clicks is telling you

We had six queries sitting between average position 3.7 and 9.9. Together they collected 127 impressions and zero clicks in 28 days.

Page one with no clicks usually means one of three things, and they need different responses. Either the query is tiny, and page one of a query nobody types is worth nothing. Or your title and description do not match what the searcher wanted, which is an afternoon's work. Or the answer is fully visible in the result itself, in an AI overview or a snippet, and the click was never going to happen.

Telling those three apart takes one look at the impression count and one look at what occupies the page above you. Nobody needs a tool for it.

What we would do with a report like ours

In this order:

1. Strike every query containing another company's brand, unless you are writing a genuine comparison you can stand behind.
2. Ask for absolute before-and-after counts next to every percentage. Delete any row where the rise is under about 50 impressions a month.
3. Sort what is left by whether the searcher could become a customer, not by volume.
4. Look at the page-one-with-no-clicks rows separately. Those are title and description fixes, not new articles.
5. Then pick two or three to write properly, instead of twenty to write thinly.

Steps 1 to 4 take an hour with the report you already have, and they are free. We wrote about how to judge the agency producing it in how to judge an AI SEO agency before you pay.

What this costs, including the part against us

The cost of this test is that your content plan gets shorter, and a shorter plan is harder to sell and less impressive to look at. Twenty articles a month reads like progress. Three articles aimed at people who can buy reads like a light month until the enquiries arrive, and it may take a quarter before anyone can tell the difference.

Here is the part that costs us directly. We sell content. A method that deletes most of a content plan is not in our commercial interest, and we are publishing our own engine's top recommendation being a rival's login page rather than quietly filing it. If applying these four questions leaves you with nothing worth writing, the honest conclusion is that content is not your bottleneck this quarter, and we have written separately about the times SEO is the wrong thing to buy.

There is also a case against us in the other direction. If you are about to start a business rather than grow one, none of this applies yet, because you have no report. The question then is whether the demand exists at all, which we covered in checking whether anyone is searching before you incorporate.

How we got these numbers, and what they cannot show

• Date and source. 21 September 2026, a 28-day window, from Google Search Console data for singrank.com read through our own content-targets tool. This is our own site's data, not a client's.
• Small numbers. The largest figure here is 233 impressions. Nothing in this report is a large sample, and we are not presenting it as one.
• Week-on-week percentages are volatile on counts this small, which is exactly the point being made rather than a defect in the data.
• Intent labels are the tool's judgement, applied by keyword matching. We read all eight write-new rows by hand and would have labelled three differently: two as navigation to another company rather than informational, and one as commercial rather than local.
• Zero clicks is a fact about our own site, not a claim about anybody else's. A different site at the same positions may well get clicks.
• No company is named. Two candidate queries carried other companies' brand names. We counted them and did not repeat them.

One next step

Send us your last monthly report, whoever produced it. We will run these four questions against its keyword list and tell you which rows we would delete and why, with reasons you can check rather than a verdict. That starts as a free visibility check through our contact page.

If the list survives the test, that is a useful answer too, and it costs you nothing to find out.

FAQ

What does a trending keyword percentage actually mean?

It means impressions for that query rose by that percentage against the previous window. It says nothing about the size of the numbers, your ranking, or whether the searcher could become a customer. A 457% rise on a small base can be a few dozen extra impressions, so always ask for the two absolute counts.

Should I write content for a query containing a competitor's brand name?

Only if you are writing a genuine comparison you can stand behind and the reader is shopping for the category. A search for a rival's login, password reset or support page is somebody already using that product. You can rank for it and you will not convert it.

Why does my site rank on page one and get no clicks?

Usually one of three reasons: the query is too small for page one to matter, your title and description do not match the intent, or the answer is already visible in the result page itself. Check the impression count first, then read what sits above your result.

Is an impression the same as someone seeing my listing?

No. Google's Search Console documentation says an impression is counted whenever an item appears in the current page of results, whether or not it is scrolled into view. It is a chance to be seen. Average position is also averaged across queries, so it can hide a wide spread of real positions.

How many keywords should a monthly content plan have?

Fewer than most plans do, in our view. We would rather see two or three pages aimed at people who can buy than twenty written thinly. The test is whether you can describe, in one sentence, how a reader of each page becomes an enquiry.