I have held three rankings across two pages. All three came from one word in the title

I write a blog for a very small browser extension. Most of what I publish does not rank on anything, which is normal and which I expected. What I did not expect was that the three rankings I have ever held would have one thing in common, and that it would be the same thing in all three cases.

Three rankings, two pages, because one page holds two of them. I nearly wrote three pages in this sentence and had to go back and count, which tells you something about how easy that particular mistake is. Two of the three still stand this morning. The third does not, and that one turned out to be the useful half of the story.

It is one word in the title. Or rather, one qualifier.

The three

Here are the three queries where a page of mine has come back in the top ten, each read twice, in a fresh browsing context, minutes apart:

The query contains	The title contains
without a signature	without a signature
without oauth	without oauth
whatsapp style	whatsapp style

Three for three. In each case the distinctive part of the query, the bit that narrows it from a generic topic to a specific question, appears verbatim in the title of the page.

That is the whole finding. It is not sophisticated and I am fairly sure it is not novel to anyone who does this professionally. It was novel to me because I had assumed something else was doing the work.

What I assumed instead

I had assumed the ranking came from the content. My articles are long, they are specific, they contain measurements nobody else has published because I made them myself. That felt like the kind of thing a search engine would reward.

The counter example is the one that convinced me. I have an article whose title is a figure of speech. It says something like what X actually gives you, and what it does not. It is arguably my most useful article. It is indexed: it comes back at rank one when I search its own exact title, so the page is retrievable, the engine has it, there is no quarantine and no technical fault.

It ranks on none of the five queries it was written for.

Those five queries all contain a specific word, the name of the feature people are looking for. My title contains a metaphor instead. The page is findable and it is not found, and the only difference I can see between it and the three that work is the qualifier.

The one I lost, which is the interesting half

One of the three rankings above is a past tense. I measured it first on 31 July, at rank two, confirmed it on two separate readings, and used it as a control to validate my own measuring script. Four days later it is gone. Not lower. Absent from the ten results, on two readings.

Nothing happened on my side. The page was not edited, moved, redirected or deduplicated. Two established editorial sites now occupy positions on that query which they did not occupy before. Whether one caused the other I have no way to know, and I am not going to pretend otherwise.

Here is why I am keeping it in the list anyway. The same page now ranks first on a different query, of exactly the same shape: generic topic plus a precise qualifier that the title carries word for word. The recipe did not stop working. One query changed hands.

That distinction matters more than it sounds. If I had only tracked the first query, I would have concluded that my approach had failed. If I had only tracked the second, I would have concluded that everything was fine. Both conclusions would have been wrong, and the difference between them is that I happened to measure more than one thing.

The part that should keep me honest

Read the table again and notice what those queries are.

Whatsapp style read receipts is the exact phrase from my own title. The qualifier without oauth is my own framing of the problem, one I use on my site. These are not queries that a stranger types after deciding they need a tool. They are queries that describe my product almost as closely as my product name does.

Meanwhile, on the two head terms of my category, the obvious phrases anyone would type, I am absent. Not low. Absent from the first ten results, checked with a positive control in the same session so the silence means something.

So the honest version of my finding is narrow: when a query is already shaped like my product, having the qualifier in the title is what makes the page come back. I have no evidence at all that the same move wins a competitive head term, and I would not believe me if I claimed it.

A measurement caveat I only found by repeating myself

I read each query twice, and I am glad I did, because one thing was not stable.

On the first query, the first reading returned two of my pages in the top ten. The second reading, about fifteen minutes later, returned four. Nothing changed in between except the reading itself.

The rank one position reproduced both times. The count of my pages in the top ten did not.

So a position is a measurement and an occupancy is an estimate. If you write "I own five of the top ten results", you are describing one reading, not a state of the world. I had written exactly that sentence about an hour earlier, and I have since gone back and marked it as what it is.

This is the same discipline as requiring a positive control, arriving from a different direction: do not treat a single reading as a fact, and do not report a derived number with the same confidence as the thing you actually observed.

What I do differently now

The qualifier goes in the title, not in the subtitle. If the query I want contains a specific word, that word is in the title or the title is wrong. My metaphor titles have become subtitles and first paragraphs, where they still do their job.

One query is not a measurement of an approach. Track several. A single lost ranking looks like failure and a single held one looks like success; only the set tells you which.

Re-read before you write it down. Twice, in a fresh context, especially before letting a number turn into a rule. I built a writing rule on two single readings this week, then went back and re-read both before letting it stand.

And keep the scope of the claim as small as the evidence. Two pages, three rankings, one engine, one product, all on queries that already describe me. That is what I have. It changed how I title things. It does not make me an authority on anything.

Disclosure

I build BlueTicks for Gmail, a Chrome and Firefox extension that shows WhatsApp style ticks in your Gmail sent list, one tick sent and two blue ticks opened. It costs 4 dollars a year and there is a free tier. It is the product whose pages I am measuring throughout, which is the only reason I have this data at all. You can find it at blueticks.io.

I have not named the search engine, the platforms or the queries in full, because a sample of two pages on one engine is enough to change my own titles and nowhere near enough to tell anyone else what to do. Run it on your own pages. Query the things you actually want to rank for, twice, and look at what your titles contain.