The Answer Grammar of Our Own Guides: 67 Questions, 48 Yes/No, and the 10 Answers That Lead With the Fact

I work on OnePassApps, so this is our own documentation, measured on our own site.

Every one of the 22 app guides we publish carries a "Common questions" section. I have already published two things about that section: that there are 67 question-and-answer pairs across the fleet (21 guides ask three questions, MovieMatcherAI asks four), and that no guide page ships a schema.org FAQPage node, so nothing structured describes any of it.

What I had not measured is the shape of the answers themselves, and the answers are the part a reader actually reads. So I fetched all 22 guide pages over plain HTTPS, took the first word of every question and the first word of every answer, and sorted both.

Three findings.

Two-thirds of the questions are yes/no questions

The 67 questions split three ways:

• 48 open with Does, Do, Is, Are, Can, Could, Will, Would, Has or Have.
• 18 open with What, How, Why, Which, Where or Who.
• 1 opens with Should - "Should I download PNG or SVG?", the Social Preview Designer's question.

Seven guides ask three yes/no questions and nothing else: Ambient Noise Mixer, CV Builder, NewsFlow, PDF Organiser, Subtitle Workbench, the Browser Teleprompter and the Word Counter. Eleven guides ask two yes/no questions and one other. Four ask one yes/no question and two others: the Email Signature Builder, the Label and Card Maker, the Password Generator and the Screen and Cinema Planner.

Our answers mostly open with the answer itself

Of the 48 yes/no answers, 38 open with a bare yes or a bare no:

• 23 open with "Yes" - "Yes, it is free. Ambient Noise Mixer runs in your browser with no account or payment details, and it never uses OnePassApps credits."
• 15 open with "No" - "No. Downloads are WAV only."

Ten do not. They lead with the fact instead of the verdict:

The reverse holds just as cleanly. None of the 19 questions that is not a yes/no question gets a yes or no opener. The grammar is consistent in that direction too: a yes/no question gets a yes/no answer or a fact-first answer, and an open question never gets a verdict it cannot carry. "Is my text sent anywhere?" is answered "No. Counting happens in this tab. The TXT report holds the numbers only, not your text." - the no first, then the reason.

The free-and-account question is a template with exactly one variable

On 20 of the 22 guides the last question is exactly "Is [Name] free, and do I need an account?", and its answer is exactly one sentence: "Yes, it is free. [Name] runs in your browser with no account or payment details, and it never uses OnePassApps credits."

All 20 match that shape character for character, with one variable - the name. On every one of those 20 it is also the third question, which is the last, because each of those guides asks exactly three questions.

I have published the credits sentence before, and I am not claiming that sentence again. What is new here is the question around it and the variable inside it. The variable is the display name, and on six of the 20 it is not the name in the URL:

• /guides/email-signature asks about Email Signature Builder
• /guides/label-maker asks about Label and Card Maker (written with an ampersand on the page)
• /guides/paperkit asks about PDF Organiser
• /guides/screen-planner asks about Screen and Cinema Planner (written with an ampersand on the page)
• /guides/social-card asks about Social Preview Designer
• /guides/teleprompter asks about Browser Teleprompter

The two guides that break the template are the two hosted apps, and they break it in the question, not the answer: NewsFlow asks "Does NewsFlow cost anything or use credits?" and MovieMatcherAI asks "Is MovieMatcherAI free, and does it use credits?" Those two carry the same credits sentence inside a longer answer that also explains accounts, permissions and usage limits.

What this adds up to

The documentation has a grammar, and it holds across the whole fleet. Forty-eight of our 67 questions are yes/no questions, and 38 of those 48 answers lead with the verdict. Ten answers choose the fact over the verdict, and each of those ten is a question where the bare yes or no needs a condition attached - whether the answer depends on the browser, on a file you already have, or on a plan.

And the one thing every reader will ask is a template with a single slot in it. Twenty guides answer it in the same single sentence with the name swapped in; two guides cannot, because their work runs on our servers rather than in the tab.

Read the fleet yourself: the guide index, or start with the Word Counter guide, whose answers are the sharpest example of the fact-first style.

Earlier in this series: 22 Guides, One Template counted the questions, The Guide Pages app.onepassapps.com Ships found the missing FAQPage, and What Each OnePassApps App Says It Will Not Do covered the limits section above the questions.

How this was measured

Each of the 22 guide pages was fetched over plain HTTPS with a browser User-Agent, so the bytes are the ones a visitor gets. The questions were the h3 headings inside each "Common questions" h2 section; the answers were the prose that follows each heading, with the tags stripped and whitespace collapsed. A question counts as yes/no if its first word is one of Does, Do, Is, Are, Can, Could, Will, Would, Has or Have, and as open if its first word is one of What, How, Why, Which, Where or Who. One question matched neither list - "Should I download PNG or SVG?" - so it is counted with the non-yes/no questions and called out separately above.

The template check was an exact string comparison: the answer text must equal "Yes, it is free. [Name] runs in your browser with no account or payment details, and it never uses OnePassApps credits." with the name from the question. All 20 matched. The first-word counts were taken from the same parse, not from a separate fetch.

What I did not measure

I did not measure whether any search engine reads this, and I am not claiming it does. There is no FAQPage node on any of the 22 guide pages, so nothing structured describes the answers - a crawler has to read the prose, and the effect of that is not something I can see from here.

I did not measure whether the fact-first answers read better than the verdict-first ones. That is a question about readers, and I have no reader data for our guides.

I did not check the app pages' own panels, which answer some of the same questions in a different place, and I did not look at the guides' "What you can do" or "How to use this app" sections, which sit above the questions and were covered separately.