Why this signal matters for Google
An expert opinion is a position you take and can defend. Not a summary of both sides. "A fixed rate gives security, a variable rate gives flexibility: it all depends on your situation" takes no skill to write. A mortgage broker who does the work every day writes something else: "I recommend the variable rate to most of my clients this year, except those planning to sell within two years, and here is why." That is what this signal measures: a reasoned opinion, meaning a clear personal position, backed by arguments a reader can check, with its limits stated out loud. In newspapers the format is called an op-ed, a guest column printed opposite the editorial page; on your own site it is simply the article where you say what you recommend. And it only counts if it carries a name: an author signature, tied to a bio that explains why this person gets to decide. When I go through a pest control company's blog, the first thing I look for is one sentence its competitor would not sign. Most of the time there is none: everyone recommends "a professional inspection" and nobody says when it is a waste of money.
Google tells its raters to look for expertise in the text itself. Human raters working for Google score page quality by following a manual anyone can read: the Quality Rater Guidelines, or QRG. That manual tells them that, for some types of pages, the level of expertise is clear from the main content itself, and it puts trust at the centre: a page has to be accurate, honest, safe and reliable, and a product review has to be written to help people decide, not only to sell. It also carries a warning aimed squarely at this signal: "reviews" from a manufacturer about its own product ("Our product is great!") are less trustworthy because of the conflict of interest. In other words, an opinion counts when it can cost you something: recommending a competitor, advising against a service you sell, saying a popular method does not work. The same manual adds that the rater must be able to identify who created the content, and notes that people often publish to show their own talent, experience and perspective. An opinion signed "The team" fails both tests at once.
Taking a side does not mean overstating it, least of all on a sensitive topic. On health, money or legal questions, what Google calls YMYL (Your Money or Your Life), the manual expects advice to stay consistent with the established expert consensus. Your opinion can break with habit, not with established facts. A physiotherapist is entitled to write "here is the protocol I favour for simple low back pain, and the three cases where I depart from it": that is a position, it is argued, and it stays inside the recommendations of their profession. They are not entitled to write "anti-inflammatories are useless" with no source. Three ingredients make an opinion credible without making it dangerous: arguments a reader can check (a study, a public figure, a case you handled), a concession (who your advice does not apply to), and a plain sentence on what you stand to gain, the conflict of interest signal. An opinion that always recommends what you sell is not wrong: it is simply worthless as evidence.
Readers and AI assistants look for someone who decides, not a site that summarises. People put their trust in those whose job is to show evidence: according to the Pew Research Center (2026), 77% of U.S. adults have a great deal or a fair amount of confidence in scientists to act in the public's best interests, more than in journalists, business leaders or elected officials. What makes the difference is not the title: a scientist is expected to show their evidence and to say what they do not know. Your opinion piece plays the same role at the scale of your site. Assistants such as ChatGPT, Claude or Perplexity prefer passages that answer a question cleanly: "Should you treat carpenter ants yourself? No, for three reasons" can be quoted; "it depends on several factors" cannot. A clear position, signed and dated, becomes a passage that can be cited with your name next to it. And because opinions age, date yours and say what would change your mind: that is what separates an expert from a salesperson.
The NeuroEEAT scoring grid (0-10)
Score one of your pages as a reader looking for someone able to decide, not as its author.
| Score | Description | Concrete indicators |
|---|---|---|
| 0-2 | No position | The text lays out the options and picks none. No "I", no author name. A competitor could sign every sentence. |
| 3-4 | Vague position | An opinion shows up ("we believe that"), but with no checkable argument, no named author and no limit. The reader cannot tell why or for whom. |
| 5-6 | Argued position | A clear stance, two or three arguments, a named author. Missing the concession, the date or the disclosure of what you stand to gain. |
| 7-8 | Complete position | Position, arguments with sources, explicit limits (who it does not apply to), signature linked to a bio, writing date. The opinion may put off some clients, and that is accepted. |
| 9-10 | Reference position | Several positions consistent over time, built on your own data or case files, picked up or debated by other experts, updated when the facts change, with a disclosure of what you stand to gain. |
How to audit it on your site
Take your five most-read articles and run them through these five questions.
- 01
Find the sentence a competitor would not sign
Read the whole article and highlight the first sentence that takes a side. If you reach the end without one, the page is an information sheet, not an expert opinion. It can still be useful, but it proves nothing about you.
- 02
Count the arguments, then their evidence
For each position, list the reasons given. Aim for three. For each one, look for something that lets a reader check it: a public figure, a study, a case you handled, a rule of your trade. An argument with no evidence is a preference.
- 03
Check who is speaking
Is the article signed by a person, linked to a bio that explains why they get to decide on this topic? "The team" or "Editorial staff" pulls the score toward zero: nobody is on the hook.
- 04
Look for the concession
An expert knows where their opinion stops. Find the sentence that says "unless", "not when", "I would change my mind if". It is missing from most articles, and it is the one that makes the rest believable.
- 05
Check the interest and the date
Does your opinion always recommend what you sell? If so, say it, or find a case where you advise something else. Then check the date: a 2022 position on a moving topic gets reread, confirmed or corrected.
How to activate or improve it
Three levels, depending on the time you have, from a neutral summary to a position people refer to.
Take a first position
- Pick a question your clients ask often and on which you hold a firm view.
- Write your answer in the first sentence, then three reasons, one per paragraph.
- Add a sentence that starts with "unless" to say who your advice does not apply to.
- Sign with your name, link to your bio, and show the date.
- Reread: if a competitor could sign the text without blushing, rework the position.
Make your positions checkable
- Back every argument with something a reader can consult: a study, a public figure, a trade rule, an anonymised case.
- State in one sentence what you gain if the reader follows your advice.
- Find at least one position where you advise against your own service: that is the one people will believe.
- Group your opinions under a recognisable heading ("My take", "What I recommend") so the line of thinking shows.
- Reread your positions every six months and note at the top of the page what has changed.
Become the reference that decides
- Build a position on your own data: your case files, your measurements, your results, with the method visible.
- Publish the same position off your site (op-ed, interview, podcast) and link to it from the article.
- Answer in public the experts who disagree, quoting their arguments fairly.
- Write an update policy: which signal would change your mind, and how you will announce it.
- On YMYL topics, have every position reviewed by a qualified peer and name them.
5 common mistakes to avoid
Sector specifics
Therapist or clinic
Google's manual expects advice here to stay consistent with medical consensus. So your opinion is about practice, not science: which protocol you prefer, in what order, and why you turn down a common request. Cite the guideline you follow and name the case where you refer to a physician.
Broker or adviser
An opinion on a financial product engages your liability and your regulator. Take a position on a method (how to compare, when to renegotiate, what to ignore) rather than on a promised return. Disclose how you are paid inside the article: it is the first thing a wary reader looks for.
Contractor or pest control
This is the easiest ground to take a side on, because your clients hesitate between two concrete options: repair or replace, do it yourself or call. Say what you would do at your own house, with a rough price and the case where you advise doing nothing. A contractor who writes "don't buy yet" wins the calls of the next six months.
Consultant or software vendor
Your buyers read ten neutral comparisons before they call you. What they want is someone who says "for a team under ten people, this tool is a mistake". Take positions on use cases, cite the projects that changed your mind, and accept that half your readers are not your ideal client.
Signals to activate in parallel
An opinion only holds if people know who signs it and what it has to lose. These signals make it credible:
FAQ
What is an opinion piece?
It is a text where the author defends a personal position on a question, with arguments, instead of reporting facts neutrally. In newspapers it is called an op-ed, a column or an editorial. On your own site, it is the article where you say what you recommend, why, and who it does not apply to. What separates it from a mere opinion is the justification: without checkable arguments, it is a preference.
How to structure an opinion piece?
Position first, then three reasons, then the limit, then the signature. State the position in the title or the first sentence, so the reader knows where you stand before scrolling. Give each argument its own paragraph with something checkable: a figure, a study, a case you handled. Add the concession ("unless", "not when"), say what you stand to gain, and close with your name, your bio link and the date. Newspaper op-eds add a news hook at the top; on your site, the hook is the question your clients keep asking.
How long should an opinion piece be?
Long enough to state a position and support it three times, which usually means 600 to 900 words; a newspaper sets its own limit and publishes it on its submissions page. Length is not the signal, though. A 400 word article that takes a clear side, backs it with evidence and names its author scores higher than a 2,000 word overview that never decides. Cut anything that does not serve the position, the proof or the limits.
How to write an opinion piece for a newspaper?
The same way, with three extra constraints: a hook tied to something in the news, the outlet's word limit, and a pitch sent to one publication at a time. Check the submissions page first, because each outlet publishes its own rules, including whether it wants the piece exclusively. Lead with your position, say in one line why you are qualified to hold it, and keep the arguments checkable. Then put the published piece on your own site too, linked from your bio: an opinion printed elsewhere is also an authority signal.
Will a firm opinion scare away clients or hurt a YMYL page?
Some clients, yes, and that is the point: a clear position turns away people who were never going to hire you and reassures the ones who will. A text that pleases everyone proves nothing about your expertise. On YMYL topics the rule is stricter, not different: stay inside the established consensus, take your position on method and use cases rather than on facts, show your evidence, and say who your advice does not apply to. The calls that follow a page that decides are better qualified, because the person already knows what you think.
Case study
Before · a broker's blog that summarises without deciding
After · the same page, rewritten with a signed and bounded position
Scenario built to show the score gap. No traffic or ranking figure is claimed: measure your own before and after.