Helpful Content Guidelines: Article Scoring Table

Most content teams have read Google's guidance on helpful content at least once, nodded along, and then gone back to publishing exactly as before. The reason is practical rather than moral. The guidance is written as a long list of questions, and a list of questions is hard to apply on a Thursday afternoon when an editor has to decide whether a draft is ready.

This guide turns the questions into something an editor can actually use: a scoring table for a single article. It walks through what Google says, converts each idea into a row with a clear test, explains how to read the result, and then covers the order in which weak articles should be repaired. It is aimed at content editors, SEO writers and agency teams who review many pieces and need the same standard applied to each of them.

One boundary comes first. The table below is a working template proposed in this article. It is not a Google product, it does not measure any internal Google signal, and a high total does not mean a page will rank. Its only job is to make the same questions get asked every time.

Google's People-First Content Questions

Google's page on creating helpful, reliable, people-first content (Opens in a new tab) says its ranking systems aim to prioritise helpful, reliable information created to benefit people, rather than content made to manipulate search rankings. Everything else on the page hangs off that sentence. The page then organises self-assessment around three questions: who, how and why.

Who: can a reader tell who made this?

Visitors should be able to see who created the content. Where readers would expect a byline, the page suggests including one, and that byline can lead to more background on the author. The page is blunt about deception here: fabricated author profiles, such as made-up names, false credentials or AI-generated headshots, are treated as a low-quality signal. For a scoring sheet this means the author row is not about whether a name appears, but whether the person is real and whether a reader can verify what they know. If your site has no author pages yet, the guide on building an author page covers the practical steps.

How: would a reader want to know how it was produced?

The guidance says readers can benefit from knowing how content was created, including whether automation or AI played a role, and recommends disclosure when it would be reasonably expected. For product reviews it gives the example of explaining how many products were tested, the method and supporting evidence. Translate that to any article by asking what a sceptical reader would want to see: where the numbers came from, who ran the test, what the date of the information is, and whether any part of the draft was machine generated.

Why: who is this for?

Google describes the why question as perhaps the most important one. Content is people-first when it exists primarily to help the people who visit directly. When the main purpose is to attract search visits, it does not match what the systems are trying to reward, and using automation mainly to manipulate rankings is described as a violation of spam policies. None of this makes search optimisation illegitimate. The same page says optimisation is fine when applied to people-first content. What matters is the order: the reader's question comes first, and findability is built on top of an answer that deserves to be found.

Four groups of questions you can reuse

Beyond who, how and why, the page lists questions in four groups: content and quality, expertise, page experience, and a set of indicators of people-first thinking. For the content and quality group, a few examples are enough to show the flavour. Does the content provide original information, reporting, research or analysis? Does it give a substantial, complete or comprehensive treatment of the topic? Does it offer insight beyond the obvious? When it draws on other sources, does it add real value rather than copying or lightly rewriting them? Does the title describe the content and avoid exaggeration? Would someone want to bookmark it or recommend it to a friend?

These are not yes or no questions in practice. An article can be partly original, or comprehensive in one section and thin in another. That is why the scoring table later in this guide uses a three-step scale instead of a binary check.

Checks for Expertise, Experience and Originality

The expertise group on Google's page asks whether the content inspires trust through clear sourcing, evidence of expertise and background about the author or site, such as author pages or an about page. It asks whether the content is written or reviewed by an expert or enthusiast who knows the topic well, and whether it contains easily verified factual errors. Each of these can be checked in a few minutes with the article open in one window and the sources open in another.

What the documentation says about E-E-A-T

The same page says E-E-A-T is not a specific ranking factor, although it is useful to consider the factors that identify it. It states that trust is the most important of these aspects, that content does not need to demonstrate every aspect, and that experience or expertise alone can make content helpful. For topics that affect health, financial stability, safety or the welfare of society, it says strong E-E-A-T signals deserve more weight. Google's SEO starter guide goes a step further and lists thinking that E-E-A-T is a ranking factor among the things you should not focus on. The article on SEO myths and what Google documentation says places this misunderstanding in a wider list.

This has a direct consequence for the worksheet. A single row labelled E-E-A-T score would invent a metric that does not exist. The sheet therefore scores the concrete evidence behind each idea: the author's identity, the first-hand detail, the sourcing and the factual accuracy.

Experience shows up in details

One of the people-first indicators is whether the content demonstrates first-hand expertise and depth of knowledge, for example from using a product or visiting a place. In an article, that usually looks like a detail that only someone who did the work would include: the step that failed and why, the caveat that a manual leaves out, the choice between two options that was harder than it looked. If nobody on the team has done the thing, the honest move is to say that the piece is a researched overview. A transparent overview is more trustworthy than simulated experience.

Originality means adding something

Originality does not forbid using other sources. The question Google asks is whether the content adds substantial value instead of copying or rewriting what is already there. A quick test: if this page vanished tomorrow, could the reader find the same thing, in the same words and with the same structure, on another result? If so, the page has no case to make. What usually fixes this is not invented data. It is your own process, a decision table, a limitation specific to your field, a corrected misconception, or a worked comparison that only your team can write.

Accuracy needs a paper trail

Because Google asks directly about easily verified factual errors, the sourcing row of the table has two parts. Does each important claim point to a primary source, and does that source actually support the sentence? The simplest audit method is mechanical. List every number and every sentence that begins with a phrase like Google says, open the cited page, and compare. If the page does not support the claim, rewrite the sentence or remove it.

Google's SEO starter guide (Opens in a new tab) points the same way. It says that providing expert or experienced sources can help people understand the expertise behind your articles, and it advises against copying other people's content in part or in its entirety.

Signs You Are Writing for Search Engines First

Alongside the positive indicators, Google's page gives a reverse list. If you answer yes to some or all of these questions, it says, you should reevaluate how you create content. In the worksheet this list becomes an inverted section, because here a yes is the bad news.

• Is the content primarily made to attract visits from search engines?
• Are you producing lots of content on many topics in the hope that some of it performs?
• Are you using extensive automation to produce content on many topics?
• Is the content mostly a summary of what others say, with little added value?
• Are you writing about trending topics your audience would not otherwise care about?
• Will readers need to search again to find better information elsewhere?
• Are you targeting a particular word count? The page answers this one itself: no, Google does not have a preferred word count.
• Did you enter a niche without real expertise, mainly for search traffic?
• Does the content promise an answer that does not exist, such as a release date that has not been confirmed?
• Are page dates being changed to look fresh without substantial changes to the content?
• Is content being added or removed purely to make the site seem fresh?

Notice the last two. The page treats cosmetic freshness as a problem, not a tactic. An update belongs on a page when something actually changed, and a short note saying what changed is a useful trust signal for the reader.

A note on AI-assisted drafts

Google's page on generative AI content (Opens in a new tab) says these tools can be useful for researching a topic and adding structure to original content. It adds that using them to generate many pages without adding value for users may violate the spam policy on scaled content abuse, so the risk comes from missing value, not from the tool. The page also explains that generative models do not retrieve facts but predict a likely sequence of words, which is why hallucinations happen and why you should fact-check by hand before publishing. That review should include title elements, meta descriptions, structured data and image alt text, since these can appear in search results.

For the worksheet, an AI-assisted draft makes the how row stricter, not different. Who verified the claims, against which sources, and which parts rest on human experience? An article that cannot answer those questions deserves a low score whatever tool produced the first draft.

An Article Scoring Table (Template)

The table gathers the questions above into one worksheet. Each row is scored 0, 1 or 2. A 0 means the signal is absent or the answer is negative, a 1 means it is partly present, and a 2 means it is clearly present and you can point to evidence. The evidence note is not optional. A score without a note cannot be reproduced by the next reviewer, and an unreproducible score is just an opinion with a number attached.

There are fourteen rows, so the maximum is twenty-eight. Do not manage by the total alone, because an average can hide a fatal row. Three rules keep the sheet honest. First, any article that scores 0 on factual errors or source support goes to the front of the queue regardless of its total. Second, a 0 on the inverse row means the article needs a new plan, since the problem is why it exists rather than how it is written. Third, total thresholds for pass, revise and rework are set once by your team and written down. Those thresholds are your own convention, not a number from Google.

How to run a scoring session

1. Have someone other than the author score the piece. Google itself suggests seeking honest, unaffiliated outside reviews, and a second reader removes much of the author's blind spot.
2. Write the evidence note at the moment you assign each score: the paragraph, the source, the screenshot.
3. Keep old scores with their dates when you re-score. The history shows whether your fixes did anything.
4. Keep scores separate from traffic data. A high score is not a traffic forecast, and an article with a weak score can still receive visits. The worksheet measures editorial quality only.
5. Review the template itself whenever Google updates the underlying documentation.

How to Prioritise Fixes for Low-Scoring Articles

A backlog of weak articles is the normal situation, not an emergency. The scoring table earns its keep here, because it sorts the work. The order below is a recommendation of this guide, and its logic is simple: fix first what makes everything else pointless.

Step one: correctness

A wrong statement undermines every other improvement. Start with unsupported or incorrect claims. Google's page says trust is the most important element of E-E-A-T, so this step is not arbitrary. When a claim cannot be verified, removing it is better than leaving it half sourced.

Step two: the reason the article exists

Ask which reader question the piece answers. If the answer is vague, polishing the prose is wasted effort. Weak articles generally end up in one of three places: merged into a stronger article, rewritten around a specific reader need, or retired. For decisions across a whole site, the inventory approach in how to run a content audit is the right companion. That guide works at site level, while this worksheet looks inside one article.

Step three: originality and depth

Next, add what is missing, which is something the reader cannot find elsewhere. That might be a decision table, a short method note from your own practice, or a correction to a common mistake. Depth is not length. Because Google lists targeting a word count among the warning signs, padding is counterproductive, and every added paragraph should answer a question the article did not yet answer.

When the root problem is shallow pages that say little, the guide on thin content goes further into detection and repair. If both the original contribution and depth rows score 0 on the same article, move to those methods.

Step four: visible authorship, method and sources

Adding a byline, a method note and a source list does not require changing the argument, so it feels like a quick win. It is deliberately placed fourth. A tidy source list on top of weak content is decoration. Add the author details, but make sure the author really has the knowledge the profile suggests, because invented profiles are named on Google's page as a low-quality signal.

Step five: title, promise and presentation

Leave the title for last, since edits to the body can change what the article promises. Rewriting the headline first and then changing the content can open a fresh gap between the two. Then deal with spelling, layout, headings and mobile reading comfort.

Step six: honest dates

Only change a published or updated date when the page really changed. The reverse-list item about dates makes this explicit. A line such as this version rewrites the second section and replaces two outdated sources gives readers, and the next scorer, something real to rely on. Google's starter guide frames the overall principle simply: review old content, update it if needed, and remove it if it is no longer relevant.

A Fictional Example: One Help-Centre Article

This example is fictional. It does not describe a real company or outcome and contains no figures. It only shows how the worksheet turns a vague feeling into a ranked plan.

Imagine a software company whose help centre has an article titled with the phrase best practices for a particular feature. The editor scores it. The byline is a team alias with no profile, so the author row gets a 0. The instructions describe a menu layout that changed in the last release, so factual errors gets a 0 too. The writing is clear and the headings are fine, so presentation rows score well. The article also repeats, almost word for word, what the product's own documentation says, so original contribution scores a 1 at best.

Under the three rules, the outdated menu instructions go first, because factual errors is a critical row. The editor corrects the steps and checks them against the current interface. Next comes the reason the article exists: support tickets suggest readers get stuck at one particular step, so the editor rewrites the middle section around that step and adds a short troubleshooting table based on those tickets. Only then does the byline get attached to a real support engineer with a short profile, and the update note describes exactly what changed.

Notice what did not happen. The editor did not lengthen the article, did not change the date alone and did not chase a score for its own sake. The worksheet kept attention on the rows that mattered.

Building the Worksheet into Your Workflow

A template that sits in a folder does nothing. Three moments are worth wiring it into: the final read before publishing, the periodic refresh pass, and writer onboarding.

Before publishing

The editor completes all fourteen rows before a piece goes live. A zero on a critical row delays publication. This feels strict until you compare the cost of a ten-minute check against the cost of correcting a published error. A useful warm-up is to write the reader's question in one sentence. If the article cannot be summarised as the answer to one question, the reader focus row will be a 0 anyway.

Refresh passes

Old articles are re-scored on a schedule with the same sheet. Comparing the new score with the previous one shows whether the edit changed anything that matters. Some articles need no update at all. If the topic is stable and the facts still hold, rewriting them just to look current runs straight into the behaviour Google lists as a warning sign.

Writer onboarding

For new writers the worksheet is a concrete answer to the question of what good looks like. Instead of handing over a rulebook, score two published articles together and discuss why they differ. Writers quickly see which rows challenge them most, and for many it is original contribution.

Agency and client reporting

For agencies, the worksheet also gives a neutral reporting language. Saying an article is weak starts an argument. Showing which rows scored low, with evidence notes, turns the same conversation into a plan. It lets you propose improvements without promising a ranking or a traffic figure, because the plan depends on raising specific rows, not on a forecast.

Common Mistakes When Scoring

• Trusting the total: A strong total can hide a single 0 on a critical row. Read the critical rows separately.
• Inventing an E-E-A-T score: Google says E-E-A-T is not a specific ranking factor. Score the concrete signals behind it instead.
• Treating length as quality: Google lists targeting a word count among the signs of search-engine-first writing.
• Reading the score as a forecast: The worksheet is an editorial control. It does not predict rankings.
• Decorating with fake expertise: Made-up names, AI-generated headshots and unverifiable credentials are exactly what Google's page calls out.
• Publishing unchecked AI output: Models predict likely word sequences, so unreviewed output can contain errors.
• Refreshing dates without changing content: The reverse list names this as a warning sign.
• Self-scoring: The author is the worst-placed person to judge their own blind spots.

Frequently Asked Questions

What are Google's helpful content guidelines?

They are a set of self-assessment questions in Google's documentation on creating helpful, reliable, people-first content. The questions cover who created a page, how it was made and why it exists, plus checks on content quality, expertise and page experience. They are guidance for creators, not a checklist that guarantees rankings.

Is there an ideal word count for helpful content?

No. Google's people-first content page lists targeting a particular word count among the warning signs of search-engine-first writing and answers the question with a plain no. Length should follow from covering the topic completely.

Is E-E-A-T a ranking factor?

According to Google's documentation, E-E-A-T is not a specific ranking factor. The page still treats the underlying signals as useful and says trust is the most important of them. Content does not have to show every element, because experience or expertise alone can make a page helpful.

Can I use a generative AI tool to draft articles?

Google's guidance says generative AI can help with researching a topic and adding structure to original content. It also warns that producing many pages without adding value for users may violate the spam policy on scaled content abuse, and that output should be fact-checked by hand before publishing.

Is an article scoring table an official Google tool?

No. A scoring table is an editorial worksheet that turns Google's published questions into repeatable checks for your own team. The total does not predict rankings or reflect any internal Google score; it only helps you decide which article to fix first.

What should I do with an article that scores badly?

Start by deciding whether it still answers a real reader need. If it does, fix it in the order described in this guide. If it no longer does, Google's starter guide suggests reviewing older content, updating it when needed or removing it when it is no longer relevant.

Conclusion and Next Step

Helpful content is not a badge or a score. It is content where the author, the process and the purpose are visible to the reader. Turning Google's who, how and why questions, plus its content, expertise and page experience questions, into a single worksheet makes that idea repeatable across a team. The value lies in the evidence notes and the repair order: accuracy first, then the reason the article exists, then original contribution, and finally presentation and dates.

If you would rather hand the rewriting work to specialists once your articles are scored, you can look at the SEO content writing service at Akua Adforce; scope and terms are written on that page. The template in this guide is a general editorial framework and does not imply that any article will earn a particular ranking or traffic.