Most content audits start in the wrong place. People open a page, read it, and ask "is this good?" — which is roughly as useful as asking whether a hammer is good without knowing if you're trying to drive a nail or hang a picture. Good for what?
Every page on your site has a job. A blog post exists to inform. A homepage exists to navigate people deeper. A product page exists to sell. A case study exists to persuade. A gated PDF landing page exists to capture an email. The single most common content problem I see isn't bad writing — it's a page trying to do a job it was never set up for, or doing none of them clearly.
So before you judge a word of it, you name the job. Then you score how well the page does that job, weighted toward what actually matters for that purpose. That's the whole method. The rest of this guide is how to run it on any page in about fifteen minutes, including the credibility signals (Google calls them E-E-A-T) that increasingly decide whether content ranks at all.
By the end you'll know:
- How to identify the single primary purpose of any page
- The specific signals that prove a page delivers on that purpose
- How to check E-E-A-T markers — bylines, citations, freshness, credentials, original data
- The engagement and clarity tells that separate read content from skimmed-and-bounced content
- A simple per-page scoring method you can apply across your whole site
Part 1 — Name the Job Before You Judge the Page
You can't score a page until you've decided what it's for. Pick exactly one primary purpose. Pages can have a secondary goal, but if you can't name one job above the others, that's your first finding — the page is unfocused.
1. Sort the page into one of five purposes
The five jobs, in plain terms:
- Inform — answer a question or teach something (most blog posts, guides, help docs).
- Persuade — change a belief or move someone toward a decision (case studies, comparison pages, "why us" pages).
- Sell — get a purchase or a high-intent action (product pages, pricing pages, sales pages).
- Capture — trade something for contact info (lead-magnet landing pages, newsletter signups, demo requests).
- Navigate — route people to the right next place (homepages, category and hub pages).
What "bad" looks like: you read the page and genuinely can't tell which one it's going for, or it's straining to do three at once — a blog post that suddenly hard-sells in paragraph two, a homepage stuffed with a 600-word brand essay.
What to do: write the job down in one sentence: "This page exists to persuade a mid-funnel buyer that we're more credible than [competitor]." If you can't finish that sentence, the page needs a purpose before it needs an edit.
2. Check that the page's structure matches its job
Each purpose has a shape it should take. An inform page wants headings, scannable structure, and a clear answer near the top. A sell page wants the offer, the price, and the proof above the fold. A capture page wants one form and almost nothing competing with it.
What "bad" looks like: a lead-capture page with the navigation bar, five outbound links, and three CTAs pointing different directions. A product page that buries the price three screens down.
What to do: ask "does the layout push the visitor toward the one action this page wants?" If the dominant visual element isn't serving the job, that's a structural miss — note it.
Part 2 — Score How Well It Delivers
Now you grade the page against its job. Use a simple 1–5 per check (1 = absent, 5 = excellent). I'll tell you which checks matter most for which purpose.
3. Clarity of the core message
Within five seconds, can a stranger say what this page is about and what they're supposed to do next? This matters for every purpose, but it's make-or-break for sell and capture.
What "bad" looks like: a headline that's a clever pun with no nouns in it. A value proposition you have to scroll to find. Jargon a first-time visitor wouldn't know.
What to do: show the page to someone outside your company for five seconds, then hide it and ask what it was for. If they can't answer, the message isn't landing — fix the headline and subhead first, because nothing below them gets read if those fail.
4. Does it actually answer the intent? (weight this heavily for inform)
For informational content, the test is whether the page answers the question someone typed to get there — completely, and ideally better than the result above you. Thin content that restates the question and links out fails this.
What "bad" looks like: 1,500 words that never directly answer the title question. A "what is X" article that's really a sales pitch for X in disguise.
What to do: read the page asking "if this question were mine, would I be done searching after reading it?" If you'd still open another tab, the content is incomplete. For a structured way to find what you're missing versus the pages that outrank you, a content gap analysis is the next step.
5. The call to action (weight heavily for sell, capture, navigate)
Every non-purely-informational page needs to make the next step obvious. One primary action, stated as a verb, visible without hunting.
What "bad" looks like: "Submit." "Learn more" floating with no context. Five equally-weighted buttons. No CTA at all on a page whose whole job is conversion.
What to do: count the distinct actions you're asking for. More than one primary ask per page splits intent. If the words on the button or in the surrounding copy are vague, that's a separate problem worth a full conversion copy audit — but at minimum, make the primary CTA specific and singular.
Part 3 — Check the Credibility Signals (E-E-A-T)
Google's quality guidelines lean on E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness. You don't need to obsess over the acronym — but the underlying signals are exactly what a skeptical human reader looks for too, which is why they matter for ranking and converting. These checks apply hardest to inform and persuade pages, especially anything touching health, finance, or safety.
6. Author byline and credentials
Is there a real, named human attached to this content, with a reason to trust them on this topic?
What "bad" looks like: "Posted by admin." No author at all. A byline with no bio, no link, no indication of why this person knows anything.
What to do: add a real byline with a one-line credential ("15 years running paid media for SaaS") and link it to an author page. For persuade and high-stakes inform content, this is one of the cheapest credibility wins available.
7. Citations and sourcing
When the page makes a factual claim or cites a number, does it show where that came from?
What "bad" looks like: confident statistics with no source. "Studies show…" with no link to any study. Numbers that look made up because they might be.
What to do: link claims to primary sources. Even a couple of credible outbound citations signals to both readers and search engines that you're not inventing things. If you can't source a stat, cut it.
8. Freshness
When was this last updated, and does the content reflect it? Stale content quietly bleeds trust — a 2019 "complete guide" to anything moving fast reads as abandoned.
What "bad" looks like: no date anywhere. A visible "Published 2018" with no update, on a topic that's changed since. References to tools or features that no longer exist.
What to do: show a "last updated" date and actually earn it — refresh the substance, not just the timestamp. For evergreen pages, a yearly review is plenty; for fast-moving topics, more often.
9. Original value — experience and first-hand data
The strongest E-E-A-T signal is something only you could have written: original data, a real screenshot, a first-hand result, a genuine opinion. This is the "Experience" that AI-generated sameness can't fake.
What "bad" looks like: content that could have been assembled from the top five results for the keyword without the author ever having done the thing. No examples, no specifics, no point of view.
What to do: add one thing only you have — a number from your own data, a photo of the actual work, a "here's what surprised us" note. One genuine first-hand detail outperforms three paragraphs of generic best-practice.
Part 4 — Read the Engagement and Clarity Tells
Finally, the mechanical signals that decide whether content gets read or skimmed-and-bounced. These apply to every purpose.
10. Scannability and formatting
Busy readers scan before they commit. Walls of text get bounced regardless of quality.
What "bad" looks like: 400-word paragraphs. No subheadings across a 2,000-word page. No bolding, no lists, no visual rest stops.
What to do: break content into short paragraphs, use descriptive subheadings every few hundred words, and surface key takeaways in bold or bullets. The page should be usable at a skim and rewarding at a full read.
11. Reading level and plain language
Match the language to the audience. For most business content, that means writing closer to plain speech than to a white paper.
What "bad" looks like: dense, clause-heavy sentences; corporate abstractions ("leverage synergies to drive outcomes"); a reading level well above your actual audience.
What to do: read a paragraph aloud. If you run out of breath or stumble, the sentence is too long. Tools like the Hemingway Editor flag this fast, but your own ear is usually enough.
12. Engagement-supporting data (if you have it)
If you have analytics, let real behavior confirm or contradict your scores. In GA4, look at average engagement time and scroll/bounce behavior per page.
What "bad" looks like: an "engaging" guide where average engagement time is eight seconds and almost nobody scrolls past the intro. High exits on a page whose job was to move people deeper.
What to do: treat low engagement time and high single-page exits as evidence the content isn't delivering — then cross-reference against the purpose-fit and clarity checks above to find why.
This is a structured manual read, and a manual read has limits. Going page by page across a real site, you'll start eyeballing the scoring and missing the cross-page patterns. A few things this checklist won't surface on its own:
- Comparative scoring across many pages — which of your forty blog posts are actually pulling weight versus quietly dragging down site quality.
- Purpose detection at scale — sensing each page's real intent from its content signals rather than your assumption about what it's supposed to do.
- E-E-A-T gaps you've gone blind to — missing bylines and uncited stats are easy to overlook on your own content because you already know it's true.
- An objective, weighted score you can hand to a stakeholder or compare before-and-after a rewrite.
If you'd rather not score every page by hand, that's exactly what the Content Effectiveness Evaluator does. Point it at a URL and it senses the page's primary purpose — inform, persuade, sell, capture, or navigate — then scores the content against a purpose-specific weighted rubric: clarity, intent-match, CTA strength, the full E-E-A-T set (bylines, citations, freshness, credentials, original data), and the engagement and readability tells above. You get an objective score and a prioritized list of what to fix first, on one page, for $39.99 per page — a fraction of an agency retainer, no subscription, delivered in under 24 hours. Comparison mode scores three or four pages side by side so you can see which content is carrying its weight. Run it on your most important page first at signalaudits.com.
Rather have it done for you? The Content Effectiveness Evaluator pulls the data, runs every check above, and hands you a prioritized, plain-English report in under 24 hours — a fraction of an agency retainer, no subscription, no upsell.