How to Use Google Search Console Data to Improve Content
Every other guide in this series ends by telling you the tool diagnoses rather than fixes. This is the page where the other half happens: which findings deserve work, in what order, plus how to tell afterwards whether any of it made a difference.
Reporting Is Not Improvement
A business can read this reporting every month for two years and change nothing. That happens constantly, it feels like diligence and it produces exactly the same result as not looking at all. The gap between reading and acting is where most of the value in this tool actually sits.
What the tool does. Describes.
It tells you what happened. Which searches, which pages, what was stored and what was passed over. All of it is past tense and none of it is an instruction.
What it never does. Decide.
Nothing in it tells you which page to rewrite, which subject to pursue or what to stop doing. Those are judgements and they need somebody to make them.
Why monthly reading becomes a ritual. It feels productive.
Numbers get looked at, a summary gets circulated and nobody has to make a decision. We have inherited accounts with two years of reports and no changes to the site whatsoever.
What acting actually requires. Choosing.
Picking a handful of pages, deciding what is wrong with each and accepting that some of them should not exist. That is uncomfortable in a way that reading a report is not.
What acting does not require. A large budget.
Most of what follows is writing decisions and removal decisions. The expensive part of search is earning visibility. This page is about what to do with visibility you already have.
How this page is ordered. By return.
Nearly working pages first because that is where the money is, then demand you have not served, then pages fighting each other, then the discipline that makes any of it measurable.
Start With What Is Nearly Working
The instinct is to fix what is broken. The return is in what is nearly working. A page already appearing for something valuable has standing that a new page has to spend months earning, so improving it is faster and cheaper than starting anything.
What nearly working looks like. Three shapes.
Appearing well and not being chosen. Appearing consistently just below where the visits begin. Appearing for the right searches with content that only half answers them.
Why these beat new pages. The hard part is done.
Google already associates the page with the subject. That association is the expensive thing. A new page begins without it however well written it is.
The first one to look at. Passed over.
Pages with plenty of appearances and few visits, which has its own guide because it is the single most actionable finding in the tool.
The second. Just below the line.
Pages appearing steadily without quite reaching the positions where visits happen. A genuine improvement to those pages has somewhere to go, unlike a page nobody sees at all.
The third. Half answering.
Pages appearing for the right searches whose content covers part of the question and stops. Those are the ones where adding two sections can change the outcome entirely. Block three is how you find what to add.
How many to take on. Five at most.
Enough to matter and few enough to keep track of. A list of forty gets abandoned in week two. Nobody ever learns which changes worked.
Queries You Did Not Write For
Filter the reporting to one page and read the searches it appears for. Some of them will be things nobody intended. Those are the most valuable lines in the whole tool, because they are demand you have already touched without trying.
Why they appear. Google understood something.
A sentence, a paragraph or an aside made your page relevant to a question you never set out to answer. Something on the page did that work by accident.
Why that is worth so much. Proven demand, proven relevance.
You know people search for it, because they did. You know you can compete, because you already partly do. Most research offers one of those and this offers both.
What to do with a good one. Serve it properly.
Either expand the existing page to answer that question fully. Or write a page dedicated to it if the subject is genuinely separate. The reporting tells you which by how much volume is behind it.
How to spot the good ones. Weak and repeated.
Appearing consistently in low positions with a handful of impressions each month. Consistency matters more than volume, since a single appearance is an accident and a monthly pattern is demand.
What this replaces. Guessing at demand.
Most content planning starts from what a business assumes people want. This starts from what people demonstrably searched for and reached you with, which is a considerably better foundation and costs nothing.
What to ignore. One offs and nonsense.
Single appearances, misspellings and searches from other markets. Our on-page material covers how to build a page around a subject once you have chosen one.
Pages Competing With Each Other
Where two of your own pages appear for the same searches, you have split one subject across both. Neither is as strong as a single page would have been. This is entirely self inflicted rather than anything Google did to you.
How to see it. Query first, then pages.
Take a search that matters and look at which of your pages appear for it. Two or three of your own addresses turning up is the finding.
How it happens. Nobody checked.
A blog post written in 2021 about the same subject as a service page. Two service pages that drifted towards each other. A location page and a service page covering identical ground.
Why it costs you. Divided attention.
The links, references and internal signals that would have made one page strong are spread across two adequate ones. Google also has to choose, possibly not the one you sell from.
The two legitimate answers. Merge or differentiate.
Combine them into one better page and redirect the loser. Or give each a genuinely distinct purpose. Doing neither leaves the situation as it is.
The version worth checking first. Your commercial pages.
Where a blog article and a service page compete, the article frequently wins and the enquiry does not happen. That is the expensive version of this problem and it is extremely common.
Which to pick. Ask what each is for.
If you cannot describe the difference in a sentence, merge them. If you can, make the pages reflect that difference properly. Our advanced material covers the consolidation mechanics.
What To Actually Change
Four kinds of change, cheapest first. Most businesses jump to the fourth because it feels like real work, when the first two resolve a surprising proportion of findings.
How the page presents itself. Cheapest.
The title and description somebody reads before deciding. An hour of writing against a ranking you already have, which is why it comes first.
The opening of the page. Next.
Whether the first paragraph answers the question that brought somebody there, rather than spending four sentences introducing your company. That single edit changes behaviour more than people expect.
What the page covers. More substantial.
Adding the sections that answer questions it appears for and does not address. This is where the unintended queries from block three get served.
What the page is for. The largest.
Where the page was written for the wrong purpose entirely, aimed at a search that does not exist or built to hold keywords rather than to help anybody. That is a rewrite rather than an edit.
What not to change. The address.
Changing addresses to improve them, without a genuine structural reason, is a common way to lose standing that took years to build. Leave working addresses alone.
Where the writing standards live. The on-page material.
How to write titles, structure a page and handle headings sits there rather than here. This page owns which pages to work on and why.
Change One Thing At A Time
If you rewrite the title, restructure the page, add two sections and change the internal links in one afternoon, something may improve and you will never know what. That is activity rather than improvement. It is also how agencies produce work nobody can evaluate.
The rule. One variable per page.
Change the presentation or the content, not both. If both genuinely need doing, do one, wait, measure, then do the other.
Why it feels wrong. It is slower.
Doing everything at once feels efficient and produces an unreadable result. Doing one thing feels timid and produces knowledge you keep.
What to write down. Three things.
Which page, what changed and the date. That is the entire record and it is worth more in six months than any report you will be sent.
Why the date matters most. Nobody remembers.
In three months nobody can say whether the change came before or after the dip. Without the date the comparison is a story rather than a measurement.
The exception. Something plainly broken.
If a page is missing a title entirely or contains an obvious error, fix it without ceremony. The discipline is for judgement calls rather than for faults.
What this buys you over a year. A method.
A dozen dated changes with known outcomes teaches you what works on your site specifically, which no general advice can. That is the compounding asset here.
How Long To Wait
Weeks rather than days. We will not give you a number. Anybody who does is guessing, since it depends on how often Google visits your site and that varies enormously between a busy shop and a five page brochure.
What has to happen first. A visit.
Google has to come back and see the change. Until then the stored version is the old one and nothing can possibly have moved.
Then a reassessment. Not automatic.
Seeing the change is not the same as re-ranking the page. That happens in its own time and may take several visits.
Then the reporting lag. On top.
The figures themselves trail reality, occasionally by weeks. So the earliest you could see an effect is later than the earliest the effect could have happened.
What to do while waiting. Leave it alone.
Not another edit, not a resubmission, not a third change. The waiting is part of the method rather than dead time.
What movement looks like early. Impressions before clicks.
A page being shown more often tends to appear before the visits follow, since being reassessed and being chosen are separate events. So a rise in appearances with clicks flat is progress rather than a disappointment.
When to conclude nothing happened. After a fair period.
Compare equal periods either side of the date, once enough time has passed that a visit is likely. If nothing moved, that is a result too and it narrows what to try next.
When The Answer Is To Delete
Sometimes the correct response to a page is to remove it. Not improve it, not optimise it, not rewrite it. That is a legitimate decision rather than an admission that money was wasted. It is frequently the finding on sites that have published for years without pruning.
The candidates. Three conditions together.
Almost no impressions over a long period, nothing linking to it internally and no answer to the question of what it offers that another page does not.
Why removing helps. Concentration.
Fewer pages means the attention Google gives your site is spread across pages that can win. It also means the internal links and references stop being divided.
Merge before you delete. Usually better.
Fold anything worth keeping into a stronger page and redirect the old address. That preserves whatever standing it had rather than discarding it.
What not to delete. Anything referenced.
Check the links reporting first. A page other sites point at is carrying weight regardless of its own traffic. Removing it without a redirect throws that away.
Why this is hard. Somebody paid for it.
The page cost money and deleting it feels like conceding that. The sunk cost is already sunk and keeping a page that does nothing does not recover it.
Where the mechanics live. The advanced material.
How to prune properly, what to redirect and in what order sits there. Everything on this tool sits on the Google Search Console guide.
Reading the
data is not
the same as
acting on it.
We have inherited accounts with two years of monthly reporting and nothing altered on the website. Reading feels like diligence and produces the same outcome as not looking. What produces something is picking five pages, changing one thing on each and writing down the date.
How we turn the data into work:
If the finding is that forty of your pages should be four, we will say so rather than quote you for improving forty.
Every guide.
One tool.
Getting set up and verified, using the reports properly, the performance and indexing data, finding what is broken, finding what is being passed over and turning any of it into work worth doing.
