A practical SEO decision

SEO success starts with the next useful fix

Before you commission more pages, find what stops an important page from doing its job. Can the right people find it, get the answer they need and complete the next step? This guide helps you choose one useful fix and check what it achieved.
Illustration of content, technical health, authority, time, measurement, and adaptation feeding a compounding SEO flywheel

One page. One reader. One useful action.

Choose a page for a service you can deliver or a question your customers actually ask. Write down who it serves and what a useful outcome looks like: a received enquiry, a confirmed booking or a completed purchase.

For example, a training company might choose its spreadsheet course page and aim for enquiries from office teams that the course can help. “More traffic” does not tell the team which problem to fix. A clear page and outcome do.

Find the failure before choosing the work

Three problems that need different fixes

Check the page and its evidence together. If an essential step is broken, repair it before trying to attract more visitors. The checks below guide an investigation; none is a ranking formula.

Google cannot access or index the page

Inspect the exact URL in Search Console. Read the reported reason, crawl permission, fetch result and indexing permission. In the indexed report, check whether Google chose another page as its main version (the canonical URL).

Correct an unintended block or fetch failure at its cause, then test again. A successful live test does not guarantee indexing or visibility. A page with no impressions is not, by itself, proof of a technical fault. Google’s URL Inspection guidance.

The page appears, but relevant searches bring few clicks

Filter Search Console to the page. Look at the relevant queries, country, device and position context, then inspect the actual search result and competing answers.

If the page or its title promises the wrong thing, correct that mismatch. Do not change a title just because the sitewide click-through rate is low. Different searches and result layouts can produce different click patterns.

Visitors arrive, but the intended action is missing

Follow the journey on mobile and desktop. Can the reader understand the offer, find the next step and complete it? For an enquiry form, test actual receipt as well as the on-screen response.

If delivery fails, fix delivery. If the form works but the page omits information needed to decide, improve that information. Do not assume that buying more traffic or changing rankings will resolve either problem.

An illustrated failure

A green check is not a received enquiry

A form shows a green check, but an envelope stops at a broken delivery connection before the empty inbox.
Illustration: the page can confirm submission while the delivery route fails. Check the message at its destination.

Imagine that the training company gets relevant visits but few enquiries. Its form shows a confirmation after submission. A test message never reaches the booking team. That is a delivery problem to investigate, not evidence that the page needs more keywords.

On a form you own, submit clearly marked test details. Check the visitor’s confirmation, any saved submission and the intended recipient’s inbox. If the message is missing, follow the delivery route and inspect the failure. Retest after the repair.

If you use Google Analytics 4 (GA4), also confirm that the intended event appears in Realtime or DebugView. A recorded event verifies the configured measurement, not email delivery or lead quality. Google’s key-event guidance.

When the journey works

Replace a missing answer with something the reader can use

Now suppose the form works, but the course page does not explain who it is for or what happens next. The following invented example shows a content decision, not a customer result.

Too little to make a decision

“Practical training for ambitious teams. Contact us to unlock your potential.”

The visitor cannot tell what the team will learn, what to bring or what information the provider needs.

Enough detail to start a useful enquiry

“A half-day spreadsheet workshop for office teams. Bring a laptop. We cover sorting, filters and formulas using sample business data. Tell us your team size and preferred date. We will confirm availability and send a quote.”

The reader can assess the fit and send a more useful enquiry. On the real page, state the actual format, requirements, price or quoting process, and relevant limits.

Use customer questions and your own experience to identify the missing answer. Add evidence where a claim needs support. Google’s people-first content guidance focuses on helping the intended reader achieve a goal; it does not prescribe a word count or guarantee a ranking.

Make the result reviewable

Record the change. Keep the measurements separate.

Before editing, keep a short note of the page, observed problem, chosen change, publication date, comparison period and report filters. Give someone responsibility for the review. Avoid several unrelated changes at once if you want to learn which explanation still fits.

Search exposure and response

In Search Console, compare impressions (appearances in results), clicks and click-through rate (clicks divided by impressions) for the page and relevant queries. Use the same search type, country and device filters in both periods. Consider position and changes in demand.

Visible query rows are incomplete: some queries are omitted, and filters can change totals. They are evidence to investigate, not a full list of everyone who searched. Performance report · Filtering and comparison limits.

The business action

Track the outcome you defined: received enquiries, suitable bookings or completed purchases. Check that your measurement still records the intended action. Review enquiry quality with the people who handle it.

Search Console clicks, Analytics sessions and enquiries are different measures. Put them beside one another; do not divide GA4 key events by Search Console clicks and call that a conversion rate. Consent, tags, time zones and reporting rules can affect the counts. Google’s explanation of the differences.

What the evidence allows you to conclude

First confirm that the edit is live and the repaired step works. Where relevant, check whether Google has processed the changed page. Then review comparable periods, including seasonal context.

A review date is a checkpoint, not a promise of results. Google says effects can take days to several months. Demand, competitors and other changes can also move the numbers. Keep, revise or continue observing; “not enough evidence yet” is a valid conclusion. Google’s guidance on search traffic changes.

Build on work that remains useful

The next improvement should start from what you learned

A working form, a clearer answer and a reliable record of changes are things the business can use again. Keep those improvements maintained, then choose the next important page or unresolved problem. Search growth is an outcome to observe, not an automatic reward for publishing more.

Promote useful work to relevant readers, but do not buy or manufacture links to manipulate rankings. Legitimate advertising and sponsorship are different: Google allows paid links when they are appropriately qualified with rel="sponsored" or rel="nofollow". Google’s link-spam policy.

Start with this sentence: “This page serves ___; the action we need is ___; the evidence points to ___; we will change ___ and check ___.” If you cannot fill in the evidence, investigate before you commission the work.

Stop guessing what to fix next.

Send us the page address, the result you need and what is getting in the way. We’ll reply in writing with a proposed first step, so you can decide what work to commission.

Illustration of content, technical health, authority, time, measurement, and adaptation feeding a compounding SEO flywheel