How to measure SEO results and decide what to improve next

Start with a specific job for each page

Give each important page a job: a service page prepares suitable enquiries, a category supports product selection and a guide helps a reader make a decision. Do not automatically judge all of them by the same number. A useful guide does not need every reader to contact you immediately.

Before making a change, record the date, URL, intended audience, expected effect and starting values. Our minimum change log also names an owner and the next review date. Without a baseline, it becomes difficult to separate the effect of a new page from a seasonal increase in interest.

• ›Which page and market are changing?
• ›Which question should the page answer better?
• ›What is a sensible next step for the visitor?
• ›Which measurement shows that step happened?

Read search data in context

Search Console reports clicks, impressions, CTR and average position. CTR is clicks divided by impressions. Use page, country and device filters for relevant comparisons. Average position is an aggregate measurement, not a fixed ranking shown to every person.

Our suggested analysis separates searches for your brand from searches without the business name and starts with your main services. Additional guides can earn broad impressions and change the site average even while a commercial page improves. Review the affected group rather than depending on the overall total.

• ›Same search type and comparable periods.
• ›Same page or a clearly defined page group.
• ›Separate review of important countries and devices.
• ›Recorded changes to offers, advertising or inventory.

Separate clicks, form starts and accepted enquiries

Our measurement plan treats these as different steps: search click, website visit, contact click, form start and an enquiry accepted by the server. Clicking a contact button is not a submitted enquiry. Agree with the implementation team which event fires only after successful storage.

Exclude test submissions, spam and duplicates from qualified enquiries. Write down a simple qualification rule, such as a relevant service, a supported market and a reachable contact. If an enquiry cannot reliably be attributed to a source, report the source as unknown rather than assigning it to organic traffic.

Use a worked example, not a success promise

Suppose a defined group records 10,000 impressions and 200 clicks in an illustrative example. Its CTR is 2 percent. If the same bounded visitor group has 8 confirmed enquiries, those equal 4 percent of the 200 clicks. This is an example calculation, not a benchmark or a result from a real client.

Search clicks and measured visits can differ in practice. Prefer attributed website visits and enquiries under the same measurement definition when calculating a website conversion rate. With small samples, report absolute counts too: one additional enquiry can create a large percentage change.

Turn the report into a next action

Our decision framework is: no relevant impressions — inspect intent and indexation; impressions without suitable clicks — compare queries with the page proposition; visits without a next step — review the offer and navigation; form starts without submissions — test the form journey. Each finding is a hypothesis to investigate, not automatic proof.

Revise titles when they poorly describe the page or its intended search need. Google can build title links from several sources, so editing one title does not guarantee a particular display. A useful monthly report needs the baseline, change, observations, uncertainties and one prioritised next action. A few traceable decisions beat a dashboard that leads nowhere.