What to look for in Google Analytics, Search Console & the tools worth paying for.

Before the first chart

You should be able to explain why you opened the report.

You log into Google Analytics to find out whether the website is working. Twenty minutes later, you know how many people used a particular browser, which country produced a mysterious burst of visits, and that something called an engaged session has increased. You still do not know whether the business is getting better opportunities.

This happens to intelligent people running perfectly good companies. Analytics software gives you far more things to inspect than your business has reasons to inspect them.

I want an owner to be able to open the right report, recognize what deserves attention, and leave with a useful next step. That might mean fixing a broken inquiry path, improving one service page, investigating an unfamiliar referral source, or deciding that the evidence is too thin to justify changing anything yet.

It may also mean keeping your credit card in your pocket. Semrush, Ahrefs and specialist tools can earn their cost. They can also become expensive ways to avoid learning what the information you already have is trying to tell you.

A landscaping company needs to know whether its website brings commercial properties it can serve. A hotel needs to distinguish someone admiring the pool from someone progressing toward a profitable booking. A specialist manufacturer needs to know whether the site is attracting buyers with a real application, even if those people take months to request a quote.

Those businesses should not inherit the same definition of success simply because they use the same software.

Inside Google Analytics 4

Start with acquisition, entrances and meaningful actions.

You do not need to learn every GA4 report before doing useful work. Begin with a small set that answers a sequence of questions: where visits came from, where people entered, what they did, and whether those actions connected to the business outcome you wanted.

Report navigation varies by property configuration and access. Use the report names below; if one is missing, an editor may need to add it to your reporting collection. The absence of a menu item is not evidence that the data never existed.

Traffic acquisition: which sources brought the sessions?

Open Traffic acquisition and inspect Session source / medium. This separates sources such as Google organic search, an email campaign or a referral. A channel grouping is useful for an overview; source and medium provide a more specific explanation of where the sessions originated. Traffic acquisition documentation.

Read sessions alongside the specific outcome you care about. If the purpose is quote requests, choose that event rather than a total containing every event the account has designated important. A channel producing many visits and few suitable inquiries deserves a different investigation from a channel producing little traffic and several substantial opportunities.

Start by asking which sources changed enough to matter in absolute numbers. A rise from two sessions to six is 200 percent growth. It is also four additional sessions. Both statements are correct; only one is likely to keep the meeting in proportion.

User acquisition: how were people first acquired?

User acquisition uses first-user dimensions; Traffic acquisition uses session dimensions. A person first discovered through search can return through email. Those reports can assign different source labels because they describe different stages of the relationship. Google's explanation of traffic-source scopes.

Use the first view when evaluating initial discovery and the second when reviewing the sessions happening in your selected period. When someone asks whether email or search deserves credit, a useful response is often to examine the roles both played before treating them as rival explanations.

Keep the scope in your report title. “New users by first source” and “Sessions by session source” tell readers what is being compared. “Our best channels” invites a much broader interpretation than either report can support.

Landing page: where did the visit begin?

The Landing page report identifies the first page in a session. It is different from a list of all pages viewed. Use it to inspect which entrances attract visits and what happens during those sessions. Filter to a relevant acquisition source before comparing results. Landing page report documentation.

This is where the homepage loses its automatic status as the center of the universe. Many valuable prospects arrive on a service page, a technical explanation or a location-specific resource. Their experience begins there, without the context you carefully placed on the homepage.

Open the leading entrances yourself. Does a first-time visitor immediately understand who the page is for? Is the offer clear? Is the next step appropriate to the question that brought the person there? Does the page provide evidence of capability, or does it merely announce that the company is excellent?

A specification article may need a link to the relevant application or product. A wedding venue page may need capacity, availability expectations and a clear inquiry path. A commercial maintenance page should make the service area and property fit easy to establish. The report identifies where to look; reading the page identifies what a visitor actually encounters.

Pages and screens: what supports the visit after arrival?

Pages and screens includes pages viewed during visits, not only their entrances. It helps you investigate material that people use after arriving, such as case studies, pricing explanations or contact information. Pages and screens documentation.

Do not condemn a page because it attracts little direct search traffic. A credentials page may help a prospect evaluate you after a referral. A detailed process page may answer the question that makes an inquiry possible. These pages can do useful work without being high-volume entrances.

Look for plausible sequences, then verify the experience. If visitors frequently reach a case study but rarely continue toward a relevant service or contact route, inspect those connections. A next-step link should reflect what the reader now understands, not simply send everyone to a generic contact page.

Engagement: did the interaction deserve closer attention?

GA4's engaged-session definition combines time and activity criteria, including a key event or multiple page or screen views. Engagement rate is the proportion of sessions meeting that definition; it is not a satisfaction survey. Check your configuration before comparing thresholds across properties. Engagement metrics in the acquisition report.

A long visit might reflect serious consideration. It might also mean someone cannot find the room dimensions. A short visit might reflect rejection, or a returning customer quickly finding your phone number. Treat duration as an invitation to inspect the task, not a verdict about the quality of the visitor.

Compare similar pages serving similar purposes. A substantial guide and a contact page do not need the same reading time. Chasing a sitewide engagement benchmark can lead you to make useful pages slower to use.

Key events: what happened that matters?

For lead generation, distinguish a button click, an attempted submission, a successfully received inquiry and a qualified opportunity. They are different stages. Google's recommended events include generate_lead and later lead-stage events, but collection requires implementation appropriate to the event. Installing Analytics does not populate your sales process by itself. Recommended lead and commerce events.

I would rather see a modest number of defensible inquiry completions than a spectacular total made from scrolls, clicks and people opening the contact page. Keep useful intermediate signals, but label them as intermediate signals.

Select the relevant key event when reviewing a rate. A session key event rate describes the share of sessions with that event; it should not be casually relabeled as the percentage of people who became customers. Event totals can include repeat activity, and a single customer may return several times.

Revenue: does it mean the same thing as the business ledger?

Commerce reporting depends on correctly implemented purchase information. Review transaction identifiers, amounts, currencies and refunds, then reconcile a sample with the order or booking system. Google's recommended ecommerce events provide the collection vocabulary; your implementation determines whether the records represent real transactions correctly. Commerce event reference.

For a hotel, a booking value is not automatically the value of a completed stay. For an online retailer, sales revenue does not subtract every cost of serving the order. For a consulting company, assigning a hypothetical dollar value to a form event does not turn that form into earned revenue.

Use the vocabulary finance uses when you discuss commercial performance. Where an analytics value is an estimate, call it an estimate. The report gets more useful when it becomes more honest.

Inside Google Search Console

Read the search that happened, then inspect the page that answered it.

Search Console is particularly useful because it lets you examine the language associated with your website's Google visibility. It can reveal that the market understands a service differently from the way you named it, or that an article is attracting researchers when the page was intended to attract buyers.

Begin in the Performance report for search results. Choose a completed comparison period, check the search type, and examine queries and pages. Add country and device filters when they affect the business question. Keep the same filters when comparing periods. Performance report documentation.

Clicks, impressions, CTR and average position need different interpretations.

Clicks count reported click activity from Google. Impressions reflect appearances under Google's rules for the relevant result type. CTR is clicks divided by impressions. Average position summarizes the relevant position observations; it is not a promise that every searcher saw you in that exact place. Google's metric definitions.

Read them together. A query with substantial impressions and few clicks might deserve a closer look at the result and its relevance. A query with few impressions may be highly specific and commercially valuable. High CTR on your own company name and lower CTR on a competitive research phrase are not automatically evidence of a problem.

Compare a page with its own prior performance under similar conditions before comparing it with an unrelated page. The searcher's task, the result layout, the device and your brand familiarity all affect what a plausible result looks like.

Separate brand discovery from category discovery.

Create a practical brand group including your business name and common variations. For an owner-led business, decide whether the owner's name belongs in that group. Be careful with names that also have ordinary meanings; a crude text filter can classify unrelated searches as brand demand.

Brand searches help you understand people looking specifically for you. Nonbrand searches help you investigate discovery around a service, problem, place or product. Both matter, but they should not be mixed when evaluating whether your reach beyond existing awareness is improving.

Keep a note of your grouping rule so the comparison remains stable. If a report quietly changes its definition of brand halfway through the year, the trend becomes less useful.

Look at query families that reflect actual customer needs.

Organize a manageable set around the questions your business can profitably answer. A specialist might use application, service, buyer problem and geography. A destination business might use experience, travel planning, booking considerations and branded research.

Do not turn every variation into a new page. “Commercial landscape maintenance” and “commercial grounds maintenance” may describe overlapping needs; whether they deserve separate pages depends on the service distinction and existing content. A new phrase is evidence to investigate, not an automatic publishing assignment.

Ask whether the right existing page addresses the whole task convincingly. Sometimes the best work is a clearer explanation, an added qualification detail or a better internal link on a page that already earns relevant visibility.

Use query-to-page checks before proposing new content.

Select a relevant query or tight group and inspect the pages appearing for it. Then inspect the queries associated with the intended page. This two-way check helps you understand whether Google and the reader are finding the page you expected.

More than one URL appearing for related searches does not, by itself, prove cannibalization. A service page and an educational explanation can satisfy different tasks. The concern becomes more credible when pages with substantially the same purpose repeatedly compete for the same demand, the wrong page appears, or the user has no meaningful reason to choose between them.

Review the actual content, titles, internal links and search results before changing URLs or consolidating pages. A hasty redirect can remove a useful resource in the name of solving a problem that was never established.

Use CTR to choose an investigation, not manufacture a promise.

If a priority page earns relevant impressions but weak click response, inspect the current search result. Does the title explain the offering precisely? Does it match the query? Is the snippet helpful? Are local listings, ads, video or answer features affecting what the searcher sees?

A stronger title may help, but a low CTR is not proof the title is defective. Record the proposed change, the affected query group and the date. Compare a sufficiently observed period afterward, keeping other explanations visible.

There is no responsible universal promise that every page should reach a particular CTR. A page can improve its wording and still receive a different mix of searches in the next period.

Respect the missing rows.

Search Console omits some query information for privacy, and table totals can differ from chart totals. Filters also affect what is shown. Treat reported query rows as a useful but incomplete view, not a complete transcript of every search. Performance data limits.

A report saying you appeared for a certain number of visible query variations should keep that qualifier. Do not quietly turn it into a count of all market demand or every customer problem your site reaches.

When another subscription earns its place

Buy the answer to a recurring question.

There is no sensible universal threshold such as “buy Semrush when you reach 10,000 visitors.” A small specialist company competing for a few valuable contracts may have a stronger research need than a much larger site whose owners are not acting on the reports they already receive.

The purchase makes sense when the free tools leave a material question unanswered, the paid product has credible coverage of that question, and somebody can turn its output into work.

When Semrush is a reasonable candidate.

Consider it when you need recurring external search research: a defined keyword set to monitor, competitor search visibility to investigate, or a broader research workflow your team will use consistently. Its Position Tracking product monitors configured keywords and targeting; its research datasets use collection and estimation methods that differ from your first-party analytics. Position Tracking · Data and metrics methodology.

For example, a company entering a new market may need to identify which types of pages appear for a defined set of commercially relevant searches, what competitors explain well, and where an existing page needs stronger evidence. That is an external research problem.

“We do not know whether our form works” is a different problem. You can spend a considerable amount on competitor intelligence without moving one inch closer to solving it.

What would make me postpone the purchase?

• The principal outcome is not measured or verified, so the team cannot judge what its current traffic produces.
• Nobody can name the decision the subscription will improve.
• The team already has more validated work than it can implement.
• The research need is occasional and can be met by a scoped engagement or a short evaluation period.
• The actual market is so specialized that the provider's data coverage needs to be tested before any long commitment.

A niche phrase showing little or no estimated volume does not establish that nobody valuable uses it. Your own Search Console evidence, sales conversations and actual search results may reveal demand that a third-party database represents poorly.

Equally, a phrase showing substantial volume does not mean your business should pursue it. Research should narrow the work to the right demand, not fill the calendar with whatever the software happens to find.

Compare the real cost of the workflow.

Look at seats, projects, keyword limits, geography, historical access, exports, connectors and any separate products needed for the job. Check current plan terms directly. Base subscriptions and add-ons change, and the package name alone does not establish that the feature you need is included. Current Semrush plan information.

Then include the person's time. A tool can save labor, but it can also create hours of export handling and interpretation. Count the time you genuinely avoid, not every hour somebody spends inside the platform.