Your SEO agency’s monthly report opens with a chart: “Google impressions are up 40%.” You read it and still can’t answer the basic questions. Which pages are people finding? What searches bring them to the site? Has your main service page dropped out of search? And does that 40% mean 40% more people saw your site?
Search Console answers specific questions about how your site shows up in Google Search: which queries and pages it appears for, how many clicks it gets, and whether the pages that matter can be found. Every number needs context: the property, search type, date range, and filters. Impressions aren’t people, and clicks aren’t customers.
This article assumes your site is already verified in Search Console and you have access.
What Search Console can tell you about your site
Search Console is Google’s free tool for site owners. It shows where your pages appear in search results, whether Google can crawl and index them, and whether there are issues Google wants you to know about.
It helps to separate three questions that different sources answer:
| Question | Where to find the answer | What that source doesn’t show |
|---|---|---|
| Does your site appear in Google Search, and for which queries? | Search Console | What people did on the site after they clicked |
| What did visitors do on the site? | Website analytics, such as Google Analytics 4 | How many leads became paying customers (unless payment data is sent in and matched to leads) |
| How many leads turned into sales? | CRM, order management, accounting system | Where the customer came from, if the source wasn’t passed along |
Search Console stops at the click. It doesn’t know whether a visitor filled out a form, and it doesn’t show revenue. How search data connects to the rest of the customer journey, and where leads get lost along the way, is covered in “Where businesses lose money between search and sale.”
Most business owners only need three parts of Search Console: the Performance report, with clicks, impressions, CTR, and position; the Page indexing report; and the URL Inspection tool for a single URL. Search Console also includes the Core Web Vitals report on loading speed, responsiveness, and page stability, as well as the Manual actions and Security issues reports.
Where to start
Wrong conclusions often form before anyone looks at the numbers. Before discussing growth or decline, check four things.
Property. A Domain property covers every URL on the domain: with and without www, both http and https, and all subdomains. A URL-prefix property covers only URLs that start with that exact prefix, including the protocol, such as https://www.example.com/. If your site runs on https://example.com but the report comes from the www property, you’ll see partial data or none at all. Language versions in subfolders like /es/ belong to the same property, so you need a page filter to look at one language.
Search type. The Performance report shows web search by default. Image, video, and news search are counted separately; for an online store that gets a lot of traffic from image search, look at that data on its own.
Date range. The default is the last three months. Compare periods of equal length that start on the same day of the week, such as the last 28 days versus the previous 28. For a seasonal business, also compare with the same period last year. If the report marks the most recent days as preliminary (they appear as a dotted line on the chart), don’t draw conclusions from them.
Filters. Active filters appear above the chart: query, page, country, device, and search appearance. A report filtered to “United States” or “Mobile” can’t be compared with an unfiltered one. Any screenshot from your agency should show the filters and dates.
How to read clicks, impressions, CTR, and position
The core metrics follow specific counting rules, and those rules determine what you can and can’t claim.
An impression counts when a link to your site appears on a results page someone loads. For a standard result, the user doesn’t have to scroll down to it; for carousels and results that load as the user scrolls, an impression counts only when the item comes into view. So an impression doesn’t mean someone read your listing. And an impression isn’t a person: someone who searches three different phrases and sees your site each time generates three impressions. Search Console doesn’t count unique people, so “20,000 people saw our site” based on 20,000 impressions is wrong.
A click is a visit from Google search results to a page outside Google Search. It isn’t the same as a session in analytics, let alone a customer.
CTR (click-through rate) is clicks divided by impressions, shown as a percentage: CTR = clicks / impressions × 100%. If there are no impressions, CTR can’t be calculated; mark the row “no impressions,” not 0%.
Average position. Position 1 is the top. For each impression, Search Console records your site’s highest position on that results page, and the report shows the average. If your site ranked second for one query and third for another, and both had the same number of impressions, the average is 2.5; a query with more impressions carries more weight. Average position doesn’t describe where your site ranks for each query. Here’s a possible scenario: rankings for your main queries hold steady, but the site starts appearing for dozens of new queries in positions 40–60. The average gets “worse,” even though nothing changed for your main queries. That’s why you should look at position query by query, or for defined groups of queries.
Illustrative example: more clicks, lower CTR
Illustrative example. The figures are hypothetical, for illustration only.
| Period | Impressions | Clicks | CTR |
|---|---|---|---|
| Previous 28 days | 10,000 | 300 | 300 / 10,000 × 100% = 3% |
| Last 28 days | 20,000 | 400 | 400 / 20,000 × 100% = 2% |
Clicks grew by a third, while CTR fell from 3% to 2%. These rows don’t show that your titles in search results got worse (lower CTR can come from new queries at lower positions), that 20,000 different people saw the site, or that sales went up. To understand what happened, break the totals down by query and page.
How to find changes by query and page
The chart shows that something changed; you find the cause by breaking the data down by query, page, country, and device. We recommend this sequence:
- Turn on date range comparison and select all four metrics. What exactly changed: clicks, impressions, CTR, or position?
- Find the queries with the biggest change in clicks or impressions. Did existing queries lose clicks, or did new ones appear?
- Find the pages with the biggest change and filter by one of them to see the queries for that page only.
- Check countries and devices. A drop only on mobile or only in one country points to different causes than a sitewide drop.
- Line up the dates with site updates, URL changes, a migration, or seasonal patterns.
A single sitewide number rarely tells you much. Make a list of the pages that matter—core services, main categories, top products, the contact page—and track them separately. Search Console supports regular expression (regex) filters, so an SEO specialist can set up one filter for the whole list.
Branded and non-branded queries
Branded queries include your company name: these people already know you. Non-branded queries describe a need, such as “boiler repair chicago” or “buy water filter.” The first group depends mostly on brand awareness, advertising, and referrals; the second on how visible your site is when people search for what you offer.
Search Console offers a branded queries filter that automatically sorts queries into the two groups. It isn’t available for sites with low impression volume, and Google notes that some queries may be misclassified. If the filter isn’t available, your SEO specialist can create one based on your company name and its spelling variations and document the rule so the numbers are calculated the same way every month. Any split like this is an approximation: misspellings, brand names made of common words, and rare queries hidden for privacy all blur it.
How to check whether your pages are indexed
For a page to show up in search, Google has to discover it, crawl it, and add it to the index. The Page indexing report shows how many of the URLs Google knows about are indexed, how many aren’t, and why.
A large number of non-indexed URLs can look alarming, but Google says outright that you shouldn’t expect every URL on your site to be indexed, and a “Not indexed” status doesn’t necessarily mean something is wrong. These URLs typically include duplicates that correctly point to a main (canonical) version; old URLs that redirect; deliberately excluded pages, such as the cart, account pages, and internal search results; and URL variations with filter parameters.
The business question is different: are the pages that should bring in customers indexed? Check each important page. Watch for these reasons:
- the page is excluded by a
noindextag or blocked from crawling in robots.txt, the file where a site tells search engines which sections not to crawl; - the URL returns an error (404 for “not found,” 5xx for a server error) or redirects to another page;
- Google treats the page as a duplicate and has chosen a different canonical URL;
- the page is listed as “Discovered – currently not indexed” or “Crawled – currently not indexed,” meaning Google hasn’t added it to the index yet.
The same reason can be normal for a utility page and a problem for a service page. Your SEO specialist sorts out the technical causes; your job is to notice that an important page isn’t indexed.
The URL Inspection tool gives two answers. Indexed version data describes the version Google last indexed: whether the page is indexed, when it was crawled, and which canonical URL Google selected. Test live URL checks whether the page in its current form can be indexed; it doesn’t evaluate duplicates or canonical selection. Requesting indexing puts the page in a crawl queue but doesn’t guarantee it will be indexed, and there’s a daily limit on requests. A promise like “we’ll submit it for indexing and it’ll be in search tomorrow” doesn’t reflect how the tool works.
Why the numbers don’t always match
Search Console numbers won’t exactly match your analytics, or even agree across different parts of the same report. That’s normal; what matters is knowing why.
Within Search Console:
- some rare queries are left out of the table to protect user privacy, and the table only shows the most important rows; the totals above the chart include this data unless a query filter is applied, so the rows can add up to less than the total;
- the totals above the chart are always calculated for the site as a whole, while rows in the Pages table are calculated per page: several links to your site on the same results page count as one impression in the total but as several in the Pages table;
- the interface and its exports limit the number of rows; a specialist can get more data through the API or bulk data export, but rare queries stay hidden there too.
Search Console and your site analytics count clicks and sessions differently. Here’s why:
- the analytics code isn’t installed on every page;
- some visitors opt out of tracking;
- standard Search Console reports use Pacific Time (the last-24-hours view uses your browser’s time zone), while analytics uses the time zone set in your analytics property;
- analytics attributes traffic sources using its own rules;
- Search Console attributes clicks to the page’s canonical URL;
- clicks to PDFs without analytics code don’t show up in analytics;
- analytics filters out some bots.
Compare trends, not exact counts. Illustrative example. For several months, you saw about 850 sessions for every 1,000 clicks; this month it’s 400. That’s a sign to check your analytics setup or cookie consent settings. The gap between 1,000 and 850 on its own isn’t a problem.
How to hand off a problem to an SEO specialist
You don’t need to diagnose the problem yourself. Your job is to spot the signal, gather context, and give the specialist a task backed by data.
Signal table
The explanations in this table are possibilities, not automatic diagnoses: a check confirms or rules out each one.
| Signal | What to check | Possible explanation | Who handles it |
|---|---|---|---|
| Clicks to a key service or product page dropped | Compare periods of equal length and the same period last year. Find the page’s queries with the biggest loss: did impressions, CTR, or position fall? Devices, countries, the start date. URL Inspection. Change log: URL, content, redirects, canonical URL | Seasonal dip in demand; lost rankings for some queries; a changed URL or canonical page; the page dropped out of the index; the results page layout changed, or the title or description Google displays changed. The title is just one possible cause and gets checked alongside the others | SEO specialist; developer, if the drop coincides with a release |
| Impressions rose, CTR fell | Which queries and pages drove the increase, and at what positions. Whether clicks grew in absolute numbers. CTR for queries that were already there | New, broader, or informational queries at lower positions; not necessarily a decline. If CTR fell for existing queries at the same positions, check how the page looks in search results | SEO specialist, if CTR fell for important queries at unchanged positions |
| An important page isn’t indexed | Whether the page should be in search. URL Inspection: status, reason, canonical URL, crawl date. noindex, robots.txt, errors, redirects. Internal links, sitemap | The page is excluded from indexing; a redirect or error; Google treats it as a duplicate; crawled but not yet indexed by Google; the page is new and Google hasn’t crawled it yet | SEO specialist for diagnosis; developer, if the cause is technical |
| Average position got worse, clicks are stable | Whether new queries appeared at low positions. Positions for a fixed list of main queries | New queries at lower positions can pull down average position without any losses for main queries | Needs a closer look only if positions for main queries fell |
| The ratio of clicks to analytics sessions changed sharply | Comparison with previous months. Analytics code, cookie consent banner, template changes | Analytics stopped firing on some pages; consent settings changed | Marketing analyst or developer |
Also flag for a closer look: a sharp increase in URLs returning server errors or 404s, a drop in impressions across all query groups at once, and any Manual actions or Security issues notices.
Handoff brief
The brief turns an observation into a task you can paste into an email or project management tool. The example shows the fields.
Illustrative example. The brief uses hypothetical data.
| Field | Value |
|---|---|
| Page or group | /services/boiler-service/ |
| Period and comparison | Last 28 days vs. previous 28; also the same period last year |
| Property and filters | Domain property; web search; United States; all devices |
| Observation | Clicks 180 → 95, impressions 6,200 → 5,900, CTR 2.9% → 1.6%; the drop started on the 14th, mostly on mobile |
| Already checked | No drop in this period last year. The page is indexed, and the canonical URL matches. The service page template was updated on the 13th |
| Hypotheses | 1) the title or description shown in search results changed after the template update; 2) some queries lost rankings on mobile |
| What to check next | CTR and positions for the page’s queries before and after the 14th; how the page looks in search results on a phone; the template changes made on the 13th |
| Owner and definition of done | SEO specialist; done when the cause is confirmed, or both hypotheses are ruled out and a follow-up check is proposed |
Next steps
- You or your head of marketing lists the important pages and confirms with your agency which property, filters, and date ranges the monthly report uses.
- The SEO specialist sets one fixed rule for separating branded from non-branded queries and one fixed list of important pages, checks their indexing, and documents the reasons for any that aren’t indexed.
- At every review, you follow the same order: the same property, search type, and filters → period comparison → important pages → indexing of important pages → Manual actions and Security issues reports → signal table. For every notable signal, your agency provides a completed handoff brief.
Your review process is set up when the report always comes from the same property with the same filters, the indexing status of your important pages is known, and every signal has a brief with an owner and a definition of done. How often you review depends on how often you make decisions about the site.
Let’s discuss which search changes on your site need a closer look.
