Back to all posts
guides 12 min read

Is Your New Website Working? Reading Early Search Console Data

Serap Gündoğdu ·
Is Your New Website Working? Reading Early Search Console Data

Is Your New Website Working? Reading Early Search Console Data

Why Small Numbers Move on Their Own

Going from 20 impressions to 35 looks like growth of 75 percent. It is 15 impressions. Search Console counts an impression whenever your URL appears in a results page, even if nobody scrolls to it. One person who runs the same query on three devices, or checks page two a few times while comparing tools, can add that much. So can a single query that happened to rank 40th for a day.

At this scale one query or one searcher can move every metric. Before you read a number as progress, ask whether it carries meaning by itself. Try removing the largest query from the data. If the change disappears, you were looking at one query, not a trend.

Seodisias gives a concrete case. From 2026-07-16 to 2026-08-12, Search Console showed 8,430 impressions, 58 clicks, an average position of 27.4 and 316 queries. Those clicks are a click-through rate (CTR) of about 0.7 percent (58 divided by 8,430). That is a modest start, and the details make it look weaker. Two queries, “seo crawler” (899 impressions) and “screaming frog alternative” (841), supply roughly a fifth of all impressions, and both sit around position 26 to 30. The brand query “seodisias” brought 7 of the 58 clicks at position 3.3, so some of the clicks came from people who already knew the name.

None of this proves the site is failing. It shows that the total is carried by a handful of rows, and that a handful of rows can swing it either way. If the 899-impression query drops 10 positions next week, the headline numbers will fall even if nothing about the site got worse.

Compare 28-day windows rather than days, and treat any jump you cannot trace to more than one query as unconfirmed. Also check Google Search Status for ranking updates or incidents that fall inside the window, since a sitewide wobble can have an outside cause. The next sections cover what to check in its place.

Check Indexing First, Then Visibility, Then Clicks

Search Console data has an order, and each step depends on the one before it. A page that is not indexed cannot appear for any query. A page that appears for no query cannot earn a click. If you read the numbers backwards, starting with clicks, a missing click tells you nothing about where the chain broke.

Start with indexing. Open the Pages report in Search Console (the Search Console Help Center defines each status) and compare the count of indexed URLs with the number of pages you published. Then paste two or three newly published URLs into URL Inspection. If a page sits under “Discovered, currently not indexed” or “Crawled, currently not indexed”, the problem is upstream of everything else, and no amount of staring at the Performance report will fix it. Check that the sitemap was processed and that robots.txt and canonical tags are not blocking the pages.

Once pages are indexed, move to visibility. Open the Performance report and look at how many different queries your pages appeared for, and at which positions. Impressions are the first thing a young site earns. They show that Google has placed your page in front of someone, even if that someone never scrolled to it.

Clicks come last, and they lag for a mechanical reason. A click needs a page that ranks high enough to be seen, a title that earns the choice, and a searcher who wants that answer. Seodisias shows what this looks like. In the same July 16 to August 12 window, its 316 queries produced a CTR of about 0.7 percent at an average position of 27.4. Position 27 sits on the third page of results, so a low CTR is what you would expect there. The 316 queries are the more useful number, because they show how widely the site had started to appear.

In the first two months, ask these questions in this order:

  1. How many of my pages are indexed?
  2. How many distinct queries did they appear for, and is that count growing?
  3. Only then: did any of those impressions turn into clicks?

If the first two answers are improving, a low click count is a delay. If the first answer is poor, fix indexing before you read anything else.

Count Distinct Queries, Not Only Impressions

Take 1,000 impressions. If three queries produce them, one ranking change or one person’s repeated searches can erase most of that number next week. If a hundred queries produce them, no single query can move the total much. The sum is identical and the two situations are not.

The Seodisias window shows the same split. Its 8,430 impressions came from 316 distinct queries, about 27 impressions per query. The concentration is visible in the table: “seo crawler” (899 impressions) and “screaming frog alternative” (841) together account for 1,740 impressions, about a fifth of the total. The crawler cluster adds 56 queries and 2,207 impressions at positions 17 to 42. The Screaming Frog alternative cluster adds 88 queries and 1,659 impressions. The meta tag cluster has 30 queries and 529 impressions, but at positions 45 to 92, where almost nobody scrolls.

So the site has breadth with no depth. Hundreds of queries show it, and nearly all of them sit too far down to earn clicks. That is a different problem from having no visibility, and it calls for different work: improving pages that are already being shown, not hunting for indexing faults.

To track this, open the Performance report, switch to the Queries tab and note the row count at the end of each 28-day window. Search Console leaves some rare queries out of that table, so the count understates the real number. Because you read it the same way every time, the comparison between windows still holds.

A rising impression total over one window tells you little. A rising count of distinct queries over several consecutive windows tells you more, because a random fluctuation rarely adds new queries three times in a row. Pages that start matching new searches are the first sign that Google is placing your content, and that sign appears well before clicks do. If the query count stays flat while impressions jump, check whether one query produced the jump. Then treat it as a spike, not as growth.

Write down the query count, the impression total and the top query’s share of impressions at the end of each window, and compare the three numbers each time.

Separate Brand Queries From Generic Ones

Distinct queries are only useful if you know who is typing them, and Seodisias’s own numbers show why that split matters. Of the 58 clicks in the July 16 to August 12 window, seven came from a single query, seodisias, at an average position of 3.3.

That query is the brand name. Nobody types it without already knowing the site exists. The people who did are mostly people we know: colleagues, friends, someone who saw a post, and the team itself checking whether the site shows up. A brand query at position 3.3 tells you Google can match a name to a domain. It says nothing about whether the site can rank for the problems it is meant to solve.

Generic queries are the opposite case. Nobody searching seo crawler or screaming frog alternative has heard of Seodisias. A click there means a stranger found a page through a topic. In the same period those two clusters produced 3,866 impressions at positions between 26 and 30, which is a very different story from position 3.3.

Split the data before you read it:

  1. Open the Performance report and filter Queries with “Doesn’t contain” and your brand name, including common misspellings and the domain without the extension.
  2. Record the brand and non-brand totals separately, for clicks and for impressions.
  3. Compare each against the previous window of the same length.

On a young site, the brand line will often be the larger share of clicks. That is normal and not a problem. The mistake is reporting the combined total as progress. Judge the site on the non-brand line: how many distinct generic queries appear, and at which positions. Brand traffic is worth tracking too, but as a separate number with its own meaning.

A Decision Table for Early Search Console Data

Look at your numbers through the same lens every time, then match what you see to the table below. Use 28-day windows at minimum and compare each one to the previous window of the same length. Daily and weekly swings on a young site are mostly noise.

For position, skip the average. It is weighted by impressions, so one deep query with thousands of impressions can drag it down while your visibility is actually growing. Count queries in four buckets instead: positions 1-3, 4-10, 11-20 and 21+. Progress shows up as queries moving from the 21+ bucket into 11-20, or from 11-20 into 4-10.

What you seeLikely explanationWhat to do
Impressions, but zero clicksPositions are deep (21+), so almost nobody scrolls that farCheck the position buckets first. Rewrite titles only for queries already in the top 10
Average position stuck at 25-30Content is indexed but has not earned a place on page one for competitive queriesLook at the 11-20 bucket for queries to improve. Leave the rest alone for another window
Sudden spike, then a dropOne query or a short-lived test placement, not a trendCheck whether a single query caused it. Do nothing unless it repeats in the next window
Average position gets worse while impressions riseNew deep queries joined the mix and pulled the average downCompare the buckets. If 21+ grew and 1-10 held, visibility improved
No impressions at allPages are not indexed, blocked, or not yet discoveredInspect the URL in Search Console, check robots.txt, canonical tags and the sitemap
Clicks come almost only from your brand namePeople who already know you are searching for you, which says nothing about generic rankingsFilter out brand queries and judge the rest separately
Impressions flat, distinct queries growingTraffic is spreading across more queries, which is a healthy shapeKeep publishing and compare again next window
Totals higher than the query table adds up toSearch Console withholds some anonymized queries from the tableTrust the totals and treat the table as a sample

Seodisias’s window fits several rows at once: an average position of 27.4, a CTR of about 0.7 percent, and 7 of 58 clicks from the brand query “seodisias”. Under the table, that means checking the position buckets before touching any titles, and judging generic rankings only on the non-brand line.

When to Wait and When to Check Indexing

A new site’s first weeks call for technical checks, not measurement. If a page has no impressions yet, the first question is whether Google can see it at all, and Search Console answers that faster than the Performance report does.

Run these checks on each new page:

  • URL Inspection: Paste the URL and read the verdict. “URL is on Google” means the page is indexed. “Discovered, currently not indexed” means Google knows the URL but has not crawled it yet. “Crawled, currently not indexed” means it fetched the page and chose not to keep it.
  • Sitemap: Open the Sitemaps report and confirm the file shows a success status and that the discovered URL count matches the number of pages you published.
  • Canonical: Compare the user-declared canonical with the Google-selected canonical in the inspection result. If Google picked a different URL, the page you are watching may not be the one being indexed.
  • Robots and meta robots: Check that robots.txt does not disallow the path and that the page carries no noindex directive, in the HTML or in an HTTP header.
  • Rendering: Use the live test to confirm that the main content appears in the rendered HTML, especially on JavaScript-heavy templates.

A crawler run over your own site covers the same ground in bulk, so you do not have to inspect fifty URLs by hand. It flags noindex tags, canonical mismatches and blocked paths in one pass.

If all of these come back clean, waiting is a legitimate decision. With a handful of impressions, any change you make cannot be separated from random movement, and rewriting titles or restructuring pages every few days erases the one thing you need, a stable baseline. Google publishes no fixed timeline for when a new page starts ranking, and the Search Central documentation only describes how crawling and indexing work, so do not put a date on it.

Today, pick your three newest pages and run URL Inspection on each. Write down the verdict and the selected canonical, then check again in 28 days against the same list.

Looking for something else?

Search every guide on the blog without going back to the list.