The first 90 days in Search Console: learn before declaring victory.

Start with the correct property

A new publication needs a reporting rhythm before it needs a celebration chart. Treat the first ninety days as an observation plan, not a promise of when search traffic will arrive. Begin by verifying the correct Search Console property and recording the public domain, launch date and major changes. If a temporary hosting address is used before a final domain, document the transition so that reports do not silently mix two different periods. This guide describes a proposed workflow; it does not claim access to NUMU's Search Console or any measured traffic results.

Separate eligibility from visibility

During the first stage, choose a small set of important article and topic pages to inspect. Google's URL Inspection tool distinguishes information about an indexed page from a live test of whether a page might be indexable. Neither a successful live test nor a request for indexing is a guarantee of search appearance. Record the status, the exact URL and the issue requiring attention. An inaccessible page needs a different response from a useful page that has not yet attracted impressions. Do not interpret every quiet report as a demand problem.

Read four metrics together

Read clicks, impressions, click-through rate and average position together. Google's Performance report defines click-through rate as clicks divided by impressions, and average position refers to the topmost result for the relevant grouping. These are search metrics, not sales or unique-customer counts. In a clearly hypothetical example, twenty clicks from one thousand impressions produce a two per cent click-through rate. That arithmetic does not explain why people clicked, whether they read the page or whether the audience was commercially relevant. Keep the denominator visible whenever presenting a percentage to management.

Compare like with like

During the next stage, compare equivalent periods and maintain the same filters. A change in country, device or search type can alter the story without any change to the article. Keep newly published pages separate from older pages with a longer observation history. Inspect query and page combinations where the data permits, while recognising that a displayed table is a reporting view rather than a complete record of every search. For a small site, a few additional clicks can create a dramatic percentage increase. Show the underlying counts alongside the percentage.

Connect search learning to editorial decisions

Use the evidence to choose a specific editorial response. A page appearing for an unexpected question may need a clearer explanation or a different internal link. A useful page with no discovery may require technical inspection or better connections from relevant published pages. A title revision should accurately describe the existing answer, not introduce an unsupported promise. Record one change, its reason and the observation window. Where several changes happen together, acknowledge that the report cannot cleanly isolate their effects. Treat the next reporting cycle as a learning opportunity rather than a guaranteed experiment result.

End with a decision, not a ranking promise

At the end of the proposed ninety-day window, produce a short decision memo: what was published, what became discoverable, which questions showed signs of demand, and which pages need repair or deeper evidence. Include genuine business outcomes only when a separate, documented measurement process supports them. Decide which editorial desk deserves further investment and which assumptions remain untested. Do not label the site number one based on a selected query screenshot or a favourable average. A responsible growth report makes the next decision easier and preserves uncertainty where the evidence is incomplete.