Google Discover optimization and traffic analysis

Discover recommends content based on individual interests rather than a typed keyword query. That makes a conventional keyword list an inadequate Discover strategy. I assess the site, editorial process and available reporting, separate evidence from assumptions, and define specific technical and content improvements with a way to review the outcome.

A Discover traffic change should be assessed separately from general technical SEO issues and the quality of on-page content, not inferred from one short-term graph.

Eligibility is not a promise of placement

Pages must be indexed and comply with Discover policies to be eligible, but there is no special submission or markup that guarantees distribution. Audience interests can change and Discover traffic can move even when a site has not made major changes. The first step is to establish whether the channel fits your readership and publishing model, rather than building a business forecast around an unguaranteed source of visits.

Read the Search Console Discover report correctly

Where the Discover Performance report is available, I examine clicks, impressions and CTR by page and time period, then compare patterns with publishing activity and site releases. Discover is not a keyword ranking report, and Search Console only exposes the report after a minimum impression threshold is reached. If the data is sparse, I flag that limitation instead of treating a few unusually successful articles as a stable baseline.

Identify genuine technical barriers

I review whether important articles are indexable, have coherent canonical signals, return appropriate HTTP responses and expose meaningful content to crawlers. Internal links, templates and mobile rendering can also affect the quality of the underlying pages. A technical defect should be documented with example URLs and a reproducible symptom; a site that is technically sound but receives little Discover distribution calls for a different conclusion.

Large, representative article images and previews

For selected articles I check image relevance, resolution, cropping and how the page declares its preferred preview image. Google recommends high-quality images at least 1,200 pixels wide, an appropriate landscape composition and permission for large previews via max-image-preview:large. The og:image tag or suitable structured data can signal a preferred image, but Google makes the final selection. Generic logos or misleading thumbnails are not a substitute for a strong editorial image.

Trustworthy headlines and original editorial work

The headline should reflect what the article actually delivers. Sensational claims, misleading previews and curiosity gaps that hide essential information may work against user trust and Discover guidance. I review original reporting, subject-matter expertise, citations, timeliness and clarity of the article. A plan based only on publishing more posts or rewriting other publishers’ work does not create a defensible long-term editorial advantage.

Choose topics through audience evidence

Rather than promising a list of winning Discover topics, I compare the site’s content themes with its audience, useful historical patterns and the expertise it can demonstrate. Recommendations can cover article formats, depth, image standards, updating practices and the editorial calendar. Both evergreen work and timely coverage may have a role, depending on the readership; the approach should not rely on a fixed publishing trick.

Diagnose Discover traffic drops without guessing

An unexpected decline needs a timeline: affected URLs, page-level impressions, click-through rate, recent releases and changes to publishing output. I also check indexing, preview imagery, article templates and relevant Google system updates. Interest shifts and distribution changes are plausible explanations, too. A traffic dip is not automatically a penalty, and a recovery date or former traffic level cannot be promised.

Deliver the audit as executable work

A useful audit identifies example URLs, supporting evidence, technical or editorial actions, priorities, ownership and validation steps. Changes within the team’s control—such as image delivery, markup, templates or editorial guidelines—are distinguished from results that cannot be guaranteed. Implementation and subsequent performance reviews can be scoped separately, so the engagement produces a practical backlog rather than just a generic checklist.