Almost everyone new to SEO tools makes the same five mistakes, and all of them come from treating a tool's output as fact rather than as an estimate produced by a model. Here they are, with what to do instead.
Third-party search volumes are modelled from clickstream panels and Google's own bucketed ranges. Two tools will report 2,400 and 5,400 for the same phrase, and both are estimates of an average over twelve months that hides the seasonality you actually care about. Use volume for relative comparison — is this term much bigger than that one — and use Search Console's impressions for anything you already rank for, because those are counts rather than estimates.
DA, DR and similar scores are vendor metrics built from each vendor's own link index. Google does not use them and does not see the same index. Two consequences follow: a page can rank above a much higher-DR competitor routinely, and buying a link because a tool scores the site 60 tells you about that tool's crawl, not about traffic. Judge a link on whether the page it sits on has real readers.
A desktop crawler defaults to fetching raw HTML. If your site renders content client-side, the crawler sees an empty shell and reports thin content everywhere — or worse, sees content that Google does not. Turn on JavaScript rendering in the crawl configuration and compare the two runs. The difference between rendered and raw is one of the most useful diagnostics available and almost nobody looks at it.
Disallowing CSS, JavaScript or framework build directories in robots.txt feels tidy and breaks rendering entirely — Google fetches the page, cannot load the stylesheet or the bundle, and evaluates a layout that no user ever sees. Restrict robots.txt to real routes you genuinely do not want crawled, and use the URL Inspection tool's rendered screenshot to confirm what Google actually assembled.
"Excluded by noindex tag" is not an error if you meant to noindex those pages. "Crawled - currently not indexed" and "Discovered - currently not indexed" are different diagnoses: the first means Google fetched the page and judged it not worth indexing, which is a content problem; the second means it never fetched it, which is usually a crawl-budget or internal-linking problem. Treating both as one bucket leads people to resubmit sitemaps for months while the actual cause goes untouched.
site:example.com returns a rough, heavily approximated figure that changes between refreshes and has never been an index count. The Pages report in Search Console is the only number worth acting on, and it comes with reasons attached.
Connect Search Console, export the queries you already get impressions for, and find the pages ranking in positions 5 to 15. Those are real, measured, and closer to converting than anything a keyword tool will suggest. Tools are for scale, once you know what you are scaling.