Tracking AI Overviews: rank tracker vs. SERP scraper
Framing: AI answer boxes don't behave like blue links, so your usual position data misreads them.

🔗 [Rank trackers with AIO flags] — tell you whether an AI Overview appears for a query and if you're cited; coverage varies by tool.
🔗 [SERP API scrapers] — pull the full AIO text and sources so you can analyze which page types get cited; more setup, more signal.
🔗 [GSC Performance] — clicks may fall while impressions hold when an AIO eats the click; the gap is your real AIO exposure.

The tradeoff: a rank tracker tells you presence, a scraper tells you why a competitor was cited and you weren't. For monitoring, trackers suffice; for strategy, you need the source-level scrape.

One to bookmark: a GSC impressions-vs-clicks divergence chart by query — the cheapest early warning that AIOs are reshaping your category.

Crawl budget: server logs vs. GSC Crawl Stats
This week on the radar — diagnosing why Google isn't recrawling fast enough after you fixed things.

🔗 [Server log analysis] — ground truth: every Googlebot hit, by URL, status, and time; nothing else shows actual crawl distribution.
🔗 [GSC Crawl Stats] — Google's own summary of requests, response times, and fetch purpose; easier, but aggregated and sampled.
🔗 [GSC URL Inspection 'last crawl'] — per-URL recrawl date when you need to confirm one important page got refreshed.

The tradeoff: Crawl Stats is fast and free but won't show you that 60% of crawl is being wasted on parameter URLs — logs will. On a large site recovering from an update, log analysis finds the crawl-budget leak; GSC only hints at it.

One to bookmark: a log-based 'crawl by directory' breakdown — it exposes where Googlebot is spending budget you'd rather it spent on money pages.

Internal linking: crawler graph vs. embedding suggester
Framing: post-update, redistributing internal links is a top lever — but the tool you use changes the recommendation.

🔗 [Crawler-based link graphs (Screaming Frog/Sitebulb)] — show real PageRank-style flow and orphan pages; structural, accurate, manual to act on.
🔗 [Embedding-based suggesters] — propose contextually relevant links by meaning; great for relevance, blind to crawl depth and authority flow.
🔗 [GSC + a custom 'pages with impressions but no internal links' join] — finds undersupported winners worth boosting.

The tradeoff: structural tools fix authority distribution; semantic tools fix topical relevance. The best internal links are both — relevant and from authoritative pages. Pick structural first when traffic concentrates on a few hubs; semantic first when content is broad and flat.

One to bookmark: an orphan-pages export crossed with impression data — it surfaces pages Google found anyway that you've never linked to.