Practical guides on Google Search Console, BigQuery, and SEO data analysis.
Remove the 1,000-row cap and get unlimited keyword access. A complete step-by-step guide to setting up the GSC bulk data export, including IAM permissions, table schemas, and the mistakes that trip people up.
BigQuery cost controls are off by default. Learn how to implement five layers of protection — maximum bytes billed, custom daily quotas, budget alerts, partitioning, and dry runs — before a runaway query wrecks your budget.
Google Search Console deletes keyword data after 16 months, permanently and without warning. Here is how to preserve your entire keyword history using BigQuery, what you lose if you wait, and how to analyse it once it is safe.
Position #1 CTR dropped from 28% in 2024 to 19% in 2025. The benchmarks most people use are out of date. Here are the current click-through rates by position, what SERP features do to those numbers, and how to read your own data against them.
There are five distinct causes of high impressions with near-zero clicks in GSC, and each requires a different fix. Most guides treat them all the same — which is why most efforts to improve this don't work.
Most CTR advice focuses only on title tags. There are six levers — and some of the others have higher ROI. A systematic guide to title tags, URL structure, meta descriptions, schema markup, featured snippets, and AI Overview citation.
Most CTR drops in 2025 have nothing to do with your snippet quality. AI Overviews, the 50-week GSC impression bug, and zero-click search growth are the three structural causes to rule out before touching your title tags.
Page 2 keywords already have Google's trust signal. A keyword at position 15 with 5,000 monthly impressions generates 32 clicks. At position 3, those same impressions produce 510 clicks. Here is how to find and prioritize them.
Keywords ranking between positions 5 and 20 offer the best ROI in SEO — the foundational work is done, but the traffic potential is unrealized. A step-by-step workflow for extracting, grouping, diagnosing, and improving them systematically.
Moving five positions up the first page produces nearly 5x more traffic from the same search demand. The interventions that actually work — content depth, title tag alignment, internal linking, and page-level backlinks — and the ones that don't.
In November 2025, Google launched a native branded queries filter. Here is how to use it, what to do if your property isn't eligible yet, and why separating branded from non-branded traffic is foundational to accurate SEO reporting.
Branded and non-branded traffic have different conversion rates, different CTR profiles, and different growth drivers. Mixing them into one organic traffic number produces a reporting error that gets worse as the brand grows.
Non-branded organic clicks is the north star metric for any SEO program. How to set up the measurement, what different patterns in the data mean, and how to connect non-branded traffic to revenue attribution.
Keyword cannibalization is one of the most over-diagnosed problems in SEO. Ahrefs found that only 1 in 80 multi-URL ranking situations actually needs action. How to identify the genuine cases and ignore the false alarms.
Two pages showing up for the same keyword is not automatically harmful — it depends on domain authority, intent overlap, and whether the pages are swapping positions or holding stable ones. How to diagnose and decide.
Before choosing between a canonical tag and a 301 redirect, you need to know which page Google currently prefers — and whether that preference is stable or oscillating. The GSC method for determining the winner and the right fix.
The median publisher lost 10% of traffic in H1 2025. HubSpot lost 70–80%. The data on who is affected, by how much, and through which mechanisms — and the three other causes to rule out before attributing a traffic drop to AI.
70% of AI referral traffic shows up as Direct in GA4 because AI platforms don't reliably pass referrer headers. How to identify, measure, and track traffic from ChatGPT, Perplexity, Gemini, and Claude in your analytics.
60% of Google searches end without a click. On mobile it's 77%. What zero-click means for your impressions data, which queries are now structurally click-free, and how to adapt content strategy to a zero-click environment.
GSC deletes all keyword data older than 16 months, permanently. The BigQuery export is the only way to preserve it — and it only accumulates from the day you start. What you're losing right now and how to stop losing it.
YoY comparison removes seasonality from performance analysis. The GSC compare mode, the impression bug that distorts 2025–2026 data, the branded traffic problem, and the right framework for accurate year-over-year analysis.
A timeline correlation between a traffic drop and an algorithm update is not proof of impact. The before-and-after analysis method, the 2024 core update reference dates, and how to distinguish real algorithm impact from volatility.
GSC experienced multiple major data incidents in 2025–2026: a 50-week impression bug, a Page Indexing report freeze, and a Performance report outage. How to distinguish Google-side problems from actual site problems.
GSC doesn't classify keyword intent automatically. How to identify informational and commercial queries in your keyword data, what the mix reveals about your content strategy, and why this matters more in a world of AI Overviews.
Mobile and desktop don't rank the same — and since Google completed mobile-first indexing in July 2024, the mobile version is now the primary basis for ranking evaluation. How to compare, diagnose, and improve device-specific performance.
The most common failure in GSC client reporting is showing raw data without context. Which metrics to include, how to frame them for non-technical stakeholders, and how to automate extraction at the volume agencies need.
GSC has no native multi-client view. Domain vs URL prefix properties, permission levels, the 1,000-property limit, API rate limits, and the operational systems agencies need to manage GSC cleanly at scale.
The native branded queries filter only works on eligible properties. For everything else, this regex pattern excludes branded variations, misspellings, and product names — so you can see your actual non-branded SEO performance.