The Bot Traffic Doesn't Lie Either

The Digital Karma Data Warehouse already imports Search Console data nightly for every active site in the portfolio, and already classifies every server log request against a maintained bot fingerprint table for security and traffic-quality reasons. Neither system was built to talk to the other.

The original question was practical: which of Krisada's 11 portfolio constellations already have enough real traffic proof, human clicks and AI reads both, to credibly sell business exposure inside that niche. Answering it meant joining GSC clicks and impressions against bot-log activity, grouped by site and by constellation, for the first time.

GSC clicks, impressions, and distinct queries were pulled from gsc_query_daily for a 90-day window. Server log requests were pulled from log_requests joined to bot_fingerprints, filtered to visitor_class ai_crawler, plus visitor_class bot rows with a fingerprint priority of 12 or lower (Googlebot, Bingbot, DuckDuckBot, YandexBot, Baiduspider). That excludes SEO tools like Ahrefs and Semrush, social preview bots, and uptime monitors, all of which also show up as bot traffic in raw logs but say nothing about search or AI discovery. The log window was 48 days, the full range currently retained.

Both datasets were grouped by site, then rolled up by constellation using the same portfolio-constellation mapping the warehouse already uses for GSC reporting.

Healthcare & Longevity is sixteen sites. Watched through Search Console alone, it looks close to dead: eighty-three total clicks across all sixteen properties over ninety days. Watched through the bot log, it is the single most AI-crawled category in the entire portfolio, over nine thousand good-bot requests a week combined, ahead of every other constellation by a wide margin.

The clearest single example is naturalherblibrary.com. Over the same window it shows 31 Google clicks and 1,411 impressions, numbers that read as thin on their own. Its server logs show 19,045 AI-crawler requests from nine distinct vendors: OpenAI, Anthropic, Perplexity, Common Crawl, ByteDance, Amazon, Meta, Cohere, and DatasetSEO's own observatory crawler. That is not a coincidence worth ignoring. NaturalHerbLibrary.com is the live test property for Real SEO™ Experiment No. 001, whose hypothesis is that a domain built for AI citability earns AI-system attention before it earns meaningful Google traffic. This is the first hard data point supporting that hypothesis: heavy, sustained, multi-vendor AI-crawler activity months ahead of any meaningful Google click signal.

Medical Marketing showed the opposite pattern and an equally important one: 10,521 impressions and consistent AI-crawler activity across all four sites, and exactly zero Search Console clicks in ninety days. That is not a visibility problem. Google is already showing those pages to searchers who see them and pass. The bot data proves the reach exists; the listing itself is what needs fixing.

Across the full portfolio, only two of eleven constellations, Healthcare & Longevity and One Mouse Marketing, cleared both bars at once: real human clicks and heavy AI-crawler reads, together. krisada.com and realseolife.com carried the two highest 90-day click counts in the whole portfolio while both pulling thousands of AI-crawler requests a week.