field notes
How to attribute every visit to real revenue: AI traffic, dark social, tracked links and privacy-first measurement.
ChatGPT sends real, high-converting traffic. Here are three ways to track it (GA4, privacy-first analytics, server logs), the full AI referrer list, and what every method misses.
Cohort analysis sounds like a data-team luxury. It is actually the simplest way to see if your product and your marketing are getting better or worse. Here is the simple version.
Stripe knows revenue. Analytics knows traffic. Connecting them is how you finally see which channel brings paying customers. Here is the setup, step by step.
Average LTV is a lie that hides your best and worst channels. Measuring lifetime value by where customers came from is the number that should set your budget. Here is how.
Your analytics counts visits. Your payment tool counts revenue. Neither connects the two, so you cannot see which marketing brings profitable customers. Here is how to close the gap.
Pageviews are the metric you optimize when you do not know what else to measure. Revenue per visitor is the one number that ties traffic to the bank. Here is why to switch.
The AI crawlers that matter in 2026: what GPTBot, ClaudeBot, PerplexityBot, ChatGPT-User and the others do on your site, checked against each company's documentation, and how to control each one.
AI bots crawl your pages; how many visitors do they send back? How to compute the crawl-to-refer ratio introduced by Cloudflare for your own site, per engine, and what it can and cannot tell you.
Adobe's retail data says visits referred by AI assistants now convert better than other traffic, after converting worse in 2025. What the numbers measure, what they leave out, and how to check on your own site.
GEO tools track brand mentions from $39/mo. GA4 counts clicks. Neither ties AI traffic to revenue. Here is what complete LLM referral tracking looks like in 2026.
Perplexity cites sources on every answer, which makes it the most referral-friendly AI engine. How to track perplexity.ai traffic and earn more citations.
A tracked link is a measurement point that survives copy-paste into chats, bios and newsletters, where UTMs die. Here is how to use them to rank your channels by revenue.
Most QR codes are dumb links to a fixed URL. A tracked QR code turns packaging, posters and events into measurable channels with scans, sources and revenue. Here is how.
Links shared in private channels lose their referrer and land in 'Direct'. Here is what dark social is, and the fix that works: stamping the source on the link.
UTM parameters work until a human touches them. Tracked short links survive forwarding, chats and bios. Here is the honest comparison and when to use each.
Meta and Google both claim the same conversion, then both take credit. The only honest ad ROI is measured independently, on your side. Here is why and how.
Your ESP reports opens and clicks, then goes silent. Real email ROI lives on your site, after the click. Here is how to attribute email campaigns all the way to revenue.
Email, SMS, social, ads, organic and AI, each measured in its own silo with its own rules, is how you end up unable to compare anything. Here is how to put every channel on one honest scoreboard.
Texts get read, but SMS clicks often land in 'Direct'. One tracked link per campaign fixes it. Here is how to measure SMS marketing in revenue, not taps.
Likes and reach are vanity metrics. The social number that matters is revenue per post. Here is how to measure it with tracked links, even for dark social.
Most "EU analytics" lists confuse EU hosting with EU ownership. Here is the list that separates them: who owns each tool, where the data lives, and what that means under GDPR.
What GDPR compliance actually requires from web analytics: personal data, consent, cookies, transfers. A no-FUD guide with an audit checklist for your current setup.
Every major ruling, fine and decision on Google Analytics and EU privacy law, in one maintained timeline: Schrems I and II, the 2022 DPA wave, the Swedish fines, and the case pending now.
A French MP took the EU-US Data Privacy Framework to court, lost at first instance in September 2025 and appealed to the CJEU (C-703/25 P). The case in plain language, its current status and sources.
The EU-US Data Privacy Framework is under appeal at the CJEU. Here is what actually happened, what history says comes next, and how to prepare your analytics stack.
A/B testing finds the best version for everyone. Personalization shows different versions to different people. They are complements, not rivals. Here is the decision.
One homepage for every source leaves money on the table. Matching the landing page to where traffic came from lifts conversion, if you measure it. Here is the playbook.
Where a visitor came from tells you what they came for. Segmenting by acquisition source is one of the most useful, lowest-effort ways to personalize. Here is how to do it.
A practical intro to website personalization: the types that work, where to start, and the measurement layer that decides whether any of it sells. No hype.
Personalization fails when it runs on guesses. Before changing what visitors see, you need to know who they are and what converts them. Here is the measurement layer that makes dynamic sites actually sell.
OpenAI started testing ads in ChatGPT in February 2026. That splits 'AI traffic' into paid and organic, two channels to track apart. Here is why, and how.