Himanshu Sharma @analyticsnerd

Check out the latest article in my newsletter: Voice AI This Week - June 10, 2026 linkedin.com/pulse/voice-ai… via @LinkedIn

My first time selling paid articles (on Voice AI). The old content model for generating leads and sales from organic search traffic is dead. Free web content has become unpaid training data for AI companies. No matter how much content you develop for free web, it all gets gobbled up by AI bots. The best you can hope to get is brand mention, which has little to no commercial value. As per Cloudflare, more AI bots now visit your website than human visitors. Bots now generate ~57.5% of HTTP requests to HTML content vs. 42.5% human. I am not feeding the scrapers for free anymore.

I don't feel the need to learn anything more about GA4 BigQuery, which helped me to move on to Voice AI. The foundational prompt library is already there. I can just tweak the prompt and build something more complex within seconds. If you are "stuck" learning the same tech/tool for too long, you are at a great risk of lagging behind in your career. You need to know a lot of things (Digital Analytics, Claude Code, n8n...), not just GA4, BigQuery, or GA4 + BigQuery.

This multilingual voice agent can reply in the caller's most recent language on every turn. I start the conversation in Spanish, then switch to English, then to Hindi, and then back to Spanish. The agent prompt and the Retell new multilingual feature made this possible. So if your caller starts in English but is more comfortable speaking in Spanish, the agent can switch languages mid-call. No need to create language-specific agents. I will share more details in my voice newsletter.

Check out the latest article in my newsletter: Google Analytics Scroll Tracking Report in BigQuery linkedin.com/pulse/google-a… via @LinkedIn

The 'AI Assistant' channel group has finally rolled out in GA4, and the advent of the new traffic medium 'ai-assistant'.

Most non-developers can't build production-ready agents even with Claude Code, but if you show them a way, give them hope, they fall into this loop. That's why you never share your proprietary prompts/systems/findings for free in the hope they will later hire you for paid work. They won't.

I have my own self-improving AI warning. It is going to be a disaster with current LLM failure modes: Hallucination, Context erosion and Overcorrection. When the base model carries unresolved failure modes, those flaws don't just persist, they can amplify catastrophically.

We just hit $20M in ARR. Here is how we did it. ………. Majority of success stories on social media are either fake or greatly exagerrated. Made up stats with an attached AI generated screenshot of stripe or some other dashboard showing ridiculous numbers. Many people don’t bother digging deep and just fall for the made up story. Don’t believe everything you see on social media.

Connect Google Business Profile (GBP) to Google Analytics support.google.com/analytics/answ…

Google Analytics planned release: In Admin, under Product links, click Google Business Profile links.

To all the founders/CEO out there: Its NOT AI vs. Humans. It AI and Humans. This hybrid work approach is superior to both pure automation and pure human work.

Yesterday, I caught someone using a fake AI-generated screenshot of GA4 UI to support his so called "study" that GA4 tracking is all wrong. As a data analyst, I don't trust any chart/graph/screenshot posted on social media that demonstrates growth in traffic, sales, leads, customers, etc. It is so easy these days to create convincing AI-generated charts/graphs/screenshots, and yet many people still fall for it.

Check out the latest article in my newsletter: Voice AI This Week - June 4, 2026 linkedin.com/pulse/voice-ai… via @LinkedIn

It is a collection of text prompts (a prompt library) that can be easily used, reused and modified to automatically generate SQL for 'N' use cases. It is one of its kind. Nothing like this even remotely exists in the digital analytics space. >> There are traditional GA4 BigQuery courses that teach writing SQL manually. >> There are scattered prompt tips or single prompt templates people share. >> There are static SQL query collections (GitHub repos, blogs). But nothing combines all of this into a comprehensive, reusable, ever-growing repository of specialised prompts specifically built for GA4 + BigQuery, designed for ongoing use and modification. It stands out because it’s not teaching people to become SQL experts. It’s giving them a permanent, adaptable prompt toolkit they can use with AI indefinitely. Something worth mentioning and remembering.

Check out the latest article in my newsletter: Google has frozen the Measurement Protocol. linkedin.com/pulse/google-h… via @LinkedIn

Check out the latest article in my newsletter: Overflow and After-Hours Voice AI Receptionist. This is it. linkedin.com/pulse/overflow… via @LinkedIn

Unpopular opinion -> Learn to manually perform a task first and do it dozen of times before thinking of automating it. So when anything breaks during automation (and it will), you know exactly what broke and how to fix it. Without manual reps, your automation workflow is unlikely to survive production load.

Check out the latest article in my newsletter: Your automation workflows are guaranteed to fail under production load without this. linkedin.com/pulse/your-aut… via @LinkedIn

If AI mentions truly compensated for lost clicks, we'd see stable or rising leads/sales despite traffic drops. In reality, many (probably most) sites experiencing significant declines in organic traffic also see corresponding drops in leads and sales. The indirect benefits from AI visibility don't offset the volume loss for a large portion of businesses.

If you see your organic search traffic dropped by more than 50% in the last 2 years and at the same time notice a significant drop in leads, rest assured, brand mentions in AI results did nothing for your business bottomline. That's the proof for you, the brand mentions in AI results have no commercial value, despite what vendor-sponsored studies (touting "visibility = value.") say. Eventually market will understand this, and the correction will take place.