# The Trillion Dollar Scam - How Google Really Makes Money

Source: https://www.youtube.com/watch?v=BfNIRyPi5QA
Recap page: https://rapidrecap.app/video/BfNIRyPi5QA
Generated: 2025-10-29T15:04:00.214+00:00

---
## Quick Overview

Google's primary revenue stream of $1.16 trillion annually is derived from advertising campaigns where the company claims credit for sales that would have occurred anyway, a practice proven by showing that running ads in a control market like Ottawa yields only 160 sales, while the target market, Montreal, sees 4,544 sales, and this discrepancy is falsely attributed to the ads, which is further evidenced by Google's CausalImpact tool, which, when applied to mock data, shows that a successful intervention (ad launch) is only successful because the control group's performance collapses, demonstrating that Google often takes credit for organic growth, a digital equivalent of handing out physical flyers.

**Key Points:**
- Google claims credit for $1.16 trillion in annual advertising revenue by falsely attributing organic sales to their ad campaigns.
- The video demonstrates this by comparing a control market (Ottawa, 160 sales) with a test market (Montreal, 4,544 sales) after an ad launch, showing the massive, misleading uplift attributed to ads.
- Google's CausalImpact tool, designed for causal inference, is shown to falsely confirm ad success when applied to scenarios where the control group collapses, mirroring the real-world data manipulation.
- The host mentions that the comparison is like the 'Toothbrush Test,' where a product's success is exaggerated because the control group's performance drops significantly, implying Google suppresses control data.
- The narrator points out that if the ads were truly effective, the sales in the control market (Ottawa) would not have dropped so severely relative to the predicted baseline.
- The video features mountain biking footage in a forest setting, specifically on trails with wooden structures, to provide a dynamic visual backdrop for the data analysis.
- The host concludes that this practice is a massive scam where ad spend is justified by showing inflated results against a suppressed baseline, rather than true incremental gain.

![Screenshot at 01:05: Google founders Sergey Brin and Larry Page are shown holding the Google logo, immediately preceding the explanation of the PageRank algorithm, setting the context for discussing Google's foundational influence on search and, by extension, advertising attribution.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-01-05.png)

**Context:** The video features the host, Ryan F9, mountain biking through dense forest trails, using the activity as a dynamic setting to explain a complex critique of Google's advertising performance metrics. The core concept revolves around how major tech platforms like Google, Meta, and TikTok allegedly inflate the success of their advertising campaigns by manipulating control group data, leading businesses to spend billions based on misleading attribution models that attribute organic sales to paid ads, a practice the host labels a 'trillion dollar scam.'

## Detailed Analysis

The video argues that Google's revenue model is based on a massive scam involving advertising attribution, where platforms take credit for sales that would have happened organically. Ryan F9 uses an analogy involving two cities, Ottawa (control) and Montreal (test market), to illustrate the deception: after an ad campaign launch, Ottawa's sales (predicted baseline) drop drastically to 160 sales, while Montreal's sales surge to 4,544, with Google claiming the difference as ad success. This is visually reinforced by overlaying a graph showing the observed vs. predicted sales, where the ad launch causes the predicted line (representing what would have happened without ads) to plummet in the control market. The host further demonstrates this potential manipulation using Google's own CausalImpact tool, showing how easily the tool can be made to report success when the control data is suppressed or collapses, which is the supposed tactic used against control markets like Ottawa. The host calls this the 'Toothbrush Test,' referencing a scenario where a product's success is exaggerated because the control group's performance drops significantly, concluding that Google's claimed $1.16 trillion in revenue from this method is based on inflated, non-causal attribution.

### Introduction & Setting the Scene

- Ryan F9 introduces the topic while mountain biking on challenging forest trails (0:00-0:15)
- He immediately frames the subject as a 'trillion dollar scam' related to big tech advertising.

### PageRank Mechanics

- Explains the original PageRank algorithm using a simple three-site link structure (A, B, C) where link importance is shared (01:09-01:21)
- Demonstrates that if links are uneven, page rank accumulates unevenly (01:27).

### The Scam Unveiled

- Argues that Google suppresses control group data when measuring ad performance
- Uses the Ottawa (control) vs. Montreal (test) example where ads allegedly cause a massive, false uplift (05:41-06:05).

### CausalImpact Tool Critique

- Shows how Google's CausalImpact tool can be manipulated to show success when the control data collapses, proving it can be used to justify flawed ad attribution (05:07-05:22).

### Real-World Implications

- The host shows e-commerce data where a $750 spend generated 73 purchases, but argues this success is falsely claimed by Google, as organic sales (4,544 in Montreal vs. 160 in Ottawa) would have occurred anyway (02:28-04:50).

### Conclusion

- The entire system is designed to make ad spend look successful, even when it is not, by manipulating the baseline (06:36-07:05).

![Screenshot at 00:03: Mountain biker riding through a dense forest trail, setting the scene for the video's outdoor, high-energy aesthetic.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-00-03.png)
![Screenshot at 00:14: Title card showing the word "FORTNINE" overlayed on a fast-moving mountain bike trail, indicating the context of e-commerce/product review.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-00-14.png)
![Screenshot at 01:05: A graphic showing Google's founders holding the Google logo letters, introducing the topic of Google's influence on search and data.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-01-05.png)
![Screenshot at 01:13: A diagram illustrating three linked websites \(A, B, C\) with arrows showing the flow of PageRank, explaining the concept of link importance.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-01-13.png)
![Screenshot at 01:27: The PageRank calculation \(0.5 for A, 0.25 for B and C\) superimposed over the trail, visually demonstrating the mathematical concept being discussed.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-01-27.png)
![Screenshot at 01:55: A statue of a historical figure wearing modern streetwear \(baseball cap and gold chain\) overlaid on the forest background, used as a visual metaphor for something out of place or deceptive.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-01-55.png)
![Screenshot at 02:25: First-person view of the biker riding down a very long, narrow wooden plank feature in the forest, emphasizing the challenging terrain.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-02-25.png)
![Screenshot at 02:46: A screen capture showing a graphic design application \(likely Photoshop\) being used to create a thumbnail for a video about 'TOP RATED MTB PEDALS!', showing the contrast between the video's content and the ad-making process.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-02-46.png)
![Screenshot at 03:37: The host standing beside his bike, gesturing emphatically while explaining the issue of ad spend vs. new customers, highlighting the video's central critique.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-03-37.png)
![Screenshot at 05:07: A highly motion-blurred, first-person view of the trail with data overlays showing 'AD SPEND' and 'NEW CUSTOMERS' lines diverging sharply, illustrating the claimed data manipulation.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-05-07.png)
![Screenshot at 05:33: A screenshot of the Google CausalImpact GitHub page overlaid on the forest, identifying the specific analytical tool being critiqued.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-05-33.png)
![Screenshot at 06:21: A shot of an older couple looking shocked at a calculator, illustrating the financial impact of the alleged scam on average business owners.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-06-21.png)
![Screenshot at 07:28: The host standing beside his bike, looking frustrated and pointing to his head, reacting to the absurdity of the advertising metrics discussed.](https://ss.rapidrecap.app/screens/BfNIRyPi5QA/00-07-28.png)