CrushPress.AI

ChatGPT Relies Heavily on Google Shopping for Carousel Products

```json { "alt": "Infographic showing ChatGPT sourcing 83% of carousel products from Google Shopping.", "caption": "Discover how ChatGPT sources most of its carousel products from Google Shopping, illustrating the impact of shopping query fan-outs.", "description": "This infographic highlights a new finding that ChatGPT sources 83% of its carousel products from Google Shopping via shopping query fan-outs. The image includes visuals of product icons, directional arrows illustrating query flow, and the Google Shopping logo. Ideal for exploring the integration of AI and online shopping platforms. Keywords: ChatGPT, Google Shopping, shopping query, carousel products, infographic." } ```

Written by

shivamcrushpressai

in

, ,

I recently stumbled upon an intriguing revelation: ChatGPT sources a staggering 83% of its carousel products from Google Shopping via shopping query fan-outs. This prompted an investigation into how ChatGPT utilizes shopping query fan-outs and what implications arise from this dependency.

In November 2025, while delving into the depths of AI research, some colleagues and I unearthed an enigmatic piece of code within ChatGPT. The field called id_to_token_map, encoded in base64, ultimately revealed parameters linked to Google Shopping, such as productid and offerid.

```json { "alt": "Top three smartphones under $500: Google Pixel 9a, Samsung Galaxy A36 5G, Motorola Moto G Stylus 2025.", "caption": "Explore budget-friendly smartphones: Discover the Google Pixel 9a, Samsung Galaxy A36 5G, and Motorola Moto G Stylus 2025, all under $500!", "description": "This image showcases three highly recommended smartphones available for under $500: the Google Pixel 9a priced at $499.00, the Samsung Galaxy A36 5G at $399.99, and the Motorola Moto G Stylus 2025 costing $349.99. These models offer a balance of performance, camera quality, and battery life. Ideal for budget-conscious consumers seeking high value, each phone is prominently displayed with a sleek, modern design. Keywords: Google Pixel 9a, Samsung Galaxy A36 5G, Motorola Moto G Stylus 2025, budget smartphones, under $500." } ```

To validate that this field pointed to Google Shopping, we attempted to reconstruct a shopping URL solely from these decoded parameters. Here’s an example from a ChatGPT carousel showcasing “best smartphones under $500,” showing how this process could replicate Google’s shopping links.

```json { "alt": "Google Pixel 9a 128GB with various buying options displayed, including Best Buy and Verizon.", "caption": "Discover the Google Pixel 9a 128GB, blending innovative features with sleek design, and explore competitive pricing from retailers like Best Buy and Verizon.", "description": "This image showcases the Google Pixel 9a with a black and blue abstract wallpaper. The product page highlights a 4.6-star rating from 2.9K user reviews. Buying options are presented on the right, with prices ranging from $300 to $634. Retailers include Best Buy and Verizon, offering installment plans. Key features include a best-in-class camera, durable design, and long battery life, all delivered under $500. Perfect for enhancing productivity and creativity." } ```

The question was whether this shopping link corresponded exactly to products shown in ChatGPT. As it turns out, it did! Yet, it raised more questions about the nature of ChatGPT’s sourcing process. Does this apply across various product categories? Does ChatGPT prefer higher-ranked Google Shopping products?

```json { "alt": "Bar chart of average QFO word count by calendar week from 2025-W44 to 2026-W04, showing normal and shopping fanout data.", "caption": "Explore the trends in average QFO word count per week from late 2025 to early 2026, highlighting normal versus shopping fanout.", "description": "This bar chart illustrates the average QFO word count by calendar week, covering the period from week 44 of 2025 to week 4 of 2026. It compares two data types: normal fanout and shopping fanout. Each category is represented in a distinct pattern, with normal fanout in a lighter shade and shopping fanout in a darker shade. Notable trends in word count variations are visible across the weeks. Keywords: QFO, word count, fanout, bar chart, weekly data." } ```

To deeply explore these queries, we investigated over 40,000 carousel products and analyzed the results. By examining the similarity between ChatGPT carousels and Google and Bing organic products, the study shed new light on ChatGPT’s reliance on Google Shopping for sourcing.

```json { "alt": "Bar chart showing average fan-outs per prompt for Normal at 2.4 and Shopping at 1.16.", "caption": "Comparing fan-out averages: Normal prompts lead with 2.4, while Shopping trails at 1.16.", "description": "This image displays a bar chart that compares average fan-outs per prompt between two categories: Normal and Shopping. The Normal category has a fan-out average of 2.4, represented by a taller bar, and the Shopping category has an average of 1.16, shown by a shorter bar. The chart uses distinct colors for each category, with Normal in green and Shopping in orange. This visual data, sourced from Search Engine Land, highlights differences in engagement or response levels across these categories, making it useful for digital marketing analysis." } ```

Diving into our findings, we see a stark difference between normal search and shopping query fan-outs. Notably, shopping fan-outs are typically shorter, aiming to fetch specific items rather than broader contextual information. This suggests ChatGPT optimizes these fan-outs specifically to compile its product carousels.

```json { "alt": "Bar chart comparing ChatGPT carousel product matches in Google Shopping top 40 between Bing and Google across various match strengths.", "caption": "Exploring the match strength of ChatGPT's carousel products in Google Shopping's top 40, this chart highlights differences between Google and Bing.", "description": "This bar chart displays the match strength of ChatGPT's carousel products, comparing their presence in Google Shopping's top 40 results between Google and Bing. Categories range from 'Exact match' to 'Very weak,' with varying percentages, such as 45.80% for exact matches in Google and 62.56% for very weak matches in Bing. A total of 43,000 products were analyzed. Keywords: ChatGPT, Google Shopping, Bing, product match." } ```

Further, the data indicates most ChatGPT carousels mirror Google’s organic shopping results. Almost 84% of similar products matched within Google’s top 20 positions, reinforcing a clear preference for Google’s top-performers.

```json { "alt": "Bar chart comparing Google and Bing's product match percentages with ChatGPT; Google at 83.24% and Bing at 10.77%.", "caption": "Google far surpasses Bing with a remarkable 83.24% product match rate with ChatGPT, highlighting a significant difference in effectiveness.", "description": "This image features a bar chart from Search Engine Land showing the percentage of strong product matches (.8+) with ChatGPT. Google achieves an impressive 83.24% match rate, while Bing is considerably lower at 10.77%. The chart uses contrasting colors to differentiate Google's and Bing's performance, illustrating the superior match capability of Google with ChatGPT." } ```

Interestingly, ChatGPT’s sourcing from Bing was minimal, with a mere 0.16% exclusive matches, indicating a predominant preference for Google’s data. This stark contrast highlights ChatGPT’s systemic approach to product sourcing.

```json { "alt": "Bar chart showing 73.81% of good product matches found by Google but not Bing, and 0.16% by Bing but not Google according to ChatGPT.", "caption": "Google's prowess shines in this analysis, finding 73.81% of good product matches that Bing missed, while Bing only helped with 0.16%.", "description": "A bar chart displays data on the overlap of good product matches with ChatGPT. It shows that 73.81% of matches were found by Google but not by Bing, while a mere 0.16% were found by Bing but not by Google. The analysis is sourced from Search Engine Land, highlighting significant disparity in search engine effectiveness between Google and Bing in this particular study. Keywords: Google, Bing, product matches, ChatGPT, search engine comparison." } ```

These findings are crucial for brands aiming to feature in ChatGPT’s carousels. Monitoring your Google Shopping rank is integral, yet understanding additional contextual factors—like product sentiment—could enhance visibility.

```json { "alt": "Line graph showing Google Shopping position match by ChatGPT carousel position with mean and median lines.", "caption": "Analyzing the alignment of ChatGPT carousel positions with Google Shopping results, this graph reveals trends in mean and median matches over seven positions.", "description": "This image features a line graph comparing Google Shopping product positions with ChatGPT carousel positions. The x-axis represents ChatGPT carousel positions from 1 to 7, while the y-axis details Google Shopping product positions, ranging from 0 to 15. Two lines indicate the mean and median values, showcasing a rising pattern. The graph is credited to Search Engine Land." } ```

For the field of AI, this study underscores that ChatGPT employs a distinct, independent pipeline for its product carousel, separate from the standard search query fan-outs. Future changes in ChatGPT’s methods remain a possibility, but for now, a systematic reliance on Google Shopping has been firmly established.

```json { "alt": "Bar chart showing cumulative ChatGPT match percentage versus Google Shopping rank from Top 5 to Top 40.", "caption": "Analyzing AI and e-commerce: This chart illustrates how ChatGPT’s cumulative match percentage aligns with the Google Shopping ranking from Top 5 to Top 40.", "description": "This bar chart compares the cumulative match percentage of ChatGPT to the Google Shopping rank, ranging from Top 5 to Top 40. Each bar represents a different Top range, with increasing cumulative percentages as the range expands. The visual highlights the alignment between AI recommendations and e-commerce rankings. Presented by Search Engine Land, it provides valuable insights into AI's performance in product matching." } ```

Inspired by this post on Search Engine Land.

crushpress.ai community screenshot

FAQs

What percentage of ChatGPT's carousel products come from Google Shopping?

The article states that ChatGPT sources 83% of its carousel products from Google Shopping via shopping query fan-outs. This highlights a strong reliance on Google Shopping for carousel sourcing.

What term does the article use to describe how ChatGPT retrieves shopping data?

The piece refers to ‘shopping query fan-outs’ as the mechanism for sourcing shopping data. It explains how this approach helps compile product carousels.

How many carousel products were analyzed in the study?

The study analyzed over 40,000 carousel products. This large sample underpins the observed sourcing pattern toward Google Shopping.

What share of good product matches did Google find that Bing did not?

Google found 73.81% of good product matches that Bing missed. This shows Google’s edge in matching products ChatGPT references.

What were the reported match rates for Google and Bing in the study?

Google achieved an 83.24% match rate, while Bing reached 10.77%. The results illustrate a significant performance gap between the two search engines.

How significant were Bing's exclusive matches according to the study?

Bing accounted for only 0.16% exclusive matches, indicating a strong preference for Google Shopping.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

More posts