CrushPress.AI

Unveiling ChatGPT’s Brand Bias: An Insightful Analysis

```json { "alt": "Illustration of digital screens with colorful bars and rectangles on a blue background.", "caption": "A vibrant digital interface showcases colorful bars, illustrating modern technology's dynamic and organized complexity.", "description": "This image features a digital illustration with multiple layered screens displaying colorful bars and rectangles. The main screen is centered, featuring blue, green, and pink bars on a clean, structured layout. The blurred effect on the surrounding screens suggests movement and depth. Set against a sleek blue background, the image symbolizes the complex and dynamic nature of technology and data visualization." } ```

Written by

shivamcrushpressai

in

,

I recently embarked on a fascinating exploration of ChatGPT’s brand recommendation patterns, and let me tell you, the findings offer a lot to chew on!

We all know that AI responses are a roll of the dice – ask the same question ten times, and you’re bound to get ten different answers. But I couldn’t help but wonder, just how varied are these responses?

Rand Fishkin’s intriguing research dives into this very question. His findings have significant repercussions for how we approach AI visibility tracking for brands.

Fishkin experimented with prompts ranging from recommendations for chef’s knives to cancer care hospitals, as well as Volvo dealerships in Los Angeles.

```json { "alt": "The CapmatchOne logo with a gradient circle and bold text.", "caption": "Discover innovation with the CapmatchOne logo, featuring sleek typography and a modern gradient circle.", "description": "The CapmatchOne logo features bold, modern typography coupled with a gradient circle, symbolizing connection and innovation. The sleek design conveys a sense of progress and creativity. This image can be used for branding or promotional purposes, appealing to audiences interested in innovative solutions and forward-thinking designs." } ```

His results showed that AI systems like ChatGPT almost never recommend the same set of brands in the same order twice.

Moreover, when asking about something specific like running shoes, certain brands tend to appear more frequently than others.

Building on this research, I zeroed in on B2B scenarios, adding some of my own twists: does the complexity of the prompt or the competitiveness of the category make a difference to AI’s consistency?

```json { "alt": "Bar chart showing average unique brands ChatGPT uses across different prompt types.", "caption": "Discover how ChatGPT sources brands with varying prompt complexities and categories. Competitive prompts yield the highest diversity, while niche prompts pull the fewest.", "description": "This bar chart illustrates the average number of unique brands ChatGPT identifies in response to different prompt types: simple, nuanced, competitive categories, and niche categories. On average, competitive category prompts result in the highest diversity with 57.8 brands, while niche category prompts have the least at 30. The chart provides insights for understanding brand diversity in AI responses, useful for optimizing prompt design." } ```

To investigate, I crafted twelve varied prompts, half of which addressed highly competitive B2B software categories, like accounting, and the rest focused on niche categories, such as user entity behavior analytics (UEBA) software.

Further, I examined simple prompts against nuanced ones that included specific personas and use cases.

Each prompt was fed into ChatGPT 100 times using different IP addresses to mimic 1,200 unique users.

```json { "alt": "Bar chart showing average brand mentions per response across different prompt types.", "caption": "Discover how different prompt complexities affect brand mentions per response. From simple to niche, see the variations unfold.", "description": "This bar chart visualizes the average number of brands mentioned per response across various prompt types. 'Simple prompts' lead with 11.7 mentions, while 'nuanced prompts' have 9.2. 'Prompts in competitive categories' show 11.1, and 'prompts in niche categories' record 9.8. Each category includes six prompts, with data reflecting 100 responses per prompt, providing insights into how prompt complexity and category influence brand mention frequency." } ```

Now onto the juicy part: the findings.

Submitting a single prompt to ChatGPT 100 times revealed that, on average, 44 different brands got mentioned. However, some response sets listed as many as 95 brands, heavily dependent on the category.

Notably, competitive categories yield twice as many brand mentions per 100 responses compared to niche ones.

```json { "alt": "Bar chart showing brand visibility distribution. Five dominant brands have high visibility, followed by 10 middle brands and 29 long tail brands.", "caption": "Discover which brands stand out! A visual breakdown of 44 brands shows how five dominate in visibility, with others trailing behind. Ideal for understanding brand awareness trends.", "description": "This bar chart illustrates the visibility percentages of 44 brands as recognized by ChatGPT. It categorizes them into dominant (5 brands), middle (10 brands), and long tail (29 brands) based on visibility levels. The dominant brands have significantly higher visibility, making up 11% of the total, while middle brands account for 23%, and long tail brands form 66%. This analysis is derived from average visibility across 100 responses and 12 prompts, useful for gauging brand prominence." } ```

Simple vs. nuanced prompts? ChatGPT typically mentions fewer brands in response to nuanced requests, but this isn’t a hard and fast rule.

When diving deeper into ChatGPT’s brand consistency, I found that in a set of 100 B2B software recommendations, only about five brands (11% of the total) were mentioned 80% or more of the time.

Dominant brands in a category like accounting software were names we all recognize: QuickBooks, Xero, Wave, and the like.

```json { "alt": "Bar graphs showing AI brand visibility in competitive vs. niche categories.", "caption": "Unlock niche success! Discover how AI visibility differs in competitive vs. niche categories with insightful bar graphs.", "description": "This image contains two bar graphs comparing AI brand visibility in competitive and niche categories. The competitive category, such as accounting software, includes approximately 58 brands, with dominant, middle, and long tail segments. The niche category, such as reverse ETL software, averages 30 brands, showcasing a variance in brand visibility distribution with distinct dominant, middle, and long-tail sections. Ideal for understanding AI market positioning, this infographic highlights the ease of achieving visibility in niche markets." } ```

If you’re not among the big guns, working within a niche offers a strategic advantage given the increased chance to be consistently recognized by AI.

For marketers, this study underscores the necessity of standing out and perhaps carving a niche if dominance in a broad category seems out of reach.

Moreover, most AI visibility tools might not give you the full picture if they’re conducting only a single spot-check. For more reliable data, multiple runs per prompt are essential.

```json { "alt": "Chart comparing brand visibility for simple and nuanced prompts, showing dominant, middle, and long tail visibility percentages.", "caption": "Exploring brand visibility: Simple prompts showcase clear leaders, while nuanced prompts level the playing field, highlighting the challenges of capturing dominant positions.", "description": "This image features a comparative bar chart illustrating brand visibility for simple versus nuanced prompts. For simple prompts, out of 100 responses, around 46 brands participate, with 14% being dominant, 20% in the middle, and 66% in the long tail. For nuanced prompts, approximately 42 brands return from 100 responses, with 10% dominant, 23% in the middle, and 67% in the long tail. This visualization emphasizes the difficulty brands face in maintaining dominance with increasing prompt complexity. Keywords: brand visibility, simple vs. nuanced prompts, dominant brands, marketing analysis." } ```

So, if you’re tracking pivotal prompts, run each a handful of times to get a better sense of your brand’s visibility.

I’m excited to share that future reports will explore ChatGPT’s understanding of brands and whether consistent recommendations reflect deeper brand awareness.

This article was originally published on Visible and republished with permission.

Inspired by this post on Search Engine Land.

crushpress.ai community screenshot

FAQs

How consistent are ChatGPT’s brand recommendations?

They are inconsistent; ChatGPT almost never recommends the same set of brands in the same order twice.

What factors influence brand mentions in ChatGPT responses?

Prompt complexity and category influence brand mentions. Competitive prompts yield more brand mentions per 100 responses; nuanced prompts reduce dominance.

How was the study conducted?

The study used twelve varied prompts; each was run 100 times with different IP addresses to simulate 1,200 users.

Which brands dominated accounting software?

QuickBooks, Xero, Wave were named as dominant accounting software brands.

What practical advice does the article offer for tracking AI visibility?

Run each prompt multiple times to obtain more reliable data; single spot-checks can be misleading.

How many brands were identified and how were they distributed?

44 brands were identified; five were dominant, ten middle, and 29 long-tail.

Comments

Leave a Reply

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

More posts