Marques Brownlee, the popular tech YouTuber known as @MKBHD, recently faced criticism following the launch of his wallpaper app, Panels, which drew backlash over its pricing model and data privacy concerns. The app, offering curated high-resolution wallpapers through a subscription plan, costs $49.99 per year, sparking disappointment from users who felt the app lacked polish and featured too many ads. Concerns over the app's request for excessive data prompted Brownlee to pledge updates, including reducing ad frequency and addressing privacy issues. Brownlee acknowledged the feedback and committed to improving the app over time. Although much tamer, this scandal fits into a pattern of YouTuber drama from blockbuster creators like MrBeast, KSI, and Logan Paul.
Marques Brownlee is renowned for his in-depth, high-quality reviews of the latest gadgets, smartphones, and technology products. With a background in technology and a highly polished production style, MKBHD's videos offer insightful commentary, clear visuals, and a balanced perspective that has earned him a large and loyal following. He often explores the latest innovations in consumer electronics, conducts interviews with tech industry leaders, and participates in broader discussions about technology trends.
To better understand this story and Brownlee's audience, we collected 7.6 million comments viewers left on Brownlee's videos starting in 2016. We found that Brownlee quickly pivoted by making dramatic changes to the pricing structure of his application. In this brief, we'll walk you through our data pipeline of how we created a custom Infegy Starscape dataset, then walk you through the negative indicators we found and help show you Brownlee's pivot.
Infegy collects millions of social, web, and blog posts across dozens of well-known websites and platforms. We add new platforms as they emerge within the social zeitgeist. However, we don't collect the entire internet.
We do offer a robust solution for those who are looking to analyze data points that are not available from our out-of-the-box dataset - Infegy's custom datasets. We've written about our custom data collection and analysis before but haven't done so at this scale. For this project, we analyzed 7.2 million comments over the last eight years - that's every public comment left on an MKBHD video (for the record, Microsoft Excel tops out at 1,048,576 rows. This means for this project, we analyzed 6.8x Excel’s max capacity). We built a custom data pipeline that took YouTube comments from this specific channel, wrangled them into an Infegy-friendly format, and then piped them via our Infegy API into a custom dataset.
Figure 1: A flowchart of our data processing pipeline. We used the open-source YT-DLP to collect comments and then Python to wrangle them and push them to our servers via our custom dataset API.
First, let's take a high-level overview of MKBHD's comment volume. Remember, this dataset is large - we looked at 7.2 million comments over the last eight years. A dataset of this size would be impossible to analyze or make sense of with traditional analytical tools like Microsoft Excel, Apple's Numbers, or Google Sheets.
Here are some high-level takeaways we found. Despite increasing view counts, the average number of comments on an MKBHD video has dropped 29% over the last eight years. YouTube comments have long been used as a punchline for increased societal negativity. This makes sense - YouTube has been consistently changing its design over the last few years, making comments more challenging to view on mobile, much to the anger of its users.
When analyzing top topics within the comments, Figure 2 divides them into positive and negative ones. Because YouTubers so often run giveaways through the comment section, it's no surprise that Giveaway shows up predominantly within the positive sections. Additionally, we also see “RED”, a popular type of cinematic camera MKBHD uses, show up as a positive topic.
Figure 2: High-level overview of the last eight years of MKBHD Comments (October 2016 through October 2024); Infegy Custom Dataset.
While getting a high-level overview of MKBHD's comment volume is excellent, we wanted to focus precisely on the Panel's backlash. To do that, we applied a date filter to look at videos posted in September 2024. We quickly see "Panels" as a negative keyword for the most important negative topics. The time-based sentiment chart shows that most negative comments dropped on September 23, 2024.
Figure 3: Zoomed-In View of MKBHD Comments (September 10, 2024 through October 11, 2024); Infegy Custom Dataset.
We wanted to understand that negative story better, so we drilled down into the comments on September 23, 2024. Figure 4 does an excellent job of explaining what angered MKBHD's subscribers so much.
Figure 4: Drill Down View of Comments Left on MKBHD's September 23, 2024 Video; Infegy Social Dataset
So, the initial launch of Panels was an unmitigated disaster. MKBHD's team waited to pivot to make the necessary product changes to the app. They clarified aspects of the AI-generated art, offered free and low-priced tiers, and fixed many intrusive ads. After those changes were released on October 11, 2024, we saw a marked difference in the reaction by his subscribers. (Figure 5). We see topics like "thank you," "respect," "accountability," and "transparency" appear. Most interestingly, his subscribers refer to KSI, one of the scandal-plagued YouTubers we mentioned above, showing that MKBHD responded better than KSI.
Figure 5: Word cloud showcasing important topics subscribers wrote after MKBHD's apology (October 11, 2024); Infegy Custom Dataset.
Infegy Starscape's flexibility allows you to dive deep into influencers' reputations, tracking real-time audience reactions and shifts in sentiment with custom datasets like YouTube comments. In the case of MKBHD's app launch, understanding how his audience responded was crucial to navigating the backlash and regaining trust. By creating custom datasets that analyze millions of comments and audience behaviors over time, Infegy enables you to track these vital narratives and understand how influencers can recover from missteps. Knowing your audience through data is critical to staying ahead in today's influencer-driven landscape.