In the vast digital landscape of social media, where every post competes for attention, have you ever wondered why some shine brighter than others?
It's easy to get caught up in the numbers game, focusing solely on metadata like hashtags, engagement rates, and audience size. Yet, there's a pivotal factor often overlooked amidst the metrics frenzy: the content itself. 🔥
At Exolyt, we're challenging the status quo with an innovative approach to social media analysis. Our mission? To revolutionize how we understand and leverage content's role in online success and learn from it.
To do so, Exolyt devised a cutting-edge AI solution to redefine how we perceive social media content, particularly on TikTok. Harnessing the power of advanced computer vision algorithms, our video analysis dives deep into the heart of TikTok videos, unlocking insights that transcend conventional analytics.
How we do it:
- Sentiment Analysis: Emotions drive engagement. With sentiment analysis, we track facial and vocal emotions expressed to gauge the tone of TikTok videos, unveiling the underlying feelings that captivate viewers and spark conversations.
- Text Extraction: Words carry weight, even in a visually driven platform like TikTok. Our tool extracts text from ‘within’ the videos, uncovering words used visually in the videos that are often overlooked during analysis. These words may contain captions and trending phrases that resonate with audiences.
- Demographic Insights: Understanding your audience is key to crafting compelling content. Through demographic analysis, we unveil the age, gender, and geographic distribution of TikTok content creators, empowering creators and brands to tailor their message effectively.
- Colour and Audio Analysis: Visual and auditory elements shape the viewer experience. By dissecting colour schemes and analyzing audio patterns, our AI discerns the aesthetic and sonic signatures that elevate TikTok videos from ordinary to extraordinary.

Video AI Analysis Examples (Disclaimer: This is not the actual screenshot of the tool)
Our video analysis tool can accurately identify objects, extract texts, analyze sentiments and demographic information, and conduct colour and audio analysis within every second of TikTok video.
This holistic approach transcends traditional social media analytics, bridging the gap between content specifics and performance metrics. It's not just about how many likes or shares a post receives; it's about understanding why!
We launched a few pilot projects with our customers to initiate this project. Here are some of the findings:
来自领先企业的试点案例
a. A Global Creative Agency
As a socially led agency that studies social behaviours within online communities, cultures, and subcultures, our client wanted to reach the right people in a strategic, relevant, and effective way. To do this, our client wanted to understand a specific topic's trends, demographics, and content patterns and decode them utilizing Exolyt’s Video AI Analysis.
To do so, they identified some key brands from the luxury industry on which they were looking to gain more insights into how people wear its jewellery while also gaining insight into the types of captions and comments seen in that particular category's content.
The Project Brief included some details like:
- Key brands: Graff, Bulgari, and Cartier, among others.
- Time window: 6 to 12 months
- Related hashtags: #luxuryjewellery
Sample Findings from This Pilot:
- Main Hashtag distribution during the given timeframe

Source: Exolyt Video AI Analysis
- Demographic Distribution by Main Hashtag

Source: Exolyt Video AI Analysis
- Most frequent age groups by Hashtag

Source: Exolyt Video AI Analysis
- Facial sentiments overtime for Cartier

Source: Exolyt Video AI Analysis
Note: Because of the aggregation, there are more negative sentiments. Negative sentiments include more emotion classes (anger, disgust, fear), while positives include only "happy" emotion classes.
- Most used words in this category

Source: Exolyt Video AI Analysis
Some Key Takeaways from this Pilot Case:
- Most of the content within these topics are produced by women aged 20-29, followed by women aged 30-39.
- Almost half of the content published under these hashtags does not feature creator faces.
- Cartier is the brand with the largest content across all age groups.
- Vancleef mentions shrink as the age of the creator in the video increases.
- The content produced by female users features more blanched brand distribution, with Cartier being the most common, followed by Vancleef.
- Over time, the number of male creators in the space has risen, while content featuring no creator's face has decreased.
- Facial sentiments indicate a sharp peak of predominantly negative emotions in creators around May 2023, followed by a peach of positive emotions in videos published towards September 2023.
b. McCann Paris
Our customer, an entity of McCann Worldgroup and a leader in the advertising sector that develops integrated marketing and communication solutions for clients in France and worldwide, wanted to study the public opinion on TikTok for one of their biggest clients from the beauty industry.
The main objective was to understand the demographic makeup of the user-generated content across TikTok for the brand, along with the colour and sound analysis of the content, to determine which skincare products are popular, which skin tonality commonly constitutes their usage, and what the users have to say about the products.
The Project Brief included some details like:
- Key brands: Large Beauty/Makeup Brand
- Time window: 3 months
- Related Hashtags: (brand name)
- Key markets: GB
Sample Findings from This Pilot:
- Age group distribution for the branded hashtags

Source: Exolyt Video AI Analysis
- Voice sentiments overtime

Source: Exolyt Video AI Analysis
- Facial Colour Analysis

Source: Exolyt Video AI Analysis
- Video Colour Analysis

Source: Exolyt Video AI Analysis
Note: This is a limited data sample based on videos that used the hashtag Loreal. The patterns found in this sample might not be generalizable to all TikTok videos across industries, markets, and regions.
Some Key Takeaways from this Pilot Case:
- Gender and age group distribution for the branded topics/hashtags showcased: Most of the content for the branded topics is produced by women aged 20-29, and creators aged 20-29 prefer tagging their videos with the #brand hashtag, while creators between 30 and 39 use the #brand hashtag more frequently.
- The creators' facial and voice sentiments maintain a stray pattern throughout the months, with a mostly positive and negative mix and very little neutral sentiment.
- The diversity of the creator's facial colours illustrates the prevalence of lighter and warmer colours, which indicate particular skincare trends.
- The videos on this topic predominantly have neutral background colours. Blacks and greys seem common, with orange-colored lighter backgrounds appearing in some videos.
- Video background colour does not substantially affect video engagement in the sample we examined in this study.
为什么选择AI视频分析?
Social intelligence consumers across various industries can find multiple applications of this data from TikTok.
其价值不仅仅局限于提升品牌曝光、互动、影响者营销或竞品分析;它同样是定性研究的高价值资源,并可作为数字人种志的有效延伸。
随着整整一代人的社交媒体沟通集中到同一平台,TikTok 为品牌提供了庞大且尚未被充分开发的市场研究机会,帮助深入理解他们的观点、体验、不断演变的文化,以及他们真正关注的议题。
增强型视频分析能力使您得以从此类人群创作的内容中汲取洞察,助力您开展更高质量的定性研究,并全面理解不同产品与服务的消费者旅程。
通过对 TikTok 视频逐帧解析,在视频的每一秒内精准进行目标识别、文本提取、情感与人群属性分析,以及色彩与音频分析,帮助解构娱乐与商业交织的 TikTok 趋势,洞察其对文化、生活方式与购买模式的影响。
它也解答了一个关键问题:为什么有些社交媒体内容更能脱颖而出?答案在于每条内容中,视觉呈现、文案表达、情绪触发与受众共鸣的精密协同。
在 Exolyt,我们致力于拆解这一复杂生态,赋能营销、分析与研究团队。欢迎与我们一同探索,让 AI 的力量与内容叙事相融合。让我们携手,以每一条更具吸引力的内容为起点,重新定义社交媒体成功的标准。
试点项目详情
谁可以参与此试点项目?
希望深入理解 TikTok 视频的视觉内容,而非只看常规指标,并有此类分析的明确应用场景的企业。
该项目包含哪些内容?
- 行业领先的 AI 内容分析模型,可逐秒分析 TikTok 视频内容。
- Exolyt 提供 TikTok 视频免费定制分析(目前处于 Beta 阶段,参与完全免费)。
如何参与该项目?
- 如果您是客户,参与流程非常简便。您可以直接与我们对接并分享您的需求;或在不影响您照常使用平台的情况下,与我们的数据科学家共同制定方案。
- 即使您不是客户,也可参与,平台按免费增值版限额使用,但可完整享有 Video AI 分析项目的全部权益。
免责声明:该功能尚未在平台上线,您无需登录 Exolyt 也可使用本项目成果。
我们将根据您的视频分析需求量身定制项目方案,并由我们主导实施,与贵公司指定联系人密切协作推进。
您的参与有望塑造 TikTok 内容分析的未来。若您希望进一步了解这项技术,并共同推进尖端 AI 工具的研发,我们诚邀您加入。

