AI helps overseas marketing and increases KV click-through rate by creating localized characters. Analyze local aesthetics, adjust character characteristics and shapes, and integrate brand elements to optimize marketing effects.
In marketing campaigns, visual appeal is key to increasing user click-through rates. In the past, we used strong brands and light textures to show selling points to the global market, and differentiated designs for the aesthetic preferences of users in different markets.
At the same time, we have been thinking about how to continue to improve the attractiveness of KV to users, and whether we can explore some new possibilities. So he set his sights on the 3B principle.
What does a product manager need to do?
In the process of a product from scratch, it is not easy to do a good job in the role of product manager, in addition to the well-known writing requirements, writing requirements, writing requirements, there are many things to do. The product manager is not what you think, but will only ask you for trouble, make a request:
View details >
01. Why can characters enhance KV attractiveness?
In advertising, there is a widely recognized “3B principle”: Beauty, Beast, and Baby. These three elements are more likely to attract users’ attention because people are more likely to be attracted to facial and body movements.
And this attraction is somewhat equivalent to the click-through rate of marketing campaigns. But it is rarely used, and there are many helplessnesses behind it:
- Team limitations: Lack of professional photography team, difficult to independently shoot people that meet the needs.
- Outdated materials: The content of the stock materials is old, and competing products are also used, lacking uniqueness.
- Time-consuming matching: For the global market, finding character models that match the cultural and aesthetic preferences of each region is a large workload and low feasibility.
With the advancement of AI technology, we have ushered in a game-breaking opportunity – using AI generative characters to efficiently create local event KVs.
02. Three moves to win, create localized AI characters
Taking the winter inbound tour to Japan at the Philippine site as an example, two generative human models can be easily obtained through simple descriptors, but they encounter new problems: generative characters have stereotypes, do not have localized character characteristics, are false and untrue, but reduce brand credibility.
How to make AI-generated characters more in line with local preferences?
The following is a study by the University of Rochester, which distinguishes between Chinese, Japanese, and Korean people by fine particles such as facial features, skin color, etc.
Based on its enlightenment, it will conduct an in-depth analysis of local stars and KOLs from three dimensions, and investigate the “appearance passwords” of 12 sites around the world.
01. Facial features: Capture the unique charm of the region
People in different regions have different definitions of beauty. In order to generate AI characters that match the local aesthetic, we analyzed a large number of local non-mixed-race stars and KOLs, because there is a certain difference between national appearance and local preferred advertising models, which can also effectively avoid the offense caused by stereotypes to users.
Taking the Southeast Asian market as an example, although users in Vietnam and the Philippines are both from Southeast Asia, their facial features are obviously different:
Through AI technology, AI characters that match the aesthetics of local users can be generated based on these subtle differences, thereby enhancing the attractiveness of KV
02. Makeup and Hair Styling: Incorporate local aesthetic preferences
Vietnamese look: clear pink base, natural eyelashes and eyebrow shape. Long hair with a low ponytail or shawl slightly curled, short and medium hair shoulder-length buttons, simple accessories. Prefer linen, T-shirts, denim, dresses and Ao Dai modifications, simple and clean.
Filipino look: three-dimensional contouring, heavy eye makeup with colored eyeshadow and thick eyelashes, bright red, orange-brown lip color, highlight the vacation feeling, and avoid being too flashy. Popular curls, high ponytails, and ball heads are common, with large flowers and metal exaggerated accessories. Prefer bright and tight tight tops such as silk and sleeveless bolero with high-waisted pants.
03. Action perspective: convey the connotation of local culture
Vietnamese market: Prefer diverse perspectives (head-up, top-down, up-to-head), rich character movements, often using full-body jumping, palm spread and other movements to convey excitement, sincerity and other emotions.
Philippine market: prefers a head-up perspective, and the character movements are mainly based on hand movements, such as clenching fists, pointing to interest points, etc., and the overall style is relatively mild.
By adjusting the character’s facial features, makeup, hair styling, and action perspective, you can guide users’ attention with a more localized image.
03. The blessing of brand elements makes KV more recognizable
In addition to the appearance and movements of the characters, the embellishment of brand elements is also an important means to enhance the attractiveness of KV.
Business elements: Add travel-related elements such as mobile phones, cameras, suitcases, passports, and sunglasses to convey the business attributes of the event.
Use of brand colors: Incorporate the main color of the brand to strengthen the user’s memory of the brand through color. At the same time, the research materials of each site are sorted out and summarized, and an exclusive material database is established, which provides rich resources and reference for subsequent design work.
Further expand the localized AI generative characters for each site, which not only improves work efficiency, but also realizes the deep integration of brand sense and localization.
Behind this, there is actually a set of practical methodological processes:
sort out research assets in various places, mark key information, and build exclusive data sets; and then use these data to train models to optimize model assets that meet the needs of various regions; After quickly producing a large amount of digital content, after rigorous screening, the high-quality content is archived into the dataset. Through the closed loop of “data accumulation, model training, content production, and result feedback”, the human model asset library is expanded and improved.
04. Clarify application scenarios to maximize the value of AI characters
Celebrities and KOLs have strong fan appeal, topic dissemination and brand endorsement capabilities, which should be applied at critical moments of the brand. Considering the scope of application of celebrity endorsement contracts, generative characters, as one of the visual design methods, can fill the gap in Class B requirements.
To ensure the safety and compliance of using AI-generated characters, in collaboration with our legal and risk control teams, we have developed a strict AI-generated work specification manual, with some core requirements as follows:
Content compliance: prohibit the use of real celebrities and IP images, and eliminate the use of unauthorized pictures for matting to avoid copyright disputes;
Output specifications: All generated images must add a 5% transparency “AIGC generated” watermark in the security area in the lower right corner to clarify the content attributes.
Review mechanism: Strictly review according to the manual specification process before exporting, screen out pictures with distortions (such as multiple limbs, facial distortions, joint deformities) to avoid visual misleading; At the same time, avoid content that touches on racial sensitivities and ensure cultural respect and equality.
Practice has proved that after batch application of AI generative characters in adaptation scenarios, the average CTR has increased by 40%, and the visual merchandising link has been successfully optimized by technical means, achieving a phased breakthrough.
05.AI empower more business scenarios
In the future, we will explore the application of AI generative characters with a broader international perspective, and strive to achieve the deep integration of internationalization and localization.
AI-generated characters are rewriting the design rules of the event KV, using real-time data feedback from different markets to quickly adjust parameters to ensure that it always meets the needs and preferences of local users. In this way, each AI generative character can enhance the attractiveness of KV in the integration of internationalization and localization, help the click-through rate of Banner, thereby releasing greater marketing value, and allowing us to see the infinite possibilities of combining technology and marketing.
