Author Orange Paint
Editor Fifi
AI cross-border, is the biggest wind mouth of this year’s e-commerce. Pai Dai learned that many cross-border sellers with the help of AI technology, to achieve efficient docking with global customers, not only significantly improve operational efficiency, but also significantly cut costs.
At the same time, the cross-border e-commerce field continues to usher in new development opportunities. For example, the TikTok Mexico station opened on March 4, local time, and Amazon also allows users to search for products and jump directly to the brand’s independent station to complete the purchase. These new developments all mean that a new blue ocean market is forming.
On March 7, Pai Dai held the Pai Dai cross-border summit in Shenzhen. The summit invited three heavyweight guests to discuss in depth the methodology of selecting explosive products TikTok Mexico station’s new opportunities AI Amazon’s practical play, as well as how to utilize the AI RPA tool in the U.S. region of TikTok low-cost and rapid start-up and other play.
Do Amazon, do not use AI probability to be eliminated
Peter, the founder of Zero Billion Overseas, has 13 years of e-commerce experience, relying on cross-border Amazon independent station Tiktok, with an annual GMV of $100 million.
Peter chose to do Amazon in 2019. He wanted to find an opportunity for long-term precipitation and accumulation, and by observing a number of domestic e-commerce platforms, including Tmall Jingdong Jitterbug, etc., but thought that there were certain limitations in the business models of these platforms, and they lacked the opportunity for long-term precipitation and accumulation. And Amazon from the earliest traffic dividend, through the operational dividend, to 2019 is in the beginning of the shift from the operational dividend to the product dividend, and superimposed on the brand dividend.
I think there should be huge opportunities in the future, such a platform is more suitable for old sellers like me, we have done business for so many years, accumulated a certain amount of money, have a deeper knowledge of the core logic of the business, and take the initiative to pursue a long-term product business. Peter said.
However, Peter also pointed out that the current Amazon sellers are facing three major problems, respectively, cost out of control efficiency bottlenecks and compliance mines. In terms of cost loss of control, Amazon sellers are currently facing many cost problems, advertising rate of more than 10 risk, including logistics and warehousing, return and exchange of goods, risk of compensation, such as encountering infringement of TRO and so on.
Selection of goods accounting often do not take into account these risks, resulting in gross margins of 30 40, while advertising rates of more than 10 that is a loss. Traditional pop-up development strategy costs an average of more than 20,000 U.S. dollars to promote a link at least 20,000 U.S. dollars, and the success rate is getting lower and lower, many sellers last year, the success rate of the development of new links is low, the old links to support the company’s operations, the boss anxious about future income.
In terms of efficiency bottlenecks, from product selection to procurement, shipping, shelving, optimization and advertising takes 4 6 months, and may not be able to do it in the end, not only investing money, but also consuming a lot of time costs and opportunity costs.
In terms of compliance mines, cross-border e-commerce compliance problems occur frequently, such as Shenzhen sellers face huge compensation for lithium battery claims children’s toy safety and other issues. The United States has many relevant laws, such as the Privacy Act, which specifically provides for the use of button batteries, Chinese sellers often do not understand. Patent issues are also prominent, TRO frequently, once the seller encountered infringement was frozen funds, do not dare to continue to sell, and difficult to fight the lawsuit, often choose to settle, pay a high cost, about 30 50 novice sellers affected by this.
The above problems can be solved by AI.Peter will compare the traditional manual mode with the AI intelligent mode.The AI intelligent bottom layer is a big model, which can understand the human intention and disassemble the task, and complete the task through the tool chain behind it, such as decompression program OCR program.
This is similar to the idea of doing Amazon itself, in the Amazon business, the three most core skills such as selection transformation operation and promotion can be completed by AI, and can also solve the problem of wind control, reduce the risk of infringement, and avoid violating the rules of Amazon shelves. There are three core strategies, one is, the dynamic thesaurus engine to capture Best Seller ABA advertising word bank to feed the content generation two is, intelligent compliance review patent trademarks sensitive words real-time scanning to avoid account risks three is, conversion enhancement system A B test product pain point matrix scenario-based graphic generation.
Peter mentioned that there is a three-step screening rule in intelligent product selection, market demand detection, supply chain capability mapping and profitability verification. Using 28 large dimensions of data to predict the single rate, predict which products are likely to become explosive, that is, the premium rate is greater than 30, the probability of infringement is less than 0 5, and the degree of logistical suitability is greater than 85.
Intelligent product selection Closed-loop model testing
In terms of low-cost traffic amplification strategy, the advertising effect can be improved by creating a precise advertising matrix. Among them, Brand Campaign can automatically grab the core big words ACoS 15, DSP Campaign can analyze the path of the user session to automatically follow the investment also includes long-tail word harvesting program, reverse generation of 50,000 long-tailed word database and automatic placement of ACoS 10.
Peter pointed out that the intelligent advertising system can not only realize the efficiency of ad spending increased by 400, and the natural traffic feedback accounted for 60%.
In addition, Peter also shared the experience of a 0 e-commerce experience team how to achieve monthly sales of 87,000 U.S. dollars in a single store. The key action chain has three steps: cold start period, explosive period and stabilization period.
Cold start period The first month, through the system to evaluate the reformulation of the shelves, AI automatic link generation Dynamic pricing, test the product order rate and the possibility of explosion, robot advertising system automatically learn advertising strategy. Explosive period By month 2 or 4, 40 ACoS 12 products were targeted, and supply chain in-depth development was initiated. Stabilization period By the 5th month to the present, realize the best moving rate through 7 24 hours automatic price adjustment, and then repurchase the results done by CM system to do long-term accumulation.
It should be noted that the above operation should include three core insights, one is the cold start golden period to focus on the monthly revenue . One of the monthly revenue formula for the monthly revenue on the shelf volume success rate customer unit price gross margin dynamic operating costs. AI can improve the success rate to more than 20, the cost will be compressed to the traditional mode of 8, to achieve cost reduction and efficiency.
When a company using AI to do e-commerce, their costs can do our 8, its success rate may be three times that of ours, the risk may be our 1 4, or even 1 10, we are actually not competitive in multiple dimensions, we either embrace AI, or be forced to go. Peter said.
The second is to pay attention to the margin of safety of capital . You can calculate the critical value through the formula to determine the operating conditions, such as the critical value is less than 1 2, indicating that the product stagnation, the rate of sales decreased, to immediately optimize the inventory structure 1 5 2 0 for the health of the operating range to reach 3 0, the cash flow profit in all aspects of the very good, with the conditions of scale expansion.
Risk isolation model
Third, pay attention to the path of asset precipitation, to have the precipitation of core assets . Doing Amazon can precipitate store assets, not only making money this year, but also sustaining income in the future. Doing Amazon has three layers of value-added models, in the data asset layer, to build a user behavior library, resulting in a selection model In the traffic asset layer, to build search weight and store reputation, precipitate evaluation and old links, many more than ten years old links are still in the head to make money, and the selection of products should take into account the long-term nature of the asset In the supply chain asset layer, the preferred supplier whitelist.
Three-layer value-added model
Peter suggests that sellers create a long-term evolutionary path, specifically from the tool chain construction, human efficiency revolution and organizational transformation of these key aspects. In terms of tool chain construction, integrate ChatGPT Deepseek Midjourney SD and Vison supervised learning model to build a strong technical support system.
In addition, Peter also pointed out that there are three definitive trends that will happen in the main body of the industry in 2025, one is the civilianization of tools The cost of using the AI selection system has dropped67 The second is the unmanned operation 90 The basic operation has been automated The third is the securitization of assets High-quality stores can apply for factoring financing.
To this end, Peter suggests that sellers should carry out knowledge density enhancement and continuous learning, and the company should pay attention to knowledge precipitation and establish a knowledge base Abstract models, etc. The speed of system iteration in the AI era far exceeds the speed of personal knowledge accumulation, and it is necessary to be adept at using a variety of systems to enhance efficiency Avoid consuming a large amount of resources to do impractical things such as research and development of large models, etc., and rationally utilize the resources according to their own situation.
The essence of explosive products is
Platform Algorithm The product of human weakness
In 2021, TikTok’s monthly activity exceeded the 1 billion mark for the first time, successfully joining the ranks of 1 billion user platforms, while its U.S. user base also exceeded 100 million. By 2022, TikTok monthly activity grew from 1.2 billion to 1.5 billion, during which TikTok shop was born and started the commercialization process, and successfully took the market of 6 countries in Southeast Asia.In 2023, TikTok monthly activity continued to grow, climbing from 1.5 billion to 1.8 billion. Currently, TikTok monthly active users in more than 1.8 billion, covering more than 150 countries.
As the first batch of TikTok sellers, Jiang Weibing operates a number of categories such as jewelry 3C accessories health care products, and all of them have achieved the head of the category, with annual sales of more than 100 million. He began to do TikTok in 2022, and in 2023, he made a full-scale effort, including the opening of the U.S. local store in September 2023, and invested a lot of resources to carry out the U.S. business, first opened about 30 fully managed stores, and then closed to focus on the local store.
When talking about the reasons for choosing the TikTok platform, Jiang Weibing said that compared with mature platforms such as Amazon, TikTok has a huge incremental development speed and is still growing. The countries that open TikTok shop are still very few, less than 20 countries, while TikTok software has covered 150 countries, the theoretical value of the above 130 countries, can go to open TikTok shop. this is a huge opportunity.
So what are the best TikTok products to sell at the moment?
Based on the TikTok platform young group female-oriented user attributes, beauty in all countries basically ranked first, clothing is also the head of the category, the ranking of other categories vary from country to country. Jiang Weibing suggests that you should first choose a big category and then go for the vertical products inside.
Jiang Weibing believes that there are two major elements to do a good job of TikTok, one is the source supply chain, and the other is short video material. For new entrants or relatively mature sellers, having a source supply chain and short video material production of these two elements is a must. If you have both, it is the best If only one of them, you can also carry out business and achieve some results. However, if there is no supply chain, and the team lacks the understanding and ability to produce short video footage, and there is no dedicated team for filming, editing and compiling, then it is pointless to enter the relevant field to participate in the competition, and inevitably will not be able to achieve results.
In addition, Jiang Weibing gave three major TikTok selection models. First, the factory mode selection, through market research data to find out the potential of the factory’s existing products of the explosion, suitable for a large number of SKU inventory and factory with the degree of sellers, the full hosting store in this mode is more common.
The second is to follow the sale mode selection, in the TikTok platform to see the products sold well, and their own supply chain resources related to high profit margins can choose to follow the sale. At present, the TikTok platform product richness is not high, the user is mainly automated shopping, follow the sale is relatively easy, but pay attention to the relatively high price per unit is better.
Third, micro-innovation mode selection, through the integration of supply chain mode for micro-innovation to create explosive products. Integration of the supply chain, not tied to a single factory, through the model of a b c, the combination of products from different factories into a new brand, you can avoid homogenized competition to improve customer orders, and train the team’s ability to combine products.
Jiang Weibing also shared his own product selection tips, that is, refusing to get high, respecting the data, controlling the source, and avoiding full support. Jiang Weibing reminds us to pay attention to six key points in the selection standard of TikTok platform.
First of all, visual impact , as a short video platform, the product short video should have visual impact.
Secondly, to price impact, the platform gold price range at 10 30 dollars, try not to break through 30 dollars, do high unit price products need to be cautious.
Furthermore, light small pieces, light small products in the logistics is more advantageous, and can reduce the user order decision-making costs.
In addition, micro-innovation, through micro-innovation can avoid fierce competition, make special products.
Also, scarcity, the product will be better if there is scarcity, Chinese cultural elements are very popular overseas.
Finally, reference to Amazon, if it is a popular product on Amazon and meets the above conditions, it is likely to explode on TikTok.
The essence of explosive products is the product of the platform algorithm and human weakness. To put it plainly, it is based on the data of the platform, the data of the mall, the data of the short video of the explosive video to do the dismantling and analysis, and must understand the algorithm of the platform. At the same time to go to the weaknesses of human nature, to solve their human pain, especially when doing festivals, to pay much attention to this point, consider how to provide consumers with emotional value based on products other than. Jiang Weibing said.
In addition, in TikTok test product rhythm, Jiang Weibing suggests that after selecting a good product, with 500 1000 number of sea transportation, 200 500 number of air transportation to try the product field. Then use their own short video matrix shooting, low cast flow test. Among them, the paid click-through rate is greater than 3 daily sales to reach 20 single for the initial success criteria, if not reached, then start to do Darren promotion. Through the above steps, to achieve daily sales of 20 single after the beginning of replenishment and iterative design, such as replenishment of the light to increase the mirror and other micro-innovations, and then the cycle.
Finally, in terms of the core logic of TikTok material, Jiang Weibing pointed out that to vigorously out of the miracle, the U.S. district short video has not reached the stage of refinement, fast and efficient out of the short video is most important. In addition, the explosive video can be copied, the same explosive video batch registration is feasible, the official account can also continue to release the same terms of the short video The weight of the self-guaranteed account is very large and should not be ignored When making short videos, we should pay attention to the details, such as the sense of persona Dialect Wisdom and the labeling of the official hot topic need to pay attention to.
How to get started quickly and at low cost in TikTok US area
According to EchoTik statistics, the GMV of TikTok U.S. e-commerce in 2024 is more than $7.5 billion, and its GMV global share is close to 30, which is the largest e-commerce market of TikTok Shop. In this context, many sellers are concerned, what is the globalization strategy layout of TikTok in 2025? In what ways can new opportunities be captured? And how to realize a quick start at low cost with the help of AI RPA tools?
In response to these concerns, CC, the founder of billion-dollar seller C Cow Overseas, answered them one by one on the scene of the cross-border symposium of Pai Dai.
CC pointed out that the future trend layout of TikTok is great. From the perspective of global layout, if a country can achieve one or two hundred million in sales, a combination of multiple countries can form a large scale, and for brands with capital, the value of the international brand will be greatly increased, which is also the opportunity we see from the map.
The TikTok market has huge potential. However, in-depth market insights are essential to succeed in this market. CC also shared the market insight logic at the conference and gave a comprehensive analysis starting from the Amazon package. He emphasized that doing cross-border business cannot simply copy the domestic selection logic. They previously applied the success of domestic Jitterbug to TikTok in the early days, expecting to drive business, but lost more than 100,000 yuan per month in the first few months. It wasn’t until January 2024 that sales rebounded to more than 900,000 yuan a month, and that’s when they realized they had to return to the basics of inbound inventory stabilization.
CC analyzed the reason for the previous loss is that with the domestic jitterbug’s understanding of content, such as the first 7 seconds of logic and algorithmic understanding to do TikTok, but found that the American understanding of content is different from the domestic. The content of the American area is on the real side, the first three seconds of the camera may be shaky No special effects subtitles Music is not prominent, but this kind of real content is very popular.
In addition, CC said that in cross-border business, market insights are crucial. All the verified explosive products must have large-scale transactions in the local area. In China, explosive products need RMB votes, but in cross-border business, they need USD votes. In the process of placement and measurement, it must be supported by data to decide the next move. Many cross-border sellers fail due to supply chain problems, so it is important to do a good job of fully evaluating in terms of the billing period, the final payment, and so on. At the same time, it is important to make appropriate short video content expression.
What is considered comprehensive market insight?
CC pointed out that there are three main aspects, the first is to analyze the data through, the industry brand single product data comprehensively pull out to see the transaction size. The second is to look at Amazon’s top 10 or top 20 brands, but why TikTok did not do, this thing must be thought out clearly. The third is merchandise, to see where the core competitiveness of their own goods.
If you have not gone to see TikTok’s market, you must look at it comprehensively, and then you must be in awe of TikTok, like Amazon and TikTok there is a big difference. CC said.
It is worth mentioning that CC showed the market insight report made with AI RPA on the spot, taking a certain category as an example, and introduced how the report is broken down and analyzed from insider stores merchandise playing style and so on. For example, through the graphic search, the AI is given instructions to analyze the basic data of the store operation play strategy positioning marketing products promotional activities, etc. It can also analyze the commodity information and give suggestions. The whole report production process used to take a long time to complete manually, but now AI RPA only takes 15 minutes, greatly improving efficiency.
Through the market insight method, it is possible to have a more comprehensive understanding of the market. In the actual product creation process, it is crucial to accurately grasp the selling points of the product.CC also introduces the method of exhaustive selling points to help you better explore the unique value of the product.
He pointed out that you can’t subjectively determine the selling points of a product, you have to rely on market data. The first step is to exhaust the selling points, grab a large number of videos of similar products, record the selling points of every frame and every second. Such as grabbing 150 breast sticker videos, analyze the selling points of each video, and discover the selling points mentioned by most people through large-scale data sample collection. Doing it manually has problems such as low efficiency, imprecision and subjective bias.
CC mentioned that AI RPA can be used to achieve fully automated analysis, where the computer automatically downloads the video, extracts the voice, uses Xunfei’s big model to interpret the dismantling and extraction, tags and registers, and quickly generates reports. You can also find strengths and weaknesses and opportunities from competitor reviews, by capturing reviews Extracting keywords Summarizing and using RPA to achieve automated operations, such as analyzing reviews to quickly identify competitor product issues, such as inaccurate size charts, so as to find differentiated selling points.
In addition to clarifying the market insight methodology and product building points, an efficient team is crucial for business advancement.CC also highlighted the importance of building a team. His company transitioned from domestic e-commerce and failed to make TikTok in the early days with domestic Jitterbug experience, realizing that it didn’t know much about American culture. In the job training, the first step is to let employees know the United States, understand its formation Population share Warehouse allocation, etc.. Different stages of the company’s genes are different, and growth strategies and strategies are different.
In terms of organizational structure, CC’s company is divided into front office, middle office and back office. The front office is related to sales, the project team has different positions such as new countries project team US project team, the middle office provides services to the front office, and the back office is responsible for calculating the administrative costs.
In terms of performance management, CC showed the internal performance management navigation bar table, which covers the performance appraisal of all positions in the front office, middle office and back office, such as trader, pitcher, BD, editor and director.
CC emphasized that a trader should have solid professional skills, project coordination skills, team management skills, landing skills, in-depth review skills, and risk control skills, as well as familiarity with professional workflows and a deep understanding of TikTok. Through the automation of forms and RPA, individual capabilities are transformed into team capabilities, thus realizing scientific management.
In addition, CC attaches great importance to talent recruitment, using multi-position coverage active recruitment and performance incentives to recruit talent. Behind talent acquisition is a commitment to the project. The company is convinced from the top down that TikTok offers opportunities, and the employees are confident,” says CC. The employees are confident,” says CC.
Finally, CC pointed out that there are great opportunities in the future, TikTok Chinese sellers have accounted for a certain proportion, China has experienced the Jitterbit algorithm has an advantage in the U.S. market. In addition, people who can do well in the bonus period should have the quality of extreme pragmatism, not empirical, not self-righteous, not arrogant, and have the determination to break through the steel plate.
The above is the core content shared by the guests of Pai Dai cross-border symposium, and I hope it can inspire you to find new growth opportunities and realize new growth goals in the cross-border e-commerce field.
Boss into the group
E-commerce bosses engage in profit See opportunities
Come on, pay for pie. E-commerce growth. Community.
15 billion, 40 million sellers have joined.
2 sessions per month
Daily precise resource matching and topic discussion
9 offline closed-door meetings per year
Limited to business owners, partners, core managers
Hot Articles
Leave a Reply