Want to change into a new hairstyle, In the field of digital marketing, it will be a great challenge for the retail industry to collect and analyze user data behaviors to achieve precise advertising and marketing. Mingliu International, which has multiple branches across Taiwan, has not set up dedicated data analysis manpower, so that valuable member information cannot be fully utilized. Fortunately, through the Digital Industry Agency of the Ministry of Digital Development, the “AI Industry Practical Application Talent Training Program” Cooperate with the AI start-up team to strengthen the advertising reach and improve the advertising efficiency.
Connect online and offline data to predict advertising effectiveness with one click
During the first three months of proof-of-concept period, Zhizao Qiji’s AI team found that the availability rate of name and retention data was only about 50%. The main reason was that various fields of the POS system were not filled in. The data, but the necessary fields such as age and gender are missing. After pointing out this problem, after modification and adjustment, the utilization rate of the data is close to 100%, which can be used by the AI model. Online and offline are connected to create an API, just enter a few items parameter, you can predict the reach of your ad serving.
For example, the main customer group of a certain branch is 18-30 years old, mainly women, who come to the store on average once a month. If such a consumer profile can be used to place online advertisements for the branch to attract potential customers with the same characteristics, the operating performance can be optimized.
Then, the AI team further integrated the historical records of Mingliu’s previous advertisements on Facebook, such as the audience, reach rate and conversion rate, and integrated these online data with the offline data of the aforementioned POS system. And hand it over to the AI model for analysis, increase the richness of data, and improve the accuracy of AI predictions.
The industry’s unique number of visitors in low and peak seasons is handed over to AI for integrated interpretation
Another very unique data change factor, which also needs to be handed over to AI for interpretation, is the time factor of the number of visitors in low and peak seasons. The data processing and analysis of the hairdressing industry requires time-series grouping, mainly because the industry’s low and peak seasons are deeply affected by holidays and seasons. For example, spring and summer are hot, and haircuts are frequent, while autumn and winter are longer, but on the eve of the New Year, there is another wave of peaks. “The unique characteristics of this industry can only be known by communicating with the questioning party. This is also the most valuable part of the AIGO platform, which is the characteristics of the matching industry and the talent collection industry.” The team of Zhizao Qiji said.
Using AI to assist in analysis has another advantage: eliminating human subjective bias. For example, a certain branch store’s advertising thinking mostly targets young women, but the AI model found that the store actually has a lot of male main customer groups, but it’s just the subjective experience of marketing personnel and designers who are used to targeting female customers. This unexpected discovery also highlights the importance of the combination of industry and AI, which can identify existing blind spots and improve marketing efficiency.
AI predicts the reach rate of online ads with an accuracy of 94.5%
After the above efforts, the AI system created by the team does not need to hire data experts to maintain. Its intuitive and friendly interface can be easily operated by marketers. Just enter the advertising time, audience age, gender, estimated advertising budget and conditions In just a few seconds, the system can predict the amount of reach, and the accuracy is as high as 94.5%.
Due to the success of the initial proof of concept, Zhizao Qiji and Mingliu International have launched the second phase of cooperation-automation, that is, the real-time information of store customers and Facebook ads is automatically imported into the AI system for analysis, so that the AI model continues learning to further improve accuracy. The AIGO team said, “Because of AIGO’s competition platform, we can open up more future cooperation opportunities like this, so that domestic AI talents can deeply understand the needs of the industry and have practical energy.”