Profile
You can use AutoML without building a data team: Transformation God teammates The retail industry is currently the industry that can best use data modeling to help companies make the best decisions and recommendations, and has achieved remarkable results, such as administrative financial personnel, turnover forecasting, performance forecasting, turnover rate forecasting, job performance forecasting, rotation, etc. ; It can also be applied to supply chain management, such as front-end inventory management, production demand, sales forecast in daily business, customer lifetime value and promotion strategy...etc. Manufacturing: all five elements are measurable The five major elements of man, machine, material, method, and environment are often mentioned in the manufacturing industry, and different predictions can also be made through models. In terms of manpower, it includes turnover rate, work performance, rotation and department transfer recommendations, etc.; material preparation can be forecasted based on production volume and demand. In terms of machines, it includes equipment and machine maintenance forecasts. For example, many companies need to use the experience of master craftsmen to carry out regular machine maintenance or maintenance operations. AI can convert the experience value of master craftsmen into evidence-based data. In terms of environment, for bad products, the bad factors that cause defective products can be quickly found through the model, and quickly eliminated. In addition, in terms of quality and energy, electricity consumption or water consumption can also be predicted, and energy consumption has a lot to do with factory cost considerations. Telecommunications Industry: Customer Demand or Sales Forecasting It can conduct quarterly mobile phone sales forecasts for store access, including forecasts of main products, shipments, and replenishment volumes; in terms of network applications, it can assist in financial management and procurement or customer cycle forecasts, such as customer churn forecasts or customer life cycle etc. Financial industry: Anti-fraud risk or credit scoring Applicable scenarios include anti-fraud risk, credit scoring, customer relationship, and financial management. Others such as fraud, money laundering, detection, customer scoring and loan default prediction, customer lifetime value, etc.; or recommend suitable financial products or financial products for customers with different attributes. After introducing the industries where machine learning can be applied, I will share two business cases that have achieved remarkable results after introducing AI. The first case: China's top ten benchmark food companies - refrigerated product shipment plan How does the company with the second largest market share in the country use AI technology to face the common point of "imbalance of production and sales" and lead the company to break through the plateau period for many years, using big data and ML models to successfully reduce the company's refrigerated products by 2.74 million yuan or about 12 million Taiwan dollars Costs derived from scrapping. The second case: FB advertising forecast The father of department stores, John Wanamaker, said: "Half the money I spend on advertising is wasted, but the problem is that I don't know which half." According to past experience, 75% of advertising needs to be closed. At this time The ML model is needed to assist the store to judge, "How to maximize the advertising benefits?" The direction of ML modeling processing is generally to collect data, put background information, index, CPM, CPC, ROAS, etc. into the training model, and the AI model produced can help the store to judge whether the advertisement meets the expected goal, and make a ROAS Simple judgment, if the ROAS is less than a certain value, it can be regarded as an invalid advertisement and used as a reference value for the next advertisement. Here are two examples. One is an e-commerce skin care product with a short advertising cycle. After importing the AI model, it was found that 96% of the advertisements were invalid advertisements. After closing, it saved nearly 35.5% of the budget and saved about 71% per month. Ten thousand. The other is an e-commerce brand with a long advertising cycle. After introducing AI, it is found that nearly 75% of the advertisements are relatively inefficient and should be closed as soon as possible. A total of nearly 20% of the marketing budget has been saved, which is about monthly savings. About 200,000 marketing budget.
Forum Role: Participant
Topics Started: 0
Replies Created: 0