Li Qiguang touched his sparse hair and went out to do some work.
"Uncle Guang, I noticed that you've been shedding a lot of hair recently." Fang Tian looked at his head. His hair was getting thinner and thinner. If this went on, it would soon become a desert.
Li Qiguang smiled and said, "Yes, you'll be smart soon!"
Fang Tian drank his coffee and smiled. "Is it because of work pressure? I'll give you a break."
"No, it's hereditary. My dad lost all his hair when he was 50." Li Qiguang smiled helplessly.
Lin Keqing said, "You can wear a wig."
Li Qiguang asked, "What good recommendations are there?"
Speaking of recommendations, Fang Tian thought of a very important thing. He opened Thousand Gold Mall and browsed the website.
He found that the recommendations on the website were not ideal.
Seeing Fang Tian frowning, Li Qiguang was puzzled. "What's wrong? Do you have hair shedding too?"
Fang Tian pointed at the computer screen and focused on the recommendations on the website. "Did you notice that there's a problem with the recommendations on the website?"
Li Qiguang walked over and looked. "What's the problem? They're all recommended to consumers by the website. "
"But I'm a man. The website actually recommended high heels for me. Don't you think that the website's recommendation is stupid?"
Li Qiguang said, "It must be so. The website doesn't know if you're a man or a woman."
Fang Tian shook his head. "This kind of recommendation is not what consumers want. I need personalized recommendations!"
For e-commerce websites, recommendation systems were extremely important. They could very considerately guess what consumers needed and then recommend products to consumers, thereby increasing the sales of the products.
In this day and age, there was no such thing as intelligent personalized recommendations.
There were basically two types of recommendation systems on all websites.
The first type was manually selected by the website editor and then placed the products on the website's recommendation spot.
The second type was random recommendations.
They were sorted according to time and the latest products would be placed on the recommendation spot.
But no matter which type it was, it was not intelligent personalized recommendations.
When consumers opened the website, everyone saw the same recommendation content.
For example, Fang Tian was now looking at the recommendations on the front page.
The first was high heels.
The second was leather shoes.
The third was a children's toy car …
Whether it was his own account, Li Qiguang's account, or Lin Keqing's account, the recommendations were the same.
Such a recommendation was very unreasonable.
Lin Keqing said, "I also feel that such a recommendation is unreasonable. If the website can be considerate enough to guess what I need, it can increase my interest in purchasing."
Fang Tian instructed Li Qiguang to call Yang Fan. Yang Fan was in charge of designing the website.
Then, Li Qiguang dialed Yang Fan's number.
Not long after, a dark-faced young man walked in. Yang Fan's skin was the same color. If one didn't know better, they would have thought that he was from Africa.
"Tian, why did you call me here?" Yang Fan pulled a chair over and sat down.
Fang Tian's fingertips kept tapping on the table. "The website needs a recommendation system."
"Is it important?" Yang Fan asked.
"Very important." Fang Tian's tone was firm. "This can greatly increase the user's activity. More importantly, it can increase the conversion rate of the products."
Conversion rate, to put it bluntly, was the number of views and trading volume. If 1,000 people opened the recommended product page and only 1 person bought it, the conversion rate was too low.
If the recommended product was of interest to the consumer, it could greatly increase the trading volume.
Yang Fan understood. To know the consumer's personalized needs, the key was in the recommendation algorithm.
"How do we do this recommendation algorithm?"
Fang Tian touched his chin and thought for a while. "The first way is to recommend products based on the consumer's browsing and search history. This is the simplest and easiest way to guess the consumer's needs."
This was not difficult to understand. When a user shopped on an e-commerce website, he searched for something, browsed something, but did not immediately place an order. The website could recommend it to the user.
Yang Fan nodded. "What's the second way?"
"The second way is to recommend products based on the consumer's purchase history. If the consumer has bought it before, and the product is not durable, we can recommend it to the consumer. "
Yang Fan thought for a while and completely understood. If the consumer has bought it and given a good review, it proves that the consumer is satisfied with the product. The website can recommend it to the consumer.
However, it cannot be durable. For example, if the consumer bought a mobile phone in the mall half a month ago, it is impossible to recommend him to buy another one.
But if it is a pack of Nescafe coffee, yes, this can be recommended to the consumer again.
Fang Tian took a sip of coffee and said, "The third way is to recommend products based on the peripheral products. For example, three days ago, the consumer bought a mobile phone in the mall. We can recommend products related to the mobile phone, such as a headset or a protective case. "
"For example, if a man bought diapers, we can recommend him to buy beer."
Yang Fan was puzzled. "Are diapers and beer related? Will the man who buys diapers buy beer? "
"You have to ask Li Qiguang this question."
Li Qiguang's son was born not long ago. He had the most right to speak.
Li Qiguang smiled helplessly. "Good recommendation. Every time I buy diapers, I feel very depressed!"
"Haha!" Fang Tian and Yang Fan laughed.
"The fourth way is to classify the product according to the user. The website will classify the consumer according to his age, purchasing power, interests, and so on. "
"Through this intelligent algorithm, we found that A and B are the same type of consumer."
"For example, Customer A bought a comic book and then an Ultraman toy. Then, when Customer B buys the same comic book, he can refer to Customer A and recommend an Ultraman toy."
Yang Fan, Lin Keqing, and Li Qiguang listened to Fang Tian's recommendation algorithm. They felt that it was very interesting. If they did as he said, the recommended products would be more considerate.
In the following time, Fang Tian said a lot more. Yang Fan recorded it all.
This recommendation system sounded simple, but it involved a very complicated intelligent algorithm. The technical aspects would be optimized and explored by the technicians.
Relying on this powerful recommendation system, the shopping experience of the mall would be far ahead of other similar websites.
In the end, Fang Tian said, "This recommendation system can not only be used on shopping websites. It can also be used on other Soft cloud applications."
Soft cloud did not only have e-commerce websites.
Soft cloud news under it, if this algorithm was used, the recommended news would be closer to the user's interest.
Soft cloud videos, Soft cloud Weibo, Golden Jade Novel Network, and so on could also use this system.
You've already exceeded your reading limit for today. If you want to read more, please log in.
Login
Select text and click 'Report' to let us know about any bad translation.
