The essence of Taobao search algorithm 2017

　　The adjustment of each search algorithm means the reallocation of traffic, which will involve the vital interests of many Taobao sellers and the interests of the platform. For Taobao, your so-called technology will eventually fail. When it fails, the losses may be greater. Therefore, I would rather believe that it is crucial to solve the essence of search algorithms. Some time ago, I made a mind map "2017 Taobao Natural Search Weight Model and Optimization Techniques", which should be helpful to you a lot, see the picture below:

　　Each search rule adjustment must be carried out in a very cautious state, and it will definitely consider various possible situations. In a sense, every rule adjustment is the platform and most sellers are beneficial (if you don’t want bad money to drive away good money). Think about the incident of Taobao being criticized during the Two Sessions, and you should be more at ease.

　　Actually, I think the so-called Taobao SEO is the process of Taobao sellers fighting with the search department. The ultimate goal of the search department is to plug all loopholes, and then you will eventually find that the cost of going to the evil path is the same as the cost of going to the right path (even higher than the cost of going to the right path). At that time, I think no one would take the evil path.

　　1. You must first figure out what kind of program Taobao’s search algorithm is

　　Therefore, Taobao’s natural search sorting is the same as that of the college entrance examination. When taking the college entrance examination, you have to take different subjects, and the product scores required for natural search sorting also need to refer to different factors. However, some universities may focus on math scores, while some universities may focus on the same English scores; some categories on Taobao may have high sales factor weights, while some categories may have high evaluation factor weights.

　　The balance of interests of buyers, sellers, and platforms is constantly changing in the search rules of promoters, but at different points in time, there may be a little focus.

　　For example, when Taobao was first started, it needed to attract more sellers to sell things on this platform. At this time, we had to take care of the interests of most sellers, and taking care of the interests of most sellers, the most typical thing was to give every baby uploaded by every seller a chance to show it to consumers. Therefore, the initial rule (even the only rule for a certain period of time) is the removal time rule: the closer it is to the removal time, the more likely it is to rank higher.

　　The core of the search rules is: recommend the most suitable products to consumers who are most likely to make a deal, while allowing the platform to benefit from the most well-known ones.

　　In fact, the process of search engine recommendation is such a process: when a user performs search behavior, the search engine will rate products that meet the conditions (the title contains this keyword, does not violate the rules, and is in line with the relevant) based on the user's search engine, and then recommend them to consumers in the order of scores.

　　2. Next, let’s take a look at what parts of the product’s search scores are composed of

　　The search score of a product mainly comes from two aspects, one is the score of the product itself, and the other is the score of the store where the product is located. The scores of the product itself include: collection popularity, purchases, conversion rate, sales volume, growth status, whether a single product is illegal, etc.; the scores of the store include: the store's sales rate, the store's DSR score, the comprehensive conversion rate, the new purchase situation, Wangwang's response speed, etc. Consumer behavior (including browsing, conversions, collections, etc.) will largely determine your product score and store score, and then determine the position your product will appear in front of consumers.

　　In the past, in the era of hot-selling models, the score of the products themselves was mainstream, so you often see this phenomenon: a certain store is not well decorated, a large number of products are unsold, and only one or two hot-selling models are available, but the ranking of these or two hot-selling models has always been very good, and the phenomenon of single products bringing the entire store is very common. But now, with the opportunity to become popular, more and more products have the opportunity, and the weight of the store will increase. If your store is not done well, the ranking of hot-selling products will easily fall off. Therefore, in the past one or two years, the weight of various indicators such as store level, sales rate, and DSR score has been increasing.

　　3. Continue to analyze: What dimensions will consumers reflect the impact of your search scores?

　　Consumer characteristic dimension: that is, what kind of consumers will come to your store, whether it is buying, browsing, or collecting and adding purchases. Some characteristic dimensions of the consumer themselves will give you bonus points or subtraction points.

　　Consumer behavior dimension: This behavior dimension has become more and more obvious now. Have you noticed a situation: when you search for some keywords on Baidu, and you open the browser, you will find that sometimes some Taobao or JD products pop up, and these products are related to your search behavior just now.

　　Output dimension: In fact, the output dimension is mainly considered from the interests of the platform (of course, the interests of the platform can be realized because it meets the interests of consumers). How to understand? On the premise of giving you the same opportunity, whoever can produce more sales should get a higher weight.

　　4. What you need to consider next is the issue of regular weights and contrast weights

　　In fact, this mainly considers data exceptions. Because every purchase behavior of every consumer is largely regular, and all data indicators are also regular when operating Taobao.

　　If you are buying a very complex and highly valuable digital camera, then normally you have to look at it for a while, then consult the store owner, and then you have to compare more sellers; on the contrary, if you just buy a pair of socks, the average order price is more than ten yuan, you may just look at one or two stores, and you will convert silently without consulting. Therefore, if most of the visitors' behaviors in your store do not match the normal behavior, the weight will drop significantly.

　　Thank you for your attention and support to Laogao Crown Club . Please indicate the source of the reprinting website www.shxuanming.net