Taobao's agent operation: the distribution principle of search traffic of Taobao

Today, let's talk about the distribution principles of mobile search traffic

After the new Taobao products are launched, the system will allocate traffic to Babies in an irregular period of time. During this period, if Babies have a high click rate and good click feedback effect, Babies will continue to get more traffic. Note: In fact, high-quality traffic has already been labeled by Taobao. The highest quality traffic will be allocated to top tier merchants. The higher the level of your store, the better the traffic will be.

In fact, this is about a problem of high-quality traffic and low-quality traffic. The personalization of online search is becoming more and more serious, and the system also stratified consumers. To put it simply, those who have strong consumption power and often buy products with high customer price are high quality traffic. Consumers with poor consumption capacity and low customer price belong to poor quality flow.

Since there is a hierarchy of consumer groups, it also corresponds to the product classification. After the launch of Tmall brand new products (brand support), Taobao has a high popularity weight (high customer price, large number of online users, high collection and purchase, and high repeat purchase of old customers), which naturally will be allocated to the display opportunities of high-quality people. Note: this is about display opportunities, not traffic! On the contrary, nameless products with low customer price can only match the flow of inferior quality with that of superior quality.

Next, we will look at the three main search algorithms.
Comprehensive search=keyword popularity weight+baby weight+store weight+weighted item
Popularity search=popularity weight of keywords=in the 7-day dimension=click rate+conversion rate (twice higher than the industry)+baby keyword is the search conversion has been growing
Sales search=keyword popularity weight+wireless end proportion (traffic, sales, conversion rate, wireless end advertising proportion)+weighting item

Here we mainly talk about popularity search and sales search. The most important factors that affect the search weight of Taobao are baby weight and keyword popularity weight. What factors will affect the popularity weight of single product? There are many factors that affect the popularity weight value of a single product, but the main impact is the sales increase ratio, click rate and conversion rate in the 7-day dimension of a single product, which is a direct factor.

Assessment criteria for manual panning search algorithm
First of all, you should know that the popularity search model is used by the mobile search algorithm. How to understand this sentence? When we browse a product through keyword search on Taobao, the search results page will involve comprehensive search, popularity search and sales search. Therefore, the more popular the product is, the more likely it is to be matched by the search system on the handheld wireless terminal.

"Some people may ask how the search system determines whether a single product is highly popular."

The main assessment criteria are: new product weight+popularity weight (popularity weight here is also generally divided into two data indicators, baby popularity year-on-year data+baby popularity month on month data. This is also the most important part of the evaluation criteria, because it belongs to the popularity algorithm in our model. You should know that the popularity algorithm is the most important one in all platforms. Popularity means that when your baby's traffic continues to grow this week, you will find that your traffic will continue to grow next week if there is no reduction in click conversion rate.)

For example, the system has given you 5000 traffic opportunities. Because the average click through rate of the industry is about 5%, and the click through rate of your new products is 1% - 2%, you can only get 1500 traffic opportunities in the end. (Note that the system here is not really giving direct traffic, but showing opportunities. However, since your product click rate is lower than the industry average, it will miss the system support period in the early stage of new products, and the weight value of the baby will not be high.)

This is mainly to assess the utilization rate of single product flow. It includes:
1. Preference of single product (collection rate, purchase increase rate, browsing duration, visiting depth, page skip rate)
2. Flow contribution value (that is, what we call flow value, that is, output sales of each flow, formula=total sales/total flow)
3. Single pit output (conversion rate, per customer commodity rate)

How to make a highly popular mobile phone
In the process of operation, it mainly controls the popularity search model in the search algorithm.
For sellers, in addition to paying attention to the decoration, page details and traffic of PC stores, they should also pay attention to the traffic of online shopping. Taobao novices may think that I haven't even done a good job in PC traffic, so I don't have the time and energy to do mobile. In fact, no matter what kind of platform, the overall operation idea in traffic acquisition is the same.

How to optimize online search traffic
The core of search traffic is actually the comprehensive ranking of all the keywords in the thousands of people

In terms of single products, if there are significant changes in the short term, it is mainly the click through rate, conversion rate and pit production that play a role, which is also the reason why some people have obvious effect of brushing real-time data. Of course, the weight of single products is also a cumulative process, and the number of clicks, views, collections, additional purchases, transactions The transaction amount, dwell time, and the praise rate, return rate, repurchase rate, average dwell time, detail bounce rate, as well as the population weight generated by these data will be evaluated periodically and cumulatively.

I will list the important data indicators and calculation formulas below. Combining the above contents, you can find an optimization direction.
(1) Single pit output=customer unit price * monthly sales volume
[Each category has a corresponding output threshold, and each keyword has a corresponding output threshold. When you reach this output threshold, the flow will explode. There is no standard formula for calculating this threshold, which is based on market changes and experience, or you can only try to touch the approximate location]
(2) Conversion rate threshold=conversion rate/conversion rate index * 1000 (the same with the keyword conversion rate threshold)
(3) UV value=sales volume/number of visitors (sales volume=number of visitors * conversion rate * per customer price)
=Conversion rate * unit price per customer
(4) PV value=sales/views
(5) Conversion rate=number of transaction users/number of visitors
(6) Collection rate=collections/visitors
(7) Additional purchase rate=number of additional purchases/number of visitors
(8) Average Access Depth: Access depth refers to the number of store pages that users visit continuously at one time (i.e. the number of pages browsed in each session), and average access depth refers to the number of store pages that users visit continuously at each time.
(9) Per capita stay time in the store (seconds): the average stay time of each successive visit to the store during the visit of all visitors.
(10) Skip rate: it indicates the proportion of the number of visits that customers enter through the corresponding portal and leave after visiting only one page in the total number of visits to the portal.

The return rate and dispute rate will not be mentioned, but they are also things you need to pay attention to.