Data analysis method and Thinking: funnel analysis

What is funnel analysis ？

Every step from the beginning of the business process to the completion of the final goal will lead to the loss of users , Therefore, an analytical method is needed to measure the transformation efficiency of each step of the business process , Funnel analysis method is such an analysis method .

Funnel analysis is a method that can visually present the user behavior steps and the conversion rate between each step , An analytical method for analyzing the conversion between steps .

A simple example , When you open e-commerce APP When you want to buy something , You will go through at least the following steps ：

• open APP Enter the home page
• Click the advertising page of a product on the home page
• Go to the product details page , I think it's good , Click to open the shopping cart
• Go to the shopping cart page , Fill in the express information , Click to pay
• Enter the payment page , Complete payment , Merchant delivery

After entering your shop , Its behavior funnel is like this ： home page → Details page → The shopping cart → place order → payment , Five steps . These five steps are indispensable , Therefore, there is a context . That is, the previous step must be completed , To continue to complete the next step .

In the process , Some people find that the object is not good-looking , Left on the details page .

Some people think the price is too expensive , Left the shopping cart page .

100 Someone walks into your APP, The only possible way to complete the payment is 1 individual , Then your payment rate is 1%.

however , If you can do 1% Payment rate , You'll make a lot of money .

The value of funnel analysis

The value of funnel analysis mainly includes : Function optimization , Operational launch , Loss of users, etc

Function optimization :

Let's take the website as an example . For example, you create a website for making pictures , You put it in SEM It's coming 10 Ten thousand people . After these people click into your website , The conversion rate of uploading pictures is only 30%, This time you will reflect , Whether the upload portal is not obvious , No one can find the upload button . therefore , You optimize the revision , hold 【 To upload pictures 】4 Words enlarged and bold , Put it in a prominent position , The next day, the conversion rate of uploaded pictures reached 90%.

If you find that the payment rate is particularly low , You may wonder , Is the payment method inconvenient , For example, people support Alipay 、 Wechat payment , And you need to pay by flash . therefore , You improve the function , Added more convenient payment methods , Then you find out , The payment rate has been improved .

Operational launch ：

Drainage to get customers , It is necessary to put in advertisements . For example, baidu advertising 、APP Open screen advertisement 、 Tiktok advertising, etc , The purpose of launching is to attract users to consume .

Therefore, the delivery funnel of the operation is ： Advertising reach —— Click to enter —— Browse —— The trial —— pay

If you put out an ad page , No one clicks , That means we are not interested in the advertising page , You need to change the advertising page .

User churn :

Take e-commerce as an example app Take Taobao as an example , If the number of orders drops this month , At this time, we need to sort out the purchase links of users , Open users from app All the links to the order are sorted out , Then the funnel analysis was used , Calculate the conversion rate of each link

If we find in the link , The conversion rate from searching products to viewing products is very low , So we need to see if a lot of searches are fruitless , Or the search results make many users dissatisfied , The result is that the user doesn't pay the bill .

That can turn the payment problem of e-commerce into the problem of search , In this way, we can do a funnel analysis on the whole transformation link of the search , Step by step to locate the problem .

Application of funnel analysis

The most famous funnel model for user growth is called AARRR, That is, get... From the user , User activation , User retention , Users pay to users spread

1. User access

User access , The most classic is pinduoduo , All the people cut , You can get it by cutting so many knives 100 Yuan cash . although 100 Yuan looks like a lot , But this 100 Yuan can reach at least dozens of potential users , And these potential users also want to take 100 Yuan cash , They will go to dozens of people to help her cut , An explosive chain reaction produces .

2. User activation
When new users are pulled , Make sure to activate him to the greatest extent . Let's say the user APP Signed up for an account , Start testing the product , The user has been activated successfully .

3. Retain

How can users open it every day APP Well ？ To improve user retention , A lot of websites 、APP Launched the sign in and punch out function , You can get corresponding discounts or gifts by punching cards for several days .

4. User payment

How to get users to place orders , This is the hardest point . Your price is cheaper than others , Your quality is still high , You can also use coupons , High praise rate of the product, etc , Can promote users to pay . yes , we have UI To stimulate payment , Also deliberately put the upper right corner of the x Do it inconspicuously , and 【 Buy now 】 The buttons of the are especially attractive .

5. User communication

spread , It can be social media communication , It can be spread by acquaintances , Word of mouth . such as , A certain lipstick has a beautiful color , Then it gradually spread among the sisters , It has gradually become a popular item .

All right. , That's all for the funnel analysis . More knowledge of data analysis , Lockable Simplilearn Data science training camp in St. Prynne . adopt Simplilearn St. pryn's data science program , You learn statistics and statistical procedures 、 Hypothesis testing 、 clustering 、 Decision tree 、 Linear and logistic regression 、R、Python Language 、 Data visualization 、 The regression model 、Hadoop、Spark、PROC SQL、SAS Macro and other skills and tools 、 Advanced analysis 、Matplotlib、Excel Analysis function 、Zookeeper、Kafka Interface and other knowledge and skills . If you follow the scientific learning path , Once a week 8 h Learning time , 21 In a week, you'll be a certified data scientist .