中文字幕欧美视频,国产成人午夜精品影院游乐网,日本二区久久精品鲁丝 http://www.dpkxx.com/en 移動應用運營平臺 Thu, 08 Aug 2019 08:50:05 +0000 en-US hourly 1 https://wordpress.org/?v=4.8 http://www.dpkxx.com/wp-content/uploads/2017/06/C512-c.png APP統計 – Cobub http://www.dpkxx.com/en 32 32 Python USES Deep Neural Networks to Identify Siamese and British Short. http://www.dpkxx.com/en/python-uses-deep-neural-networks-to-identify-siamese-and-british-short/ Mon, 05 Feb 2018 03:25:59 +0000 http://www.dpkxx.com/?p=7317 Let's take a couple of pictures and see how the cat is Siam? Which cat is short?
First Siam

Python USES Deep Neural Networks to Identify Siamese and British Short.,首發于Cobub

]]>
Let’s take a couple of pictures and see how the cat is Siam? Which cat is short?
First Siam

Second short

Have you ever been able to identify Siam and British short? Probably, it doesn’t seem to work. This is because there are too few materials, and we can see that these two pictures can be extracted from each other with too few short features. What if we Siam short put 100 picture, the short put 100 picture for your reference, and give a Siamese or English is just a photo can be identified is that a cat, if not recognized, also has a 90% can guess may be right. So if you provide 500 pictures of Siamese 500, are you more likely to guess right?
How do we identify Siamese and British short? It is first summarized the characteristics of the two cats such as facial color, eye color, etc., when have a picture to identify short, we will see if facial color, eye color can be characteristic of Siam.
Will computers be able to identify the two cats as well, after learning how to identify Siamese and English short?
So how do computers recognize images? Let’s look at how computers store images.

The image in the computer is a pile of Numbers in order, 1 to 255, which is a black and white picture, but the color varies from three primary colors – red, green and blue.

In this way, a picture is a cuboid in a computer! A cuboid with a depth of 3. Each layer is a number between 1 and 255.
To get a computer to recognize a picture, you have to let the computer know that it wants to recognize the features of the short image. Extracting features from images is the main task of identifying images.
Here is the main character, roll and neural network.(Convolutional Neural Network, CNN).
The simplest convolutional neural network looks like this.

It is divided into input, convolution layer, pooling layer (sampling layer), full connection and output. Each layer compresses the most important identifying information and transmits it to the next layer.
Convolution layer: to help extract features, the deeper (multi-layer) convolutional neural network will extract more specific features, and the more shallow network extraction will be more obvious.
Pooling layer: reduces image resolution and reduces feature mapping.
Full connection: flattening the image feature, treating the image as an array, and using the pixel value as the characteristic of the value in the predicted image.
Convolution layer
The convolution layer extracts the features from the picture, and the image is stored in the computer according to the format we mentioned above (cuboid). First, extract the feature and how to extract it? Use the convolution kernel (weight). Do the following short operation:

You look at the left and right matrices, and the matrix sizes are from 6×6 to 4×4, but the size distribution of the Numbers seems to be consistent. Look at the real picture:

The picture seems to be blurry, but what about the size of the two pictures? It’s in the following way: same padding.

You add a circle of 0 around the matrix of 6×6, and then you have a 6×6 matrix, and why you add a circle of 0 is related to the size of the convolution kernel, the step length and the boundary. Do it yourself.
The above is a demonstration of using a 3X3 matrix on a 6×6 matrix. What does it look like to convolve in a real picture? The diagram below:

A 28x28x10 activation map (activation diagram is the output of the convolutional layer) is obtained by convolution of a 32x32x3 graph with 10 5x5x3 filters.
Pooling layer
Reduce image resolution and reduce feature mapping. How do you reduce it?
Pooling is done alone on each depth dimension, so the depth of the image remains the same. The most common form of the pooling layer is the maximum pooling.
You can see that the image is obviously getting smaller. As shown in figure:

A new graph is obtained by extracting the maximum value of 2×2 on the two-dimensional matrix of each layer of the activation graph. The real effect is as follows:

With the increase of convolution layer and pooling layer, the characteristics of corresponding filter detection are more complicated. As you accumulate, you can detect more and more complex features. There is also a problem of convolution kernel optimization, and multiple training to optimize the convolution kernel.
The following USES apple’s convoluted neural network framework, TuriCreate, to distinguish Siamese and English short. (first of all, I have been working late in win10 to reload the computer more than 3 times. The system should have WLS, and it is convenient to install turicreae under the enterprise version, MAC system and ubuntu system.)
First of all, prepare to train with 50 pictures of Siam, 50 long. The test USES 10 pictures.
Code :(development tool anaconda, python 2.7)

The data is placed in the image directory of the h disk, and I am installing ubuntu in win10, so the h disk is hung in MNT/down.

Test files :(x refers to Siam, y refers to short, so the name is to distinguish the cat type from the test pictures in the code)

test_data[‘label’] = test_data[‘path’].apply(lambda path: ‘xianluo’ if ‘x’ in path else ‘yingduan’)
The first results are as follows:

The accuracy of the training accuracy is 0.75 and the accuracy is 0.5. Well, it seems that the study is too little, and it will take three years to simulate the five years of the college entrance exam, which will increase the number of Siam and English short pictures to 100. I’m looking at the results.

The accuracy of the training was 0.987, the accuracy of the test was 1.0, and the accuracy was 1.0.
See the results of turicreate recognition:

Our actual picture of the cat is :(red is the type of the real cat – in the code, according to the image name, green is the type of cat identified)

You can see that the two are consistent. The cow forces the training data only 200 pictures, can achieve this effect.

Python USES Deep Neural Networks to Identify Siamese and British Short.,首發于Cobub

]]>
How to Improve the User Churn Rate ? Here are 10 Strategies. http://www.dpkxx.com/en/how-to-improve-the-user-churn-rate-here-are-10-strategies/ Wed, 24 Jan 2018 01:17:37 +0000 http://www.dpkxx.com/?p=7295
Outline:
? What is user churn?
? Reducing user loss is a top priority for businesses.
? Ten strategies for reducing user churn.

How to Improve the User Churn Rate ? Here are 10 Strategies.,首發于Cobub

]]>

Outline:
? What is user churn?
? Reducing user loss is a top priority for businesses.
? Ten strategies for reducing user churn.
Many companies do not have specific plans to activate silent users and recall lost users. Some companies try to replace lost users with new users, while others spend a lot of time analyzing problems and trying to prevent users from losing them.
Before we discuss the methodology for reducing user churn, we need to understand how user churn is defined.

What is user churn?

User churnrefers to the number of users who leave the product within a specified period of time. According to different products and different businesses, the time period is different. The user loss indicator shows the ability of the product to retain users from the side. Companies will pay a heavy price for the loss of their users.

Data talks

When it comes to the impact of user churn, research has done the statistics, and the cost of getting new users is five times the cost of retaining existing ones. The harvard business school report also points out that a 5 percent increase in retention rates can lead to a 25 to 95 percent increase in profit. There are also data suggesting that retained users are major contributors to corporate revenue.

According to gartner, a consultancy, 20% of retained users will contribute 80% of the company’s future revenues. At the same time, the chances of successful sales to the remaining users are 60-70%, and the chances of successful sales to new users are only 5-20%.

The above data shows that reducing user loss and improving retention of users is the most important and beneficial to enterprises (especially e-commerce enterprises).

What are the adverse effects of user loss on the enterprise?

? Reduce enterprise income, affect enterprise performance.
? Lower corporate profitability.
? Improve enterprise marketing and user recall costs.

There are 10 strategies for reducing user churn.

1. Analyze the reasons for user loss.

Why do users lose? The answer also needs to be found in the user, the most direct way is to talk with the user. Phone calls are the quickest way to talk to users. By telephone, we can know that our products do not solve the users’ pain points, what problems we have caused to users and so on. In addition to phone calls, we can also find out the reasons for the loss of users by sending emails to users, inviting users to comment on the website, or interacting with users on social media.

2. Maintain user engagement.

Maintaining user engagement can help prevent user loss to some extent. In order to maintain user engagement, we need to continue to prove the value of the product to the user. In addition to letting users know the main functions of the product and updating the content of the iteration, we can also show users new transaction information, special offers or recent preferential activities, etc.
Previous face-to-face interactions have been the main way for users to participate, but new research shows that websites and social media are becoming the main ways for users to participate.

Getting new users involved in the product is also a good way to reduce user churn. For example, we can ask new users about the first impression of the product, which helps us understand the initial impact of the product.

3.Give users adequate guidance.

Reducing user churn can be achieved by providing users with high-quality guidance/support information. These guidelines include but are not limited to free training, online BBS, video guidelines or product demonstrations. Good product functionality with enough guidance not only gives users the tools to solve problems, but also gives users a guide to the tools they use. The greatest potential of our products and services is to let users feel that we are paying enough attention to them and that it is not so easy for users to leave their products.

4. Timely detection of users on the edge of loss.

Data from the past loss of user behavior through the analysis, we can come to the conclusion that some loss of user joint behavior, such as they lost before that time is not as active as ever, loss and puts forward some problems to us before but didn’t get our feedback and so on. By identifying these common features, we can predict users who are on the edge of loss and take appropriate measures to retain them.

5. Identify high-value users.

Identifying high-value users and prioritizing the needs of these users is important because our revenue is primarily contributed by these high-value users. There are two ways to identify high-value users: one is to analyze the engagement of users in each stage of the product life cycle, and the other is to group users according to user behavior. We can have a high degree of engagement at each stage of the product life cycle, and the user group that often implements the purchase behavior is considered a high-value user group.
When analyzing user engagement and the behavior of different user groups, we can not only identify high-value users, but also identify users who are on the edge of loss. In other words, we can also predict the loss of users as we determine the high value users.

6. Incentive

There are many ways to motivate users, such as preferential activities, points exchange, etc. In incentives to retain on the verge of losing users or recall has been lost, we must ensure that these measures cost consumed less than these users to our contribution to profits. We can’t afford to waste a lot of human resources to save those who won’t contribute a dime.

7. Target users.

In As A Product Manager, Do You Really Understand the Data Analysis?,We have a product vision – the product must be targeted. After all, the target user is barking up the wrong tree, even if we make the most of it and we can’t let the user stay. If we use words like “free” and “cheap” to attract new users, our new users may not be our target users at all. Those who collect freebies are the most likely to lose. Our target users should be users who value the long-term value of our products, not those who are greedy and cheap.

8. To provide better service to users.

Poor service can lead to loss of users. There are two main reasons for the loss of users: one is incompetent and rude staff, and the other is that the service is too slow to be tolerated. More than 70 percent of the lost users are lost to bad services.

The image above shows the importance of user service. Any service that makes users dissatisfied can lead to loss of users.

9. Value user complaints

The product problem exposed by user complaints is only the tip of the iceberg. According to the survey, 96% of users will not be satisfied even if they are dissatisfied with the product, and 91% of them will leave without a word. Only 4% of users will complain about the product!

Therefore, we must take users’ complaints and complaints seriously and give timely feedback. Research shows that users who receive feedback and solutions are more likely to become loyal users, who can spread our products or services to form a good reputation.

10. Show our competitive advantage.

We need to make it clear to the user that we are different from our competitors; Where we stand out; If users don’t use our products, they will miss something. Knowing these problems, we can see what our competitive advantage is. After knowing the competitive advantage, we can hype it up!
We’ve talked about 10 strategies to reduce user loss, and we want to help the people who run the APP.

How to Improve the User Churn Rate ? Here are 10 Strategies.,首發于Cobub

]]>
Why Can’t You Help Clicking It? http://www.dpkxx.com/en/why-cant-you-help-clicking-it/ Fri, 10 Nov 2017 09:26:00 +0000 http://www.dpkxx.com/?p=7138 According to the statistics, 79 percent of smartphone users will check their phones within 15 minutes of waking up in the morning.A 2011 study by a university found that the average person watches 34 times a day.However, recent figures from the industry have been much higher - almost 150 times.We have to admit, we're addicted.Faced with this high-tech product, we're not addicted to it, and we're at least obsessive-compulsive.We can't wait to see WeChat, weibo, access to mobile phone taobao, jingdong, was just going to have a crush on for a few minutes, an hour to find their own fingers slide is still on the phone screen page.This desire may be with us all day, but it is seldom perceived.

Why Can’t You Help Clicking It?,首發于Cobub

]]>
According to the statistics, 79 percent of smartphone users will check their phones within 15 minutes of waking up in the morning.A 2011 study by a university found that the average person watches 34 times a day.However, recent figures from the industry have been much higher – almost 150 times.We have to admit, we’re addicted.Faced with this high-tech product, we’re not addicted to it, and we’re at least obsessive-compulsive.We can’t wait to see WeChat, weibo, access to mobile phone taobao, jingdong, was just going to have a crush on for a few minutes, an hour to find their own fingers slide is still on the phone screen page.This desire may be with us all day, but it is seldom perceived.
How exactly is this habit developed?
Why do we habitually open an App?
Why do some products stop us from losing our addiction to other products?
Is there any secret to making users more comfortable with your product?

According to cognitive psychologists, the so-called habit is a kind of “unconscious behavior that is produced in the context of the situation”, which we make almost without thinking about it.These days, the products and services we take for granted are changing our every move, and that’s what the product designers were trying to do.That is to say, our behavior has been designed unconsciously.
How do these companies manage to influence users’ habits and control their thinking with just a few encoded characters on their electronic screens?What are the factors that make people want these products?
Making users develop habits and dependencies is an essential element of many products.Nowadays, more and more enterprises have realized soberly, with only has a huge customer base is not enough to constitute a competitive advantage, the user the product a high degree of dependency is the key to decide its economic value.
So, are you ready to learn more about cultivating positive user habits?And then, looking down, you’re going to get a whole new understanding of the addiction model.

The four stages of the addiction model – trigger, action, variable reward, commitment

Trigger: remind people to take the next step

Triggering is the first stage of the addictive model, which triggers the user to trigger action.There are two types of triggers: external triggers and internal triggers.Let you produce habitual rely on those products are often the external trigger the first to play a role, it is through the information penetration in all aspects of your life to guide you to take the next step, such as email, website links, or the application icon on the phone.
Use an external trigger is just the first step, trigger is the inner core, it through the user memory storage in various association to remind them to take the next step, negative emotions can often act as internal trigger.Develop habits to develop product designers need to consider the user’s psychology, to understand those likely to be internal trigger emotions, and to know how to use external trigger to help users to take action.

Action: the direct response of people when they expect a reward

The trigger is action.If they fail to act, the trigger fails to take effect.Dr Fogg, of Stanford university, argues that to get people to act, three elements are essential: motivation (M), ability (A), triggering (T).I’ll write it in terms of the formula, which is B= MAT.
Triggers remind you to take action, and motivation determines whether you are willing to take action.The motivation for users to use products is based on people’s pursuit of happiness and the avoidance of pain.The pursuit of hope, the avoidance of fear.And because people want to be identified, they hate being rejected.So, as long as your product can give users happiness, hope and identity, it can be the motivation for user action.
With an inner itch (trigger), there is a willingness to tickle, and users can easily “scratch”.The ease of use of the product directly affects the user’s usage of the product.To successfully simplify a product, we have to clear the way for users to use the process.Fogg, summarizes the influence degree of task difficulty of six elements: time required to complete this activity, economic investment, physical, mental, and others on acceptance of this activity, and the degree of match between the activity and routine activities or contradiction.When designing the product, figure out what is preventing the user from completing the activity.Is the user no time, or no money?Is it a busy day and don’t want to think again, or is the product too difficult to operate?To win hearts, you have to make your product easy and easy to navigate.
Therefore, to increase the incidence of preconceived behaviors, the trigger should be obvious, the behavior should be easy to implement, and the motivation should be reasonable.

Changeful reward: satisfy the user’s demand, arouse the desire to use

In the third stage, your product will be able to satisfy users’ needs and arouse their desire to use more.What drives the user to take action is not the reward itself, but the desperate need to be rewarded.The difference between an addiction model and a normal feedback loop is that it can inspire a strong desire for something.The feedback loops around us are not uncommon, but the feedback loops that can be seen in the results do not help to create the inner desire.
The variable reward of “installation” of products is a magic weapon used by the company to attract users.Fundamentally, variable rewards must meet their needs while appealing to users.There is often more than one reward for products or services that can kill users in seconds.Products that don’t have the advantage of being in a variety of variability have to be constantly updated to keep up with The Times.

Social raise reward

The interpersonal rewards obtained through interaction with others.For example, when a friend gets married, he sends out a circle of friends and receives a wave of blessings, which is a social reward.We like our “circle”, enjoy the “likes” of others, and expect “comments” from others.Social rewards can make users crave and expect more.

Prey to reward

Get resources or information.Such as microblogging, microblogging began to attract people, because people as long as the repetition of a “rolling” behavior, can search to the interesting information that oneself like, this is hunting reward mechanism, content variety provides users with unpredictable attractive experience.

Reward yourself

Experience the sense of manipulation, the sense of accomplishment, and the sense of closure.The “upgrade” in the game affects the evaluation of mastery and ability of the self, and the game rules of upgrading and acquiring privileges can satisfy the players’ desire to prove their own strength.Even mundane emails, the number of unread emails in mailboxes are like tasks for many people, a task that needs to be done.
The variable reward is a powerful tool to attract users.Insight into why people habitually rely on products helps designers to design their products well.

Input: through the user’s input to the product, cultivate “repeat customers”

But it’s not enough to get the product to the user’s mind.
A one-night red product often has a good trigger, an easy operation, and a rich social reward.But without the ability to “invest” in a long time, hot style will lose users’ attention over time.
It turns out that the more we invest in things, the more likely we are to think of it as valuable, and the more likely we are to be consistent with our past actions.Finally, we change our preferences to avoid cognitive dissonance.This is the last stage of the addiction model, and it is a stage where users have to invest.The input phase is related to the customer’s expectation of long-term reward, which has nothing to do with timely satisfaction.
When a user provides their personal data and social capital for a product, spending their time, energy, and money is what has happened.In other words, investing doesn’t mean letting users spend money; it means the behavior of the user can improve the quality of subsequent service.Add attention, add to the collection, expand the virtual assets, understand the new product functions, all these are the input that users pay to improve the product experience.These inputs have an effect on the first three stages of the addiction model, and triggers are more likely to form, and action is more likely to occur, and reward is more attractive.
You must say, everything has a routine, then know the routine, how to reverse the pattern?
As a product manager, you can use the addiction model to compare your products:
What do users really need?What pain can your product alleviate?(internal trigger)
What attract users to your product?(external trigger)
What is the simplest operation a user can take when expecting a reward?How do I simplify this operation?(action)
Are users satisfied with the reward, or do they want more reward?(variable reward)
What are the “drips” that users make to your product?Do these inputs help to load the next trigger to improve product quality during use?(input)
If you are a user and learn about these “design patterns”, you can take targeted “anti-design, anti-pattern”.Check your daily actions: what situations do you automatically run on an App?When is the easiest time to swipe your phone?What are the driving forces behind it, to kill time, or to relieve stress?Make a note of it.Find your own behavioral patterns and intrinsic drivers to consciously control your behavior.

文/ 小歐 微信公眾號:中歐國際工商學院
本文改編自《上癮:讓用戶養成使用習慣的四大產品邏輯》一書

Why Can’t You Help Clicking It?,首發于Cobub

]]>
The 5W Rule for Perfect Push Messages http://www.dpkxx.com/en/the-5w-rule-for-perfect-push-messages/ Thu, 02 Nov 2017 09:22:42 +0000 http://www.dpkxx.com/?p=7103 APP operators know the importance of push messages to improve user engagement.The good news is that the user engagement will improve significantly, and the number of users will be greatly reduced.
How do we get the push messages to the desired effect?APP operators must pay close attention to the behavior of users, have a clear preference for users' interests, and push the content of interest to different groups of users at the right time.The perfect push messages are valuable to the user, which can help the product to improve the user experience and increase the user's liking.
Who: Object
What: Content
When: Time
Where: Situation
Why: Reason

The 5W Rule for Perfect Push Messages,首發于Cobub

]]>
APP operators know the importance of push messages to improve user engagement.The good news is that the user engagement will improve significantly, and the number of users will be greatly reduced.
How do we get the push messages to the desired effect?APP operators must pay close attention to the behavior of users, have a clear preference for users’ interests, and push the content of interest to different groups of users at the right time.The perfect push messages are valuable to the user, which can help the product to improve the user experience and increase the user’s liking.
Who: Object
What: Content
When: Time
Where: Situation
Why: Reason

Figure out these 5 Ws, then we can push messages satisfying users’ need.

1. Why: Why do we push this message?

“Why” always comes first, with every push message having a clear goal — to make the user login for the first time or for the user to upgrade?Not only do we need to know the target of the push message, but we need to know what behavioral data we want from the user.These behavioral data are used to measure the impact of push messages on users.For example, after the user logs on to a social APP for the first time, it will receive the following information, which is the push we made to get users to improve their personal data.

2. What:What do we push?

The content of the push messages have the following three characteristics:

(1)Targeted:

Sometimes small details can play a big role.In the example above, we added the user’s name before the message was sent to the user. “Hi, XXX, welcome…”Instead of sweeping to the point: “welcome…Plus the name makes the user feel more approachable.

(2)Ensure relevance and timeliness:

Push messages must be timely and highly relevant to users.Netease news, for example, is pushed to the user’s nanjing tomorrow weather news, reflects the correlation between geographical position, this message is to push on Friday afternoon, Saturday’s travel plans to give the reference for the user to remind, the right time.

(3)Accurate and rapid direct user pain points:

The user’s time is precious, and the message we push must get the user to the maximum value.Do not push spam messages to users, or do not match the user’s needs.For example, messages for users who don’t have a car.

3. Who: Who are we going to push the news?

The push message object cannot be cut and cut. We need to divide and manage the user through user behavior data.When it comes to user behavior, here are a few more common user behavior analysis platforms, such as friends union, baidu statistics, etc.But these SaaS platforms also have problems, and the raw data is difficult to export, and the data of their own products cannot be owned by us.It is also easy to recommend Cobub Razor, data private and backstage construction.
Different push messages receive different recipients.
We provide personalized news feeds based on the user’s behavior habits, preferences and so on.For example some music APP users to listen to music for different style, favorite star and collection of song single push related update alert, such precision push has greatly raised the proportion of the user to open the message.

4. When do we push the news?

(1)Time of day:

Can select the user free time, such as before going to work in the morning, at noon to eat, after dinner, specific choice which push period, can according to user’s use time was used to determine, in short not to interrupt the user when user busy or rest.

A time period for users of a financial app

(2)Push frequency

Too high push frequency can result in two kinds of results — the user clicks on the push and immediately turns off, or the user simply ignores the message and doesn’t even look at it.The number of push frequency should be determined according to the type of application, generally speaking, the social App can be pushed daily, the information class can be 3-4 times a week, the tool class 1-2 times a week.The number of times should not be too much, otherwise the user will not open up, and probably close the message push, even uninstall the application.

(3)Local time:

Push messages need to be determined according to the user’s local time, this needs us according to the user’s location to determine if your users from all over the world, Beijing time at four o ‘clock in the afternoon when Washington is at four o ‘clock in the morning.If we unified Beijing time to push the news, users in Washington would be disturbed.

5. Where: What situation are we pushing in?

Push scenes can also affect the effect of push messages.We need to consider where the user is when the push message arrives, what he is doing, or what device the user is using when he receives a push message.
We, as users, receive a lot of notifications every day. For example, the taxi app will push express coupons for me on Friday after work.The group purchase app will push the group purchase information of the restaurant near me at mealtimes.If the user is in a disturbed state when they receive a push message, then the user will not be able to look at the push message and take action that we expect.

The messages in the application are pushed based on user behavior, so the effect is better.Regardless of the user, lie sofa eating, no matter the user on the PC, mobile end, as long as we clearly understand the user behavior, according to their behavior push related news, the user is there is a strong possibility we expect action.

Conclusion

The perfect push message must be pre-defined by 5W.Through 5W, we make clear the target, content, object, push time and scene of the push message, so that the user can see the value of the push message to the user.Pushing is the best user touchpoint, and it’s more important to make a thoughtful product that fits the user’s habit than it does daily.

The 5W Rule for Perfect Push Messages,首發于Cobub

]]>
The Implementation of Cobub’s Codeless Capture Technology http://www.dpkxx.com/en/the-implementation-of-cobubs-codeless-capture-technology/ Wed, 01 Nov 2017 08:36:28 +0000 http://www.dpkxx.com/?p=7091 With the advent of the era of big data, data mining has become more and more important.Front end point buried as a more mature data access method is widely used.Currently buried point is divided into two ways, and will come with no points.Sets a buried point is easy to understand, is called the SDK API, in the code inserts buried point related code, user behavior acquisition.Because we are in development projects, buried point are manually, every business needs change to buried point everywhere, and no burial code, which does not need to be manually inserted into the code, just prior to related configuration, the SDK automatically collect user behavior, avoided because of the change of demand, and buried the great degree error causes such as to bury some heavy and complicated work.This paper mainly introduces the technical implementation of codeless capture technology.

The Implementation of Cobub’s Codeless Capture Technology,首發于Cobub

]]>
With the advent of the era of big data, data mining has become more and more important.Front end point buried as a more mature data access method is widely used.Currently buried point is divided into two ways, and will come with no points.Sets a buried point is easy to understand, is called the SDK API, in the code inserts buried point related code, user behavior acquisition.Because we are in development projects, buried point are manually, every business needs change to buried point everywhere, and no burial code, which does not need to be manually inserted into the code, just prior to related configuration, the SDK automatically collect user behavior, avoided because of the change of demand, and buried the great degree error causes such as to bury some heavy and complicated work.This paper mainly introduces the technical implementation of codeless capture technology.

The Implementation Process of Codeless Capture


1. Visual view circle, the page will appear floating circle, drag the circle to want to configure the control of events, play box will pop up input event.
2. In the previous step play box type in the name of the event the custom, the name will and view viewPath bind.ViewPath is the unique identifier view, in the heart of the below in detail.
3. The user clicks on the controls, whether controls binding events, such as binding events to upload.

The Technical Points in the Process

The Implementation of Visual View Selection

Custom UIWindow subclass, as suspended small circle, add UIPanGestureRecognizer gestures, according to the displacement of hand gestures, to set the displacement of suspension box.Gestures stops for floating window center coordinates.
Traverse the child views on the main window, find the window containing the suspension center and can respond to user interaction of the layer view, is the user can select the view.
Refer to iOS control message transmission chain, a core method.UIView hitTest: (CGPoint) point withEvent: UIEvent *) event.This API automatically traverse child views, find a point of view, the event nil.Since the event parameters is nil, eventually found the view is not necessarily can respond to user gestures view, if you can’t response to traverse the father view, until you find the view can respond to user behavior.

Selecting the View Binding Events

View a unique identifier viewPath generated, the above steps have been given a select the view.How do you determine the view viewPath is key.ViewPath need the whole application only, just can be the difference between different events.Because it is no, so can attribute to analysis from the view itself.We can put the App’s understanding of the concept of spanning tree view structure, root node of the tree is UIWindow, one of the branches of the tree is composed of UIViewController and UIView, leaf nodes are UIView.Then from the root node to leaf node path can be seen as the only one.Also is the viewPath view.Achieve logic, introduces below viewPath consists of two parts, the first part is a node path, another part of the matching node index.Node path is made up of each node Class patchwork, node index, index in a node in the parent node, such as child views in the parent view’s subviews array subscript.This is the logic diagram of the traverse the nodes.

Index of computing nodes, this step, there is a special view to note, reusable view index is associated with a data source, such as UITableViewCell, such a view of the index cannot use the parent view’s subviews subscript instead, you should use subscript representative of data sources, such as cell indexPath. Section: indexPath. Row.Here is a simple view and reusable view viewPath example.TestViewController UIView – UIButton & 0-0-0 and TableViewController UITableView – UITableViewCell & 0-0-1-0.
How to detect the user triggers the binding event ID view is also the point, the use of core technology is the Method the runtime Swizzle.To introduce the following according to different types of controls, how to hook the corresponding method.
1. The UIControl type control hook – (void) sendAction (SEL) to: (id) target forEvent: UIEvent *) event
2. UIScrollView, UITextView UITableView, UICollectionView types of controls, hook – (void) setDelegate: (id first) the delegate method, and then the hooks to collect event agent method, such as textViewDidBeginEditing, tableview: (UITableView *) tableview didSelectRowAtIndexPath: (indexPath NSIndexPath *), etc.
3. View the hooks with gesture events – (void) addGestureRecognizer method, and add new objects to gesture in the method to realize the target and action, – (void) addTarget (id) target action (SEL) action.

Conclusion

Codeless capture technology is the analysis of the above points, first of all, through visual select the view to need binding events, and generate a unique identifier viewPath, through the hook system control method, get the user’s view, the views generated viewPath list with local events, events will upload viewPath corresponding than success.

The Implementation of Cobub’s Codeless Capture Technology,首發于Cobub

]]>
How to Acquire New Users with Lower Costs? http://www.dpkxx.com/en/how-to-acquire-new-users-with-lower-costs/ Mon, 30 Oct 2017 02:20:01 +0000 http://www.dpkxx.com/?p=7045 Emerging industries, like sharing a bike or like payment of the ancient and the industry, the new is always going on, new, retained, active, transformation, erosion, activation is operation of the whole process, such as new is the first step, the new do bad, the product is lack of the soul, the soul is the user and no user, to fine the perfect products are meaningless.New is so important, as a new era of qualified operation, how should we pull new users?

How to Acquire New Users with Lower Costs?,首發于Cobub

]]>
Emerging industries, like sharing a bike or like payment of the ancient and the industry, the new is always going on, new, retained, active, transformation, erosion, activation is operation of the whole process, such as new is the first step, the new do bad, the product is lack of the soul, the soul is the user and no user, to fine the perfect products are meaningless.New is so important, as a new era of qualified operation, how should we pull new users?
Outline of this paper is illustrated below:

A touch, precision up to the user, the exposure is large enough

1. The advertising precision


Part of the listed companies advertising expenses (data source results)
Product is completely no advertising, from the table we can see that even have won 15 international gold prize of 1915 China’s national drink moutai also need large amount of advertising.According to Wind information data organization, according to data of 2014 listed companies in China a total of 65.929 billion yuan worth of advertising promotion cost, even the industry giant AD is also in constant, increase the user’s perception of the brand, we have what reason to expect no money input can be a guest?

Advertising is essential to promote contact means, advertising are accurate enough, blind pursuit of big traffic platform as part of public money to throw distance platform for the small and medium-sized, frequency, cycle, small test results and accuracy and hard.

Here lists online, by the way of the mobile Internet traffic on the channel (incomplete)
Decision must be tested before big advertising, not after drop test, never know one stage which the flow of a platform is the high quality, the most suitable products.The direction of the big, big platform flow and spread, more suitable for search class advertisements;Vertical class platform also to consider the industry each competing goods delivery situation, on a vertical platform and a number of conversion were rare.

In addition to the online and offline advertising, advertising metro bus station, district school board, medium rolling screen, mall office building elevator and window, etc., can be the choice of offline advertising.

Case study:

In this play amount exceeds 2 billion douban 9.1 distracted in white night after fierce, standing and love money into two P2P creative advertising with the plot development, is said to be a price is 1 million, if is to pull a new guest, the advertising is not precise, of course if it is in order to expand the brand influence is understandable.

In contrast, another financial companies advertising (to brand name) :

Advertising objective: power, an online financial low cost access to high quality customers
Advertising background: an online finance is an Internet comprehensive service platform, with excellent and professional service ability, comprehensive strength in numerous financial wealth management services in the best posture.At the same time, as the Internet blossoming in the financial services agency, the rise of consumer finance financial guest cost rising, so how to accurately for the guest, low-cost new is that they desperately needed.
Target: effect of guest oriented, ROI more than the industry average.
Delivery period: March 2017 – June 2017
Financial management target groups: 20-35 years old, bachelor degree and above in group of high quality business people, crowd of finance and economics, finance, young consumers.70% of men, women accounted for 30%
Delivery types: finance, news, web portal, PC;Automobile, IT, education APP side.
On the implementation: through data analysis, they found that the target audience to pay attention to the weather every day, lest delay flights and choose to drive or take a taxi to go out;At the same time, they focus on news of finance and economics dynamic, care about their wealth accumulated depreciation and investment opportunities;Among them, the male money users generally focus on sports, sports news, have a favorite sports star and the team, and are willing to go to the scene to see a ball game.To via “weather”, “sohu news, tencent news”, “sina sports channel” to try, the results showed that the financial news of user intention is very strong, and then to financial news website and APP advertising.
In conclusion, advertising to analysis the behavior of the target user, now almost every Internet user is transparent, find some user behavior and its platform, the correlation of the test, according to the effect of increasing drop ratio, and control the cost of the new guest as possible.

2. Artificial precision touch up to push forward (online and offline)

Don’t really have money for advertising?That can only be human flesh pull customers, of course, AD is rich, artificial way also is very good, can more direct understanding of their products in the market reaction.If precision is a guest, not brand shaping, artificial forward and to push, should let the contact information of target users, rather than the cast the net wide and full.

Parking APP promotion, for example, the famous American traffic information data company Parking INRIX is among the first to launch the software, it is a Parking assistance service phone APP – “INRIX On – Street Parking”, they offline promotion scenarios have a “Street intersection”, “village office Parking lot”, “brand auto repair shops and stores”, while On the road to promote only choose private car, rather than a bus and taxi.As much as possible to the parking lot talking to property on the parking lever to promote cooperation, if not to push in to the parking lot during the rush hours.

To sum up, start doing promotion will be thinking about how to be more rapid, more accurate for the guest.

3. Spontaneous for users to share

In the new users at the same time, don’t forget the old users to the new role.Users can share, spontaneous reward can also be affected by platform to share, means, though different, but the result is the product to share with people around you.

On the users to share this topic, the activity box running clubs a “humble one activity, but it can achieve user viral growth” in a case I feel very fit, quote here, good ideas worth repeated exposure.

Users in the process of using the APP, meet interesting things, will share it.Worship at the bike just to market when using, we have to try.At that time, often in WeChat friends see friends to share your trip, or even a friend sent than the number of travel, riding time circle of friends.The sharing behavior of users spontaneity, in a short period of time for the bike has brought a lot of users.

Users find products is very good, the spontaneous products to share with friends.Product with payment for knowledge is the APP and NuMa APP on the sharing of product design and have such different, inside get APP, you can see the “recommended” [get] to friends this feature, users can according to their own requirements, generate posters, share with their friends, or is a key to directly share to their friends.And NuMa APP without this feature.

Share button in addition to the APP itself, users will also be spontaneous recommended in the chat, said didn’t stick to force others to read, for example, I would recommend “netease snail reading to her,” keep reading every day at least 1 hour, 1 hour later have no you want to read permissions (unless published a book review, if she published a book review, she gradually habit even fell in love with reading it).

Also there will be the UGC in the APP is very good, a picture, a sentence, an article is very touching, users will spontaneously to share.Such as netease cloud music, white space, everyone is the product manager, etc.

In addition to the users to share, spontaneous and induced by interests users to share, such as the product of spell group mode, Meituan hungry? A red envelope, drops Uber bus ticket incentive mode, etc., basically all consumer, financial APP users share incentive mechanism.

4. Different industry flow swap

Traffic exchange platform between cannot have direct conflict of interest, and to have the certain user overlap.In the absence of expense cost, take platform has its own traffic and conform to the standard platform to exchange flow, look for the best personal connection to the head company, with their cooperation, to talk about the flow of other home will have a better chance of cooperation.

Refer to baidu encyclopedia (the world’s largest Chinese encyclopedia) explanation for swap “flow”, which has described concepts and technical way of the cooperation of very clear, there is no need to do.
Case study:
Remember qihoo 360 Lou easily share their traffic exchange case, citing “Lou loosely blog” video search products business case: at first I did a lot of preparation work, want to how can to promote the flow of the product.First of all, I some of the traffic data analysis before cooperation.Put some effect is very poor cooperation stopped, only keep two flow effect is done, so I will take a lot of traffic to do swap BD position.

First of all, I would like to collect a lot of traffic exchange cooperation is willing to do the entertainment sites contact, and then a a to talk about, our product has less traffic at the beginning, since less change.But light on BD position, will never be able to transfer flow rate higher.Because access to video search users, mainly come in search of the video, can click cooperation links to sites on the BD position will not too much.

Then I thought of video search the features of this product, visit our web site users will eventually run into some video sharing website, such as “youku” “potato” cool “6” and other sites.Why do I want to go to their flow, should let these sites also to our BD, return to our traffic.

Then I put ALEXA1000 name at the time of video sharing site to talk about it again.At first two private relations and good video sharing site agreed cooperation.

Is better to talk with the back of the case, I’ll tell some video sharing sites, so-and-so video website also cooperation with us, and when a lot of video sharing sites also have do and some entertainment website traffic exchange cooperation.We are video search traffic more targeted in the past, if entertainment website can cooperation, it should be more cooperative with us.

When the video sharing site cooperation policy is you give me 1000 IP, I will return you 2000 IP.The more traffic in cooperation with us, they are more cost-effective.So when ALEXA1000 within the video sharing site nearly eight all have cooperation with us, and several of them were tens of thousands of IP traffic exchange, such as “youku”, “6”, “popcorn”, “mofile” and “mating” is the main partner.

Was the same kind of video search site also has several, such as “love asked video search” “yahoo video search” video “baidu search” because they haven’t asked the video website return flow, so I went to talk to the video website traffic exchange cooperation, has the certain difficulty.So I through the video search products features, give A video website over flow guide to video website, B B video website traffic guide to C video website.Cooperation with video sharing site, traffic.Flow rate of the exported to BD position also more and more, I also with some entertainment website also increased the swap.

Do remember the highest peak, one day there are hundreds of WPV, than when I took over turned over several times.And it is supported by no company resources to do, let me feel a sense of accomplishment at that time.On qihoo video search do promotion experience, has also been think is the most happy in my career for a period of time.

5. The media exposure

Tencent business friend once said:

“Many start-ups have no resources, voice is too weak.Sometimes, clearly project is good, but the spread of the articles are not attractive, not suitable for platform to report.”

I usually read product report more platform is a “dark horse”, “the tiger sniffing net” and “36 kr,” they reported each time all is around the story of the founders, thought to packaging feelings;By analyzing the good parts of the product or service differentiation, the sweet high standards of service, complementary advantages, founder of the team (in the industry has a strong professional ability, have a sharp insight to market) to expand coverage, etc.

Most of the time of media resources and it is not how much you pay, media platform also need high quality content material.They want those ideas, attitudes, dry entrepreneurs, they would express themselves about the industry, about the market, on the understanding of the user.So, good PR, first not ask what resources, but how do you think of PR, how to understand the people who help you do the PR.Otherwise, give you more resources, with no eggs.

6. Channel operation management

Because I am in the financial industry, the Internet, so it is understand that the assets of the financial sector is control in the hands of the mediation, financial platform in addition to the proprietary assets, but also cooperate with the mediation, thereby gaining assets, in this process, the platform can’t contact with real borrowers, contact only the mediation, industry present situation is so so, so, multi-channel operation management is quite important.

The so-called channel is provide business support for product transport users, partners.Financial platform to acquire assets (borrowers) and (investors) to get the funds, assets and intermediary cooperation in a wide range of industries, is to have them to provide various financial platform risk control into a standard customer;Also from the platform for various flow of funds, have long-term cooperation.

Constantly on a regular basis to analyze channel resources into effect, and to practical communication channels, defective rate is high to and channels of cooperation, in a timely manner to inform or to terminate its channels to prevent cash or brush single risk.

To attract users attention

1. The precise touch is the basis of the resonance caused demand

We always think about how to make the article more propagation force, how to give more people play the video H5 push, so large quantities of a variety of channels, but the effect may not be good, serious money spent, the effect not beautiful, so a lot of under pressure.Back to reflect on, we did not push the high quality content to the right people, with net fishing in the sea than to take a small mesh plug small fish touch xikou.

2. The impact of the visual or graphic feel

(1) the differentiation
Comparison of different things or to emphasize different aspects of the same things, and comparing, in order to produce a strong contrast.Contrast can be intuitive image, such as color, shape, texture, quantity, area, the direction of contrast;Can also be internal, such as properties, function, emotional appeal, connotation, characteristics, etc.Contrast to the elements, or produce the fierce conflict, or complement each other to be in harmony are an organic whole, constitute a strong difference teaching fruit.Contrast the creative purpose, is to highlight advertising, with a stimulus information affect the audience.

Case study:

(image from “creative design” oritive.com)
High contrast of cosmetic advertising: advertising scenario of graduated in 1969, a group of old man’s homecoming.Among the old woman thought the young woman in red face and figure, like their granddaughter, he showed confidence and a smile with the eyes of the other male students, female students embarrassing expression also contrasts, to fit the youth posture of the success of cosmetic surgery.

Contrast methods of the 2008 Olympic Games in the poster: the first is expression on the contrast, athletes is photography, the crowd is drawing techniques;Followed by color contrast, the athlete is color, is black and white people;Finally, the composition of a picture on the contrast, the athlete with the blank as the background, while the crowd numerous, very compact.The contrast condensed into a huge screen tension, the tension in the Beijing Olympic Games draws near, and ten thousand people shocked the aura of incisively and vividly express in my heart.

(2) zoom in selling point
The following several advertisements do you think of “soil”?

Malatang businesses: “mountain spring hot food, one hour in a pot, malatang also can be so healthy”

Native cafe: “flagship” let you eat to 40 years ago, the taste of chicken”

Organic tea merchants: “other people’s tea drink NongYaoWei, the organic tea to drink only fragrance” in our house

We eat malatang is worried repeat with soup pot, one with even more than a day?Native is missing under the modern feed for aquaculture has not used to the taste of the early?For black tea, poking fun at “bitter” a NongYaoWei?Our pain points is the selling point of merchants, the merchants selling point directly root of publicity, the customer first whether flocculant head, will at least to try.

Case study:

OPPO R11: “before and after 20 million, take pictures more clearly.”
Candy phone F11: “50 million pixels, than the more clear and more clear.”
Oppo direct selling point “more clear”, and enlarge the selling point “clear” candy “than more clear, clear, and specifically on the advertising of the oppo, both the interest, and has a strong selling point into effect, this completely easing the bit in competing goods ads, clothing!

(2) the connotation of the creative
We have seen a lot of creative advertising design, have to say, some slightly the connotation of the exotic advertising instead more impressive.In advertising design industry, creative seems to be spring never dried up, all sorts of funny, surprising, exaggerated, the connotation of creative ads emerge in endlessly.

Case study:

Your hand sanitizer: “one day what are exposed to, finally you do with your hands?Consider very fear.So, you will choose not to recall, or choose Siribuncha instant hand sanitizer?”.

NBA players: “sports need passion, need like animals are aggressive”.

Dry cleaner: “whatever he is lip seal, oil or wine stain, all get rid of all”.

3. Strong direct user demand, and puts forward the solution

Stamp in the user’s pain points is not enough, more simple and direct offer solutions, such both neither are hurting users feelings, and can drive the new product.


Double tenth a JPM medicine in 2016 with its logo and the advantages of the high recognition, deep recognition, advertising, with the attitude of a clever into its poster, can not only improve the visibility, and remarkable success.Pure and fresh style, warm words, let a person feel anything disgust, instead of multiplication.It not only box out of the users in the hand chop and the accident harm may occur in the daily life, also can in time out of their delivery to your door “20 minutes” solutions, is a sweet little quilted jacket.

4. Bold promises, eliminate user finally

Users never know “you” to want to know you, is a big step, she needs to break his heart security defense, this step will often turnover rate is higher.Chasing girls like you, the first meeting about what cinema or sparsely populated mountain, who will trust your own security to you?

We all said taobao fake goods, is the seller, return have to pay the freight, laborious also thankless, give a bad review and even cause the seller harassment, but there are many sellers in efforts to promote trust, use safeguard mechanism to reduce the order cost, make products and logistics cooperation at the same time bold promises, don’t believe you let users to trust their products?The gap between that began with a number of shops.

Light user demand, actively seek products

1. To provide access to, step up to

Careless mistakes never lost jingzhou made, very not easy to use content, advertising, attracted by any way users, but did not provide download the entry of the product, or make entrance can’t see don’t watch carefully, that this is a new, just made a brand publicity.

2. Reduce the user the opportunity cost


A user choose to do and not do an operation is a kind of decision-making behavior.In the perspective of economics, the decision-making behavior of “cost” is the opportunity cost.Users get products behavior brought about by the “experience is good”, “information security”, “will you won’t get involved in traffic”, “whether it takes a lot of time”, “will not work”, etc.Some we can be resolved, some are not solve, we reduce the user’s opportunity cost as much as possible, let the expected return is greater than the opportunity cost, the user access to products of this step loss can be reduced.

3. Participate in steps can decrease, it decreases


Pull the new process, to prevent the user when registered products deserted it, so make the registration procedure can be simple, simple, and even do tourists mode, don’t have to log in to enter the product of a few interface, to limit the function of the most characteristic after registration to use, to reduce turnover registration.

Case study:

WeChat reading is tencent products, have developed a “trial”, because there are a few people have this requirement, is they don’t want what WeChat friends to see their reading a book, or are not willing to take WeChat to land any products other than WeChat APP, maximum efficiency to optimize the attrition rate of the user login interface.Tencent did so in their own product system, we have what reason to let the user must log in?

4. With a hammer “affordable” limit use to conclude business

If the user has not registered, the couple’s exclusive discount and new methods, and if the user has already registered products, has been receiving limited-time discount can give users a sense of urgency of time, an impulse, which occurred some spending or transformation behavior (such as WeChat reading the length of the money to buy books, etc.).

New exclusive activity is almost every product standard, to do the page more impact, such as Vipshop preferential payment pages and pages of the countdown, can produce time sense of urgency, in the case of not fully considered just finished the action.Users also have the feelings of people, impulsive, also easy to change into rational conservative thinking, consider more long more conducive to action of product.
Article source: Everyone is a product manager author: wang liang

How to Acquire New Users with Lower Costs?,首發于Cobub

]]>
As a Product Manager, Do You Really Know about Data Analysis? http://www.dpkxx.com/en/as_a_product_manager_do_you_really_know_about_data_analysis/ Mon, 23 Oct 2017 07:20:54 +0000 http://www.dpkxx.com/?p=7022
Every product manager knows that data analysis is important, but can you give a clear answer to both of these questions?
1. What is data analysis?
2. Why is data analysis so important?
It doesn’t matter if you don’t know the answer, because this article is about to answer it from the following aspects:
1. What is data analysis?
2. Relevant concepts of data analysis
3. How to conduct data analysis?
4. How to measure and collect data?
5. How to do data analysis report?
6. The relationship between data analysis and products

As a Product Manager, Do You Really Know about Data Analysis?,首發于Cobub

]]>

Every product manager knows that data analysis is important, but can you give a clear answer to both of these questions?
1. What is data analysis?
2. Why is data analysis so important?
It doesn’t matter if you don’t know the answer, because this article is about to answer it from the following aspects:
1. What is data analysis?
2. Relevant concepts of data analysis
3. How to conduct data analysis?
4. How to measure and collect data?
5. How to do data analysis report?
6. The relationship between data analysis and products

What is data analysis?

In short, data analysis represents product status, user behavior, and what users click on and so on. Although the data represents the state of the product, it does not indicate the cause of the state of the product. Data analysis cannot rely solely on a single measurement data, which should be based on a series of converging data.
For example, if you want to analyze the state of an object, we can’t use temperature only, you have to combine other data like position, velocity, composition, environment temperature etc. to implement data analysis. Suppose the velocity is 0, the object is 1 meter above the ground, and the surrounding temperature is the same as the object, and you can conclude that the object is static.
Similarly, when you analyze product status and user behavior, the more metrics you gather, the better for you.

Relevant concepts of data analysis

To get the most out of data analysis, you need to be very knowledgeable about the concepts of data analysis. These concepts include:
? Data points
? Segmentation
? Funnel
? Cohorts

Data points

Data points, the single point of data. A data point measures a particular project, including measurement data and time.

Accurate data points are the premise of our chart of product development trends.

Segmentation

Segmentation is based on user’s common characteristics and product usage pattern.

Segmentation is based on but not limited to:
? Technology (browser, operating system, equipment, etc.)
? Behavior (first time visit, return visit, etc.)
? Demographic (language, country, etc.)
If you do not do segmentation by the above characteristics, you have to do it by the ones that can be measured. For example, user gender is a feature that can be measured. As long as we add gender to the user profile, we can collect relevant data.
You can understand the user’s potential behavior patterns through segmentation. Average data masks these patterns of potential behavior. For example, you add first visits vs return visits as the segmentation feature, although the average page views is 2, you find that for first visits, the average page views is 1.2, the average page views of return visits was 3.4. Without segmentation, the difference in page views of first-time visitors and return visitors will be covered by the average page views.
With segmentation, you can focus on the main target user groups. For example, if your target users are located in eastern China, you only need to analyze the user behavior in different provinces in eastern China instead of analyzing users from all around China to optimize the product to adapt to the targeted users’ needs.

Funnel

The funnel model is mainly used for daily data operation and data analysis, such as flow monitoring and product target transformation.

To achieve the goal, the user performs a series of actions. For example, on the e-commerce platform, users will perform the following operations to achieve the purpose of purchase:

Through the funnel model, we can find out where the user is losing and find out the reasons for the loss through the investigation and analysis.

Cohorts

Cohorts is similar to segmentation, the difference is that cohorts is to compare and analyze user behavior over time.

Cohorts is helpful for measuring the long-time value of users.
Cohorts can compare differently, for example, we can compare a week ago registered users and registered users, a month ago can also compare the registered user of a particular date. If we don’t have a group for a week ago and a month ago, the new user will disturb our analysis of the user behavior during these two periods. When compared to a user of a particular time period, we can measure the impact of a marketing campaign or a product’s function update on user behavior.

The figure above is a retention diagram based on user registration time. On October 8, the number of users remained significantly higher compared with other users. When we look at this data, we can explore what is causing the change in user retention.

How to conduct data analysis?

Product managers are exposed to massive amounts of data, so how do we perform data analysis? We need to make the following plan:
1. Define product vision
2. Define KPI that meets product vision
3. Definitions allow us to achieve the metrics of KPI
4. (Through user behavior logging) define the funnel that affects measurement
In order to make a better plan, we need to understand the relevant concepts in the plan.

Product vision

Product vision refers to product usage and target users. In short, “what problems does the product solve for users?” Without product vision, all of our next action is a waste of time.

KPI

KPI measures product performance. La new, retained, active, transmutation and so forth are the categories of KPIs. We can also use KPIs to set product development goals, such as increasing user registration by 20% or increasing the purchase conversion rate by 30%. KPIs should be suitable for the stage of the product. If we are just starting out, the main KPI is the user registration, not the user activity.

Metrics

Metrics are the means to achieve KPI. Metrics generally have conversion rate, purchase rate, and so on. By computing two or more data points, we can get metrics. At the same time, the change trend of measurement indicators is also the basis of product improvement.

Funnel

The important funnel will change the metric in some way. After establishing the product use process/user behavior log, we develop the relevant funnel model based on the metrics and user behavior. The funnel model produced by the registration rate and the conversion rate is not possible.

Getting data points

Obtaining measurable data points to achieve KPIs, calculate measurement indicators data, make funnels significant.

The plan is not static, we need to update the plan according to the product vision and KPI.

How to measure and collect data?

There are two methods: establishing an internal analysis system or relying on a third-party analysis system. The internal analysis system can be customized based on metrics. The downside is that we need to build and maintain resources individually.
External analysis systems, such as Google Analytics, Mixpanel, KISSmetrics, etc., are good options. Third-party analysis systems are easy to implement without wasting the resources needed to build and maintain them. Cobub Razor is a professional APP data statistical analysis tool in China, which supports privatization deployment. The data is flexible and safe. It is a good choice.

How to do data analysis report?

Usually we do data analysis reports by making comparison charts and trend charts.
A comparison chart shows a change in a metric between two point points, such as a metric that changed between last week and this week. It allows us to see whether there are large fluctuations in the metric between two time points.
A trend chart shows changes in a metric over a period of time, such as changes in a metric over the past month. It shows the direction of the measure, indicating whether the product is performing — better, worse, or not?
The report identifies the problem and then tries to answer “why does XX happen?” “Why does YY change?” These problems, we can optimize and improve the product.

The relationship between data analysis and products

We improve the product according to the data analysis results. If there is no data analysis, we can easily change the product blindly and make head decisions. If there is no data analysis, we cannot know the effect of product change. In the process of product development, we need to constantly conduct data analysis to ensure that our products are developing according to our expectations.
To ensure that products are in a leading position, product managers must know what data analysis is and the importance of data analysis. Hope this article can be helpful to the general product manager.

As a Product Manager, Do You Really Know about Data Analysis?,首發于Cobub

]]>
Actual Combat of Apache NiFi Processor http://www.dpkxx.com/en/actual-combat-of-apache-nifi-processor/ Mon, 16 Oct 2017 03:27:49 +0000 http://www.dpkxx.com/?p=7004 1 Introduction

What is Apache NiFi? NiFi's website explains: "an easy-to-use, powerful, reliable data processing and distribution system." Popular, namely the Apache NiFi is an easy to use, powerful and reliable data processing and distribution system, its designed for the data stream, it supports highly configurable indicator diagram of data routing, transformation and mediation logic system.
To NiFi can describe more clearly, through NiFi architecture to do a brief introduction to below, as shown in the figure below.

Actual Combat of Apache NiFi Processor,首發于Cobub

]]>
1 Introduction

What is Apache NiFi? NiFi’s website explains: “an easy-to-use, powerful, reliable data processing and distribution system.” Popular, namely the Apache NiFi is an easy to use, powerful and reliable data processing and distribution system, its designed for the data stream, it supports highly configurable indicator diagram of data routing, transformation and mediation logic system.
To NiFi can describe more clearly, through NiFi architecture to do a brief introduction to below, as shown in the figure below.

According to the official website of the individual components, do the translation:
? WebServer:The goal is to provide an HTTP command and control API.
? Flow Controller:This is the core of the operation, with Processor for processing unit, the expansion of the provided for running threads, extension in receiving the resource scheduling and management.
? Extensions:In other document describes various types of NiFi extension, Extensions is the key to expand in the JVM operation and execution.
? FlowFile Repository:FlowFile library is NiFi track record is active in the flow of a given current state of the stream file, its implementation is pluggable, located on the specified disk partitions the default method is a durable log before writing.
? Content Repository:The Content library’s role is the location of the actual Content bytes of a given stream file, and its implementation is pluggable. The default method is a relatively simple mechanism to store blocks of data in a file system.
? Provenance Repository:The Provenance repository is where all the source data is stored and supports pluggable. The default implementation is to use one or more physical disk volumes, which are indexed and searchable in each location event data.

2 The Introduction of NiFi Processer

Said in the previous section so much, mainly through NiFi NiFi architecture diagram introduced the basic concept, the concept is a Flow Controller is the core of NiFi, then the Flow Controller specific what is it? Flow Controller plays file communication processor roles, maintain a connection and management of multiple processors each Processer, Processer is actual processing unit. So, let us through NiFi UI see NiFi Processor contains?

By above knowable, the Processor contains various types of components, such as amazon, attributes, hadoop, etc., can be easily identified by the prefix, such as Get, Fetch beginning on behalf of the acquisition, such as getFile, getFTP, FetchHDFS, execute on behalf of the execution, such as the ExecuteSQL, ExecuteProcess, ExecuteFlumeSink can be easier to know its easy use.

3 Actual Combat of NiFi Processer

Having said so much, introducing NiFi’s architecture and Processor, what about the actual combat? Then, this article takes the author’s actual demand as an example, carries on the actual combat of the Processor. The requirements are as follows: Select a data processing scheduling tool to implement custom scheduling for server scripts. The script of the server involves scheduling of environment variables, oracle databases, and Hadoop ecosystem components. When the server script scheduler is completed, it returns the script run state and provides the failed re-run interface.
In order to achieve the requirements, ever scheduling scheduling tools, such as Apache Oozie, Azkaban, Pentaho, finally compares the various pros and cons of trying to use Apache NiFi as a try, by looking at NiFi Processor API, can better support Processor for ExecuteProcess remote operation. Below are in demand for actual combat.

3.1 Add and configure the Processor

1. Add and configure the Processor

2. Right-click on ExecuteProcess and select Configure Processor to Configure the Properties TAB. Each of these configuration options provides related instructions, as shown below.

As the figure above shows, there is a need to explain the options.
? Command: sh.
? Command Arguments:-c;ssh user@ip sh js/job/job_hourly.sh `date
? Batch Duration: Don’t set. // we need to schedule regularly, rather than at intervals.
? Redirect Error Stream: Don’t set.
? Argument Delimiter: ; / / to; Split the parameters.

3.2 Processor Dispatch

NiFi support three scheduling policies, including Time Driven (drive), CRON Driven (CRON) and Event Driven (Event Driven, not optional), according to the actual demand we choose CRON Driven, personal understanding of CRON is the application of Crontab, the parameters of the CRON meanings respectively: second, minute,, day, month, week, years, when the need to cooperate with *,? And L perform together (* representative works on the value of the field; ? Representative for the specified field is not specified value; L on behalf of the long plastic). For example, ” 0 0 13 * *?”The representative wants to have a dispatch at 1 PM every day. Therefore, the scheduling configuration of parameters is based on our requirements. This is shown in the figure below.

3.3 Operation State Monitoring

NiFi is available for developer scheduling through Rest apis, where we monitor the running state with the Processor API (state parameter acquisition, Processor startup and stop).
1. Operation status monitoring parameters:
Command is as follows: the curl ‘http://IP/nifi-api/processors/processorsID’, get the following results can be interpreted through json parser, and access to state.

2. Start and stop of Processor:
NiFi’s Processor startup stops with its Put method. The most effective action of Put is to change its operation state. There are three states of the NiFi Process, namely Running, Stopped and Disabled.
Then we’ll start and stop the two command Rest apis that are executed in the script.
? Start the command (using the Rest API’s Put method) :
curl -i -X PUT -H ‘Content-Type:application/json’ -d ‘
{
“revision”: {
“clientId”: “586ec1d7-015d-1000-6459-28251212434e”,
“version”:17},
“component”: {
“id”: “39e0dafc-015d-1000-918d-bee89ae2226e”,
“state”: “RUNNING”
}
}’ http://IP/nifi-api/processors/processorsID
? Stop the command (using the Rest API’s Put method) :
curl -i -X PUT -H ‘Content-Type:application/json’ -d ‘
{
“revision”: {
“clientId”: “586ec1d7-015d-1000-6459-28251212434e”,
“version”:17},
“component”: {
“id”: “39e0dafc-015d-1000-918d-bee89ae2226e”,
“state”: “STOPPED”
}
}’ http://IP/nifi-api/processors/processorsID

4 Summary and postscript

This article first introduced the Apache NiFi, then took the actual requirements of the author as an example, and explained the actual combat of the core component Processor of NiFi. Because NiFi still belongs to a top-level project Apache launch time is not long, is very powerful, but can access resources are still limited, in this paper, it is more of a throw brick process, its really powerful functions in data processing, welcome each other discussion of interest to you.

Actual Combat of Apache NiFi Processor,首發于Cobub

]]>
How to Improve APP Retention Rate? http://www.dpkxx.com/en/how_to_improve_app_retention_rate/ Mon, 14 Aug 2017 06:43:59 +0000 http://www.dpkxx.com/?p=6895 The sole purpose of business is to create and retain customers

Our APP marketers are always better at getting more and more people to use our products and become our users. But how do you keep them when you have them? It may not be hard for a new app to attract hundreds of millions of users. But it's never easy to keep these users and increase the number of users to 100 million.
Although users retention needs us to spend a long time to complete the work, we can still take some measures to improve it in the short term and can improve user activity, such as some marketing activities or according to the user's usage of history to their push precise related news. This is a good way to interact with the users to let them participate, users can be obtained from these activities and news to get more valuable things. Therefore, the app retention rate will be increased.

How to Improve APP Retention Rate?,首發于Cobub

]]>
The sole purpose of business is to create and retain customers

Our APP marketers are always better at getting more and more people to use our products and become our users. But how do you keep them when you have them? It may not be hard for a new app to attract hundreds of millions of users. But it’s never easy to keep these users and increase the number of users to 100 million.
Although users retention needs us to spend a long time to complete the work, we can still take some measures to improve it in the short term and can improve user activity, such as some marketing activities or according to the user’s usage of history to their push precise related news. This is a good way to interact with the users to let them participate, users can be obtained from these activities and news to get more valuable things. Therefore, the app retention rate will be increased.
In a project management APP, for example, we can send those who download the APP but not create a project an introductory tutorials, or inform them of our new app functions. These messages not only create value for APP users but also increase the APP retention rate. Doing all of these tasks won’t cost us much energy and time, but it will give us a lot of unexpected rewards.

Three stages of user retention

It’s not a quick fix to start using our APP from users to be loyal to our APP.If we don’t do anything, we don’t know when the user will be in the arms of a competitor. Therefore, this requires us to understand the three stages of user retention through data and take effective measures to retain them.
Here are the three stages of user retention:

  • Early retention: let new users take action to start using our products.
  • Medium-term retention: get existing users into the habit of using our products.
  • Long-term retention: let long-term users achieve more value through our products to enhance their loyalty.

We need to constantly show our customers the value of our products, otherwise the user will sooner or later turn to our competitors.

At this point, just remember that user retention is our goal, and then break it down into phases. This is more conducive to creating more targeted and effective lifecycle management.

Early retention — inspire “aha moment”

We encourage users to use our products usually within 7 days after the download, if we failed to arouse the curiosity of the user during this time, let he is interested in our products, after that he may never use again. That’s why we want to make sure that when we register our products, we can immediately see what our products are worth to us. This is known as the “aha moment” when users see the value of our products and think, “oh, that’s it, I get it.” New users can only get value from our products through actual operation. For impression notes, new users can create a to-do list. For Youdao, the user can be able to translate through a photo, via voice query. Your true love users will do these things themselves, but for the less active users, we need a little push. The message push is the most effective tool at this time.

Medium-term retention — encourage users to get into the habit

“No trigger, no action, no habit, no retention.”
WeChat, alipay, baidu search these products all have one thing in common – they let users develop usage habits. When we get into the habit of using these products, it’s impossible to imagine what it would be like to have one day without these products. So how do you guide users to develop these habits in our products? There are triggers behind every habit.

There are a number of ways to boot a user’s triggers, such as notifications, alerts, and so on — but E-mail and in-app push messages are especially effective. Take a blind date software as an example:

As a dating APP, if you don’t upload photos and keep up-to-date personal data, it’s hard to get romantic offers from the opposite sex. No one falls in love with an anonymous avatar. Therefore, it is necessary to edit an email or push a message in an application, telling the user to improve the information. Trigger success often depends on depth of notifications and nudges. Send an email rather than in the outside of the scene, is not as good as the user is often visit key pages for a reminder, the operation guide: “have no face, you will not be the objects of search, quick upload avatar” to a more direct and effective, if the user cannot be gained from such actions he felt something of value. No one will spontaneously complete the personal data. So back to the product itself, what can the user get out of it, what can they do here? This is the reason why they continue to stay. If your message doesn’t connect to the user’s intent and goals, it doesn’t have any effect.

Long term retention- keep the spark alive

If we are lucky enough to develop a large group of users, we will feel that the day is worth a lot of money. The mobile Internet industry is very competitive, and once we make mistakes, users are likely to lose it to their competitors. We must continuously show the value of our products to the user so as to maintain the stability and growth of the retention rate.
There are a number of ways to improve user retention, such as rewarding users, but getting older users to recommend new users through push messages can also help us to keep them.
Recommending new users looks more like a new strategy than that, but its benefits go far beyond that. Domestic mutual gold products often use this strategy. Here, we take a taxi software as an example:

Inviting friends rules here has the following 3 benefits:

  • The credible reputation recommend or bring in new customers.
  • Habits begin to form because a group of friends are beginning to rely on the same product.
  • Bilateral recommended projects encourage existing users to share, while reducing the risk of new users.

(right: the new and old customers can get a coupon, is a win-win situation) such a rule would encourage existing users form habits, also get new users at the same time, these new users are likely to repeat our old customers recommend behavior, so a virtuous circle, to increase new users at the same time improve the retention rate.

Turn our users into lifetime users

Today’s APP operators must keep winning and retaining users to survive. It would be great if you could turn existing users and new users into lifelong users.

How to Improve APP Retention Rate?,首發于Cobub

]]>
4 Effective Strategies to Improve the Engagement of APP Users http://www.dpkxx.com/en/4_effective_strategies_to_improve_the_engagement_of_app_users/ Tue, 25 Jul 2017 09:38:31 +0000 http://www.dpkxx.com/?p=6857 The definition of user engagement varies from product to product

For example, for a bookkeeping app, an effective user should be logged in every day to add his new income or spending. For an operator's app, an effective user may be logging in once a month, charging a fee, or ordering additional traffic packets. Therefore, user engagement has no uniform quantifiable definition standard, but it is possible to form the standard of vertical industry. For Banks, for example, start times the average of 1.7 times the user is a reasonable value (the value come from Cobub’s long-term observations, Cobub is a domestic open source mobile application of statistical analysis tools).
Engagement is not like page views (PV), visitors (UV), users pay or conversion rate these indicators as easy to measure, not a data statistical analysis tools to directly reflect product user engagement. However, ignoring user engagement is very dangerous.

4 Effective Strategies to Improve the Engagement of APP Users,首發于Cobub

]]>
The definition of user engagement varies from product to product

For example, for a bookkeeping app, an effective user should be logged in every day to add his new income or spending. For an operator’s app, an effective user may be logging in once a month, charging a fee, or ordering additional traffic packets. Therefore, user engagement has no uniform quantifiable definition standard, but it is possible to form the standard of vertical industry. For Banks, for example, start times the average of 1.7 times the user is a reasonable value (the value come from Cobub’s long-term observations, Cobub is a domestic open source mobile application of statistical analysis tools).
Engagement is not like page views (PV), visitors (UV), users pay or conversion rate these indicators as easy to measure, not a data statistical analysis tools to directly reflect product user engagement. However, ignoring user engagement is very dangerous.
Although user engagement is more difficult to measure and establish a unified standard, we can obtain the following indicators from the mobile application analysis platform:
1. Average startup times
2. Average length of use
3. APP interval
4. Number of visits
5. Retention rates
6. Response scale to customer contacts
Using the following radar map to represent user engagement.

Why is user engagement important?

The market is often used to measure the performance of an application. But more than 20% of these downloads have only been turned on once, according to empirical data. As a result, APP operators are getting a lot of registered users, but they can’t all turn into “customers” effectively. So how to keep users engaged continuously is the most important problem that APP operators should solve.
By the way, distinguishing between “user” and “customer” is an important link in Internet product operation, especially for non-tool applications. It is also important to note that “users” can be turned into “customers”, and simple and simple is wrong.

As can be seen from the above: user engagement is one of the difficult problems, but it is often overlooked, so there are four strategies to improve user engagement:

1. Make a deep first impression

Every day, potential target users learn about our products, download apps from various sources, and see the product interface for the first time. However, in the design team of mobile application, it is often forgotten that the interface that users first see should include the following three functions:
A. Explain the function of the product
B. Motivate the user to start using
C. Let users know how to get help

If one APP lacks any of these features, users will be able to write off the APP and never look back.
There are many good ways to welcome newly registered users and show them our hard work results. For example, new welcome page, product use tutorial, novice guide or related data related to the product. When users first open their mobile applications, they should see a nice welcome page and inform the main features and features of the product before they use it. The first impression will greatly affect the subsequent product use experience.

There are many kinds of user contacts, such as SMS, push, email, etc. APP operators can use SMS or email to greet newly registered users or push marketing activities. But the data suggest that messages within the application are more likely to be read, clicked, and responded. You can add a button to the application on the last page of the boot page, which is very useful for entering the app home page. You can also tell users where to ask for help by in-app messages. Keep a reply box at the bottom of the help page to facilitate user connections and improve user experience to improve conversion rates. The message sent by the APP operator is the first step to establishing a connection with the user, so don’t ignore it.

Simple communication with users is a great way to encourage them to ask questions, and you need to try new ways and stick to them. Through communication, the APP’s product manager can find out what the product lacks, or what the user’s opinion or dissatisfaction is. Product managers can better understand user needs, better design, and operate products to help mobile apps win higher user satisfaction and more customers.

2. Gradually expose the product depth

Any valuable product has some less obvious but useful features. These features include mail notifications and reminders, third-party integration, and export functions.
Often, these more elaborate designs are not immediately discovered and used by users. Just like before using data, who CARES about exporting the data? Or who would consider using keyboard shortcuts before using the unique features of a product? As a result, operators of most products tend to expose these product features through untimed tweets, emails, documents, or frequently asked questions. But these methods are not very effective.
When these messages are not available and are more likely to disturb the user than to interest them, the result is counterproductive. The user’s response is likely to be to throw information into a spam mailbox and stop receiving all the information sent later.
If the common issues or help module to prompt the user product promoted which features, also means that the information was found to be the only chance is: users in the use of products encountered a problem. But it’s not an appropriate time to learn about new features, because the purpose of their trip is to solve existing problems instead of discovering new lands.

Define a timeline for a message

It is recommended that APP operators create a news feed schedule based on user behavior analysis data, and gradually improve some functions according to the user’s usage. When you have a deep understanding of the user base, we can see which of the more secondary features in the product makes the user experience better, and in what scenarios it is useful to them. Once you have done this work, the rest will be timely communication. Can according to the behavior of the individual users send him a series of E-mail, SMS push and application in automatic message, to carry on the fine operation, marketing activities, each of us a message will prompt the user working towards a common goal. Every piece of information will let users know that they can come to us with any questions during their use. This is the key to getting user feedback, which can help us adjust our marketing strategy and increase user engagement.

3. Release function and improve within the application

Users in the process of using the products, they are very concerned about the product to have continuous improvement, or have developed a new function, we need to do is to continuously optimize product and timely release new features, to attract users, increase user engagement.

When it comes to attracting users, in-app messaging is a hundred times better than E-mail. This push message should include an overview of the feature, as well as a novice user tutorial that allows users to test it with the click, swipe, and so on, and get the relevant response. It’s hard to imagine a similar effect in an email.

4. Let the target user participate in product testing


The products of startups generally fall into two categories: those that address the pain points, the problems they encounter, and the ones that people are happy to have or use. These two categories are like pain killers and vitamins. If products are painkillers, then we can learn a lot from test users.
During a period of public testing, we will find that there are two types of users: one is to know the product on social platforms or other channels, and these people may choose to download the product. But they don’t have a pain point for the product, just curious to see a quick solution here. These people can’t become customers or generate valuable feedback on the product.
Another type of registered user is someone who is genuinely interested in the product, but they need to be promoted to become a customer. If we’re not sure what the user’s target is, we can look directly at some of the ideas that come from using the product. This may seem difficult, but in-app messages can help us do this.

Summary

We should manage the user and product life cycle in a reasonable way, from “understanding the user” to “enhancing the user experience” to “precision marketing”. In this process, it is necessary to establish a reasonable index system for user engagement. There may be different situations in each APP, but the APP operators must set up such indicators. This paper gives an example and some experience summary.

4 Effective Strategies to Improve the Engagement of APP Users,首發于Cobub

]]>