In our daily work, data analysis is used in many times, and it can play a lot; as designers, there must be some understanding of the basic knowledge of data analysis, and can better and Team communication and strengthen its own capabilities; the author shared some basic knowledge about the data, let’s take a look.
As we all know, the Internet is a high-speed development industry. The arrival of the data era is one of the future trends that everyone can be underestimated. It will not only affect all aspects of social life, and will also bring the sake of the Internet. Variety.
The design of big data is no longer subjective. Currently, more and more companies advocate designers to participate in the overall process of products in the project, and use data to better promote business development.
First, why do desiors understand data?
In the early Internet team, everyone’s division of labor is more clear, and all functions are responsible for their own business. The designer is only participated in the project as a link in the entire link, and the whole project is not enough, lacks the global view of the project. .
However, with the rapid development of the Internet, the company’s ability to designers becomes a special, many enterprises UI / UX posts have become a position for experience designers with higher requirements, and we can see To the industry; in order to get better with the times, we should establish a consciousness of the owner and take the initiative to help your own career better development.
The role of data is to understand the product, the user’s basis, the output of the measurable plan, reaches the target, and business design needs rationally observation and thinking, data is a way to be rational; if you want better Solution, create business value, then data is a nice entry point.
1) The designer can use data as the roots of rational design. As a designer, we should also try to start from the source of the problem. By using data to understand the user, use data to understand user behavior, provide auxiliary reference to design, help designers are correct The direction is promoted, and there is also a chance to discover new business opportunities and product outbreak points, and use the help of the product to take results by designing the revision.
2) In actual work, we are not only completed by mission. Sometimes, we also need to have a product team and operational chat demand, talk to the development and talk indicators, and talk about cost-reward ratios, if we don’t have a concept of data indicators, it will give you a certain impact on communication.
3) In addition, as the company’s ability to designers is getting higher and higher, designers need to understand business, design and empowerment, can do growth, do not understand data, designers who do not understand data may be more difficult to get better job opportunities And better career development.
Second, what is data
Understanding anything needs to understand his definition. The data is the result of facts or observation. It is the logic of objective things, and simple data is a set of numbers, and there is no value.
However, after people associate a series of data with the observed entity, the entity’s attribute gives the data imparts a meaningful information, which has become a valuable information.
This paper mainly refers to data related to users, products, marketing.
Third, how do designers use data?
The design value of the designer is to provide visual support for products. If we want to get greater growth, further development, then in actual work, we can use data to understand our users, understand the essence of demand, and clear design goals. .
At the same time, you can also guide our design, let the design have reasonable, and go more calm.
Using data can be used to use data to help us prove the value of the design in the project and improve the words of the design team throughout the project.
Finally, we can also understand our product’s use, explore products that can be improved in products, and improve the product experience, and also reflect your own greater value.
1. Close to our users, deep digging demand nature, clear design goals
Although the designer’s ability to solve the problem is very important, understand the essence of demand is a prerequisite for solving problems, and the direction of design is very confused without clear demand.
And when docked with the business side, because the differences in posts, it will lead to the understanding of the two parties, which leads to the state of designers to change the change; therefore, in depth, explore the nature of demand before design implementation, and Other business parties are necessary to be more necessary.
The data is quantified for user behavior, and behind the numbers are real users.
By designing qualitative analysis or quantitative analysis, research and analysis of user behavior data or online activity data, understand how the user uses our products, whether the user’s experience process is the same as the idea of ??our design product, Find the confusion of users in poor performance, find and discover design opportunities, and achieve the effect of optimizing user experience through such analysis and optimization methods.
2. Promote the feasibility of design programs to promote design programs
When we are internally, external opinions or programs are not recognized, we have proved that our design ideas and design is correct, how to prove our optimization solution is the improvement of products, how to prove our design Is the plan worth promoting?
The designer who is not good to say is often under the wind in the debate with the product, operation and other related party, but we can prove the feasibility of the design of the design. Data will not be deceived, we can find the existence of product design. Questions and verify the problems, optimize the problem of solving the existence.
3. Data assisted us to make decisions, determine the optimal programs
In the actual project design, multiple designers often participate together, give a variety of solutions, or a variety of designers out of the same designer, then in this case, how do we choose the most in line? What is the design plan for the current demand?
We can’t predict which one program is 100% viable. In addition to experience, some data can also give us a certain decision support; for example, we can refer to the relevant or similar modules to do auxiliary decision, and even we can pass through A / B Test to verify which design is more suitable.
4. Verify the effectiveness of the design, prove the value of the design, and the data results are oriented, constantly adjust the design strategy
After the new features are online or function iteration, the effect is unable to see very intuitive, so we need to make the relevant data analysis, observe user data, traffic after the online data is taken. Change, click rate, conversion ratio, etc. We can know if the new feature online user is recognized, and the data after iteration is improved than before.
By data, it can not only find problems, but also measures the target completion. If there is still a poor performance, continue to analyze the data can be further optimized, which in turn can promote the optimization of the next round of design, and enter a benign circulation.
Fourth, data classification and source
Before you want to figure out the source of data, we first understand the classification of data. Data has different categories from different dimensions. We simply distinguish from data analysis dimensions, and the data of Internet products is currently more commonly used data types. It is mainly divided into: Quantitative Data and Qualitative Data.
Getting more biasing methods of qualitative data is usually the acquisition of the subjective feelings of the participant experience system availability, and it is judged that there is a problem with which aspects of the design.
More common source of data is investigating data. The research data is mainly sampled data obtained by users; through collecting user subjective feelings or attitudes, such as user interviews, availability testing, message board mining, etc., it helps us to determine the problem, combined Our own interaction and design knowledge to determine if it changes or continues to optimize product design goals.
Quantitative data refers to those measurable numbers with scientific. Quantitative data provides an indirect assessment of design availability, or by quantitative data verification qualitative research results; more common data sources are product log files, background data, investigation data, embedded data, etc.
PS: The dock data is mainly referred to by applying data on the application of the point statistics. It is a common data acquisition method. It is mainly through technical means to help us obtain behavior data within the application; simply, it is in the application Enter a code, monitor the user’s behavior event, which can provide us with the real situation of users using the application, such as the click rate, transformation, and hop rate, etc. of a certain activity.
These two data are visible from different angles, and they play an important role in the design iterative cycle, which is complementary. Qualitative hypothesis cause, then verify it by quantification; quantitative excavation data, then analyze the reason.
5. What are the common data indicators?
Common user indicators can be divided into user data, user portrait data, traffic data, user behavior data, income indicator data, channel indicator data, research data.
User data is relatively easy to understand, that is, “Reaction user attribute, and user-related data”; user data contains and excavated content, there is a lot of small aspects, there are users active, retained, new, etc., from big aspects Speaking, there are user portraits, user properties, preference properties, user behavior habits, etc.
The points in the quantitative study of user data are mainly the number and quality of the user, common like: active users (DAU, WAU, MAU), new users, user retained rate (subtle retained rate, stay on the 7th Rate, monthly retained rate), cumulative registration user, etc.
DAU: A daily number of users. WAU: Visit the number of users per week. MAU: Visit the number of users per month.
These three indicators are to remove active users after repeating users, and a user has access to the product in the same channel within a computing cycle, and only one, generally used to measure the viscosity of the product to the user.
It is mainly to study the storage of new users, refers to the first time the APP in a certain cycle has been launched by a period of life after a period of life, and the length of the user’s life cycle and the room we can improve.
Among them, two other indicators: keep the next day and 7 days. Through these two indicators, it can reflect how the user quality has been added through a certain launch channel, as well as in accordance with these two data indicators and other data, determine the subsequent guest approach.
Storage rate = number of new users to log in / new users * 100% Second day retained rate = new user new user new user login number / new users in the same day * 100%
7-day retained rate = new users after seven days after seven days * Add users * 100%
Cumulative registration user:
The number of users who log in to history, the same user multiple access does not repeat.
2. User portrait data
The user’s portrait data is also a user data. It is usually from the perspective of the user attribute. It is usually the data obtained by qualitative research, reflects the features of the existing user population, and the configuration can be divided into three classes: basic attributes, user behavior properties. , Preferring; such as age, gender, new user / old users, user level, purchase preference, region, terminal and model distribution.
3. Traffic data
Flow data is mainly based on user access to the product / page, reflects the overall situation of the page. Comparable, such as: page PV, page UV, per capita access, traffic source, traffic direction, etc.
UV: Number of page access, describes the number of users accessed, UV can come from various ways, such as external advertising drainage, internal resource bit allocation, user active return traffic, activity drainage, etc.
PV: The number of page access, that is, how many times the page is viewed, PV repeats the number of users’ access, and the PV is more accessible multiple times in a certain period of time.
Per capita access: Per capita access is the number of times each user accessed within a statistical cycle, that is, the number of access times / access.
4. User behavior data
User behavior data refers to a series of behavior data indicators related to the user’s operations in the product.
Click rate (CTR) = Module Click to click on the number of modules
The user’s click rate refers to the number of clicks of the user on the module, accounting in the traffic of the entry page.
Exposure rate = module exposure number / page view number
The exposure can see how many users have seen how many pages depth are seen in the browsing depth on the page.
The length of use refers to how long it is on the page in a certain cycle. Using time long attention to three data, use total length, single use time, long-term use.
The hopping rate is that the number of access times that only access to a page accounts for the total number of access times.
UV value = GMV / traffic
UV value refers to the average amount of each UV output, which can reflect the quality or traffic of traffic and traffic from the side.
5. Revenue indicator data
The revenue indicator mainly refers to an indexed ARPU associated with income in the commercial process, such as GMV, conversion rate, customer list price, order number, number of paid people.
ARPU = total revenue / number of users
It refers to the average income generated by each user in a certain cycle.
Total sales GMV = order quantity ¡Á guest price
GMV refers to the total amount generated by the order, including payment amount and payment amount, is a common indicator in e-commerce products, usually reflects the volume of an e-commerce platform.
Order quantity = Access UV ¡Á order conversion rate
The order amount refers to the total number of orders generated, and there may be multiple items in one order, which is different from the sales volume of goods.
Transformation rate = order quantity / access UV
The conversion rate refers to the ratio of the number of transformation behaviors, such as a certain activity of a certain activity, is referred to as a statistical period. The transaction volume of the product accounts for the proportion of the number of commodity exposure.
Guest price = total sales / order quantity
The customer price refers to the average purchase of the goods, which can reflect the purchasing power of the user. However, the price of the customer has strong category. Different merchandic passenger bills is large, in some operational strategies, usually through full reduction, etc. Marketing activities, stimulate users to buy more products, and increase customer price.
6. Channel indicator data
Through statistics from the delivery data of each channel, there is a probably awareness of different channels, and the quality and value of new users are analyzed by user follow-up behavior, reflecting the value of various channels.
The designer does not come into contact with this type of data in actual work, so simply introduced below.
Channel flow refers to the exposure rate brought by a certain channel.
The number of channels of channels refers to the number of users added through different channels, such as electronic market channels, digital marketing channels, and offline channels.
Channel user conversion is the exposure of a certain channel, accomplishing a new user accounting of the number of downloads and registration numbers.
Channel ROI refers to the output ratio of a certain channel, such as how many times, how many times in transformation, and ultimately bring how much installation / sales, put it to calculate how much actual user of each advertising fee or Income, this is ROI.
7. Research data
It refers to the sampled data refining through the user. Such as users expect value, satisfaction ratings, subjective evaluation, net recommendation value NPS, return / restriction, EEG test data, eye dynamometer test data, etc.
Research data can help us find problems, refine the problem, find a specific reason behind the data. Six, data collection principles
Many times the designer often doesn’t take the initiative to watch the awareness of data, more common is that the business or product is synchronized according to the situation, and the data they have already analyzed, such as click rate, conversion rate, etc.
These processed data are often difficult to see specific problems or reasons, we need to see data from more dimensions, which is conducive to our more in-depth analysis, research issues.
Original information: Data that has not been filtered through any person (not abstract data that others have handled, is a real data performance or user investigation of one-hand information).
However, when the information is collected to a certain extent,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
In the information collection phase, do not make judgments, avoiding other content in the back process for judgment to judge.
Seven, how to find problems through data
Trend analysis is the most commonly used analysis method in data analysis, as long as it is related to data, almost everyone, always is in use. The greatest benefits of trend analysis methods are simple and provincial things!
Because it doesn’t need any ainci-foundation, there is no need for any expertise, and there is no need for a lot of data. Overall change, find influencing factors.
Isolated data is meaningless, there is a difference in comparison.
Comparative analysis is a very important analysis method in actual data analysis. We need to find the same type of comparison, you know how to get bad, you can improve where you can improve.
3. Multidimensional analysis
Multi-dimensional analysis is subdivision analysis, doing multidimensional analysis first to clarify 2 directions: dimensions and indicators.
Indicators: Refers to the unit or method for recording key processes, measure the target, such as DAU, retained rate, conversion rate, etc. Dimension: Refers to the angle of observation indicators, such as time, source channels, geographic locations, product version dimensions, etc.
Multi-dimensional analysis is to disassemble the problem in multiple dimensions, observe indicators in comparative segmentation dimensions, and find more problems by subdivision of a comprehensive indicator.
For example, a resource-bit conversion rate is the most serious problem we face, then we can continue to refine the factors affecting the conversion rate, and give these factors separately, give a solution.
You can also look at the problem from the dimensional direction, such as a time-to-time data, or the user data from a certain channel is relatively poor, then our re-targeted gives a solution; usually we look at a comprehensive indicator, The total value, but these total values ??are usually unable to find problems.
In actual operation, we need to disassemble the total value to disassemble the total value to discover the problem, find the problem that can be found.
4. Formula dismantling
The formula dismantling method is for an indicator, and the influencing factors of the indicator are quickly found in the formula hierarchy, so that the factors affecting the indicators are quickly found.
The formula dismantling method does not have a fixed standard, and a target variable is not the same as in different scenarios, or to solve different problems, and need to utilize formula dismantling. E.g:
5. Data model
The above is to analyze the data through some simple methods, but the dimensions of the data tend to have a lot, complicated, in order to help us better data analysis, we can also help us with the data model.
There are many types of data models, and each data model has different focuses. This is not expanded here. It only lists some commonly used data models. Everyone is interested in searching for itself from online search. For example: Google’s GSM Analysis (Design Verification), Value Matrix Analysis (Determination Direction), OSM Model (Target Verification), AARRR User Growth Model, DuPont Analysis, RFM Model, Funnel Analysis, etc.
Eight, how to cultivate awareness of data analysis
Finally, we talk about how to develop awareness of data analysis, I believe read this article and see the designers here are well aware of the importance of data designer, I personally think that learning anything, in fact, are interlinked method .
First, we must first establish a sense of self-learning, in their daily work often remind myself to pay attention to these relevant content, and actively seek and grasp the opportunity to participate or to learn;
Then, by studying books or the Internet system of relevant knowledge to guide us through a reasonable method to establish our awareness and knowledge;
Another is to take action, through constant practice, constantly trying to work actively sort out the problems encountered, summarizes the problems encountered, iterative own knowledge base;
Finally, we have to teach to learn, by sharing their own summary, to sort out their knowledge base, share to more people, to improve their own, through feedback received continue to improve themselves.
This article from the original release @ gentle giraffe on.