This article will share my personal judgment experience, hoping to draw everyone's attention to judgment itself. Also, because decisions are often dynamic, with varying styles and experiences, they are indeed "for reference only".
We are making judgments and choices all the time: which type of fruit to buy, which movie to watch, which driveway to take... Daily choices are always so random, and we can say "be careful next time" if we make a wrong choice , with the possibility of early termination. But when you enter the work, you will find that judgment will become difficult.
The judgment at work, whether it is the selection of technical framework, the focus of work content, the choice of partners, etc., will affect the process and results of follow-up work, and ultimately affect: happiness, growth and harvest, and material incentives. This makes us cautious when judging and full of hesitation when choosing.
As we grow older, the cost of trial and error will increase. We are eager to find relevant experience: to be able to "hit the ten rings with one shot", and get higher results with fewer times.
judgment information
Just like "a clever woman can't cook without rice", if there is no input, it is difficult for us to judge and produce output. In addition, from the perspective of programming, "program = data structure + algorithm", if the data information itself is missing, then judgment as a behavior is difficult to meet expectations.
Therefore, it needs to be clear that the prerequisite for better judgment is to obtain more information. Some commonly used message types are listed below.
▐Case : looking for templates
If you can find similar problems and cases, it is like having a "God's perspective": you can quickly understand the types of problems and common strategies, and after matching, you can quickly derive the methodology. For example:
Question type judgment
When doing code questions and daily optimization, you must first judge the type. After the judgment, you can determine the algorithm template and start quickly:
Recursion problems: abstract recursive methods;
Sorting problem: determine the sorting algorithm;
Shortest path: define the path weight.
character judgment
When working with people, judge the style of your partner:
He is the dominant type, so he needs to listen and assist more;
He is a follower, so he needs to invest more and take the lead horizontally.
strategic judgment
When a new company enters the market, it will study the existing company's products, organization, strategy, etc., and by copying the business strategy model, it can avoid "detours"
Quick response: Use a flat organizational structure to improve response efficiency;
Seize the market: use the low price strategy to pull the scale.
To learn from other people's experience, the key is to find the "source" of learning. In terms of influence frequency: Although excellent partners and related sharing can give people strength, the coverage is limited after all, and it is unlikely to become the source of strength for most people. So, you need to look for "more general" input:
Chinese culture with a long history
The Doctrine of the Mean: "Too much is not enough" in "The Analects" teaches us to do things just right;
Falling behind and being beaten: The development of modern history has taught us to vigorously develop science and technology and the economy;
False success: "The Unspeakable Academic Overlord" reminds us to do things with one bang.
The development history of each country
The transfer of labor to cheaper places: from Japan and South Korea to China, and now to Southeast Asia;
Suppression of rising stars by leading countries: The previous suppression of Japan by the United States has now shifted to a game of great powers with China.
Knowledge accumulation during schooling
Focus on intrinsic value: discuss the relationship between price and value;
Look for everyday patterns: the circumference of a circle is equal to 2 π R;
Higher theoretical level: the relationship between the theory of relativity and Newtonian mechanics.
We are actually very clear about these phenomena, and if we abstract them further and make connections, we can solve many problems. But we often lament that we "don't know how to solve the problem" because we lack sufficient awareness: we need to train the ability of association and association, so as to "stand on the shoulders of giants and practice the existing conclusions".
Standing on the shoulders of giants
▐Cornerstone : Bottom-up
Sometimes what to do and what to do may be very broad, but just like the saying "keep your feet on the ground and look up at the stars", thinking about uncertain things can be based on certainty. So, the cornerstone under your feet is also an input. What is the underlying foundation stone? This depends on where you're going:
The cornerstone is code logic: as a business development, it is necessary to undertake business requirements, and the cornerstone is the business code logic in hand. You should look at the code more, understand more restrictions and constraints, and find out the space that can be expressed.
The cornerstone is the basic ability: As a product, product design needs to be carried out. The cornerstone is the basic ability of the cooperative team. You should know more about the current situation, know the risks and the areas that need to be improved.
The cornerstone is team experience: As a team TL, when developing new markets, it is necessary to divide personnel responsibilities. The cornerstone is the experience and direction of the team members. You should observe and communicate more to judge which places are more suitable for everyone to exert their efforts.
However, the foundation is complicated, and how to understand it is often difficult to make a good balance:
The granularity is too coarse, and there is no feeling: just pass it in general, just read it, don't ask for a deep understanding, it took a lot of time to read it, but looking back, I feel like I didn't remember anything.
The granularity is too fine, too local: In-depth traversal, every detail is looked at. After a long time, you may not see much, and because it is too local, it is difficult to match the problem.
How should we break the situation, how to understand the underlying foundation, some strategies can be referred to:
take notes
Notes are indexes: A good memory is not as good as a bad pen. Notes are actually indexes, and we don’t need to memorize them all.
The definition of remembering: within a certain period of time (such as within 1 minute), being able to find specific logic can also be considered as "remembering".
Do a good job of classification
Distress of not classifying: If there are so many knowledge points, if they are not classified, they will become more and more confusing and difficult to find, which will cause an embarrassing situation where the more you look, the more confused you become.
Classifiable categories: For development, it is a good entry point to classify according to the perspectives of "application", "business activity", and "field".
focus on
Look at the more complex ones: because it is simple, you can also look at it when the time comes;
Look a little more at high frequency: because this can solve 80% of your problems;
Take a closer look at the closed loop: because this can give you a complete thinking.
▐Decomposition : top-down
Decomposition looks easy, but it is actually very difficult. Because the upper layer may only give a strategy and direction, how to do it specifically? , Can you do it? , Will there be good returns? Most are unclear.
Although this is very general, it is indeed much better than "finding a needle in a haystack" and "nothing", and it is an important input: the general direction is generally very cautious, it is impossible to completely fail to land, and it has potential guidance.
When decomposing from top to bottom and making a judgment, you also need to pay attention to several points:
search for original intention
Difficult to understand in part: At our place, it may have been decomposed into several layers. If you only look at one or two things, it is difficult to understand.
Searching for intentions at the upper level: think to go up, explore the complete strategy and logic, and try to think: what is the right level to achieve, how much manpower to invest, etc.
Understanding Decomposition Requirements
Preliminary research: Some are just sorting out and understanding the background, and it is not yet time for real decision-making and implementation;
Prototype verification: Some require some demo verification to obtain practical secondary input, but it does not say that it must be very complete. If you do encounter difficulties, you can discuss them again.
Resolute execution: Some tasks are deterministic, even if you encounter difficulties, you must overcome them, and you need to grit your teeth and persist if you invest more time and manpower than expected.
have your own judgment
Be skeptical: it doesn't mean that top-down decomposition is ironclad;
Timely feedback: During the practice process, if you find something that is not suitable for the production law, you must give feedback in time to avoid investing in the Great Leap Forward and entering a more embarrassing situation.
In many cases, it is more important to "observe words and feelings" and find the key points. Although it is difficult to be as direct as "this is the key point" that the teacher said in class, we can still find some characteristics: repeated mentions, key releases, key concerns, etc.
Even the distant starry sky can point to
▐Environment : general trend
We often hear that we should "follow the trend, take advantage of the trend, and build up the momentum".
When the development is good, it can be done: when the enterprise is developing well, many projects can be launched regardless of whether they take advantage of the trend, take advantage of the trend, or create momentum, because even if the value is not so obvious, there are still some in the context of large development. .
When the development is difficult, follow the trend: when the momentum is slowing down, if you can’t follow the trend, then building up the momentum may become "going against the trend", a dilemma, and even lead to a situation of "cutting the mess quickly" at the upper level.
To make the above judgments, the key is to "assess the situation", so what are the momentums:
Infrastructure is a momentum
The value of efficiency improvement is obvious: When the infrastructure is not perfect, the value of efficiency improvement will be obvious, and the construction of infrastructure will continue.
The level continues to rise: from basic traffic load balancing, to the construction of various messaging, RPC and other middleware, to the establishment of various middle platforms, and then to the accumulation of various industry products, etc., it is getting closer and closer to the business itself.
The good life is a momentum
Quantity turns to quality: When the product construction reaches a certain level, the pace of development will slow down significantly. Everyone will start to think about how to better improve the R&D environment, and will pay more attention to the overall cost of the software life cycle. Quantity support will also gradually shift to quality construction.
mostly a momentum
Short-term head: After a lot of competitors are reduced, it is expected to make extensive profits through top customers, and the direction of support has also been adjusted.
Long-term majority: But later we will find that the new wave comes from most of the non-head customers, because the people represented there have more frequent changes. In the long run, most of them are a kind of momentum.
Cost reduction and efficiency increase is a momentum
Contraction after expansion: When the enterprise develops to a certain level, the growth momentum will slow down, and it will pay for the previous expansion. Cost reduction and efficiency increase will become the overall strategy of the enterprise from a major event in individual departments, and this will be a long-term matter.
Everyone can indeed feel the momentum, but to link these momentums at work, an important judgment is: whether these momentums have approached or even entered our working environment. If there is still a certain protection area, inertia may still continue to do things according to the previous thinking, and may be paid in the short term. However, the word "changing with each passing day" should be something worthy of awe. If the mind does not move forward, when the tide comes, it is indeed "avoidable".
The Laws of Nature Behind the Waves
▐ Surroundings: cognition of others
When discovering information, there will be constraints such as vision and work commitment, and the information that an individual can mine is limited. Just like the concept conveyed by "1+1>2", we should communicate more and obtain information and opinions from partners around us. How to better obtain the cognition of those around you:
looking for target
Start with the end in mind: If you have clear target customers, then their views, opinions, requirements and other input will be a good direction and key constraint for your product.
Good at listening and summarizing
Core Judgment: You will hear different voices. You must be good at hearing which are the core judgments, and summarize them while seeking common ground while reserving differences.
Tolerance is great: it should be inclusive, and when information and opinions are inconsistent, they cannot be forced to output.
wide range of colleagues
Industry views: "Others" do not necessarily refer to work colleagues, you can find external conferences, books, etc., to obtain research and opinions from industry colleagues.
After obtaining various sounds, they need to be processed to dig out deeper rules:
cause of thinking
5W1H: When you hear that there is a certain direction behind the team, don’t just focus on what to do, but think about why you started this matter at this stage, and think about deep-seated direction adjustments such as the relationship between production and technology and development strategies behind it, so as to better Nice to stick to the main line.
don't just see the good
Dialectically look at the problem: In meetings and books, I saw that others have gone through a similar path, and mentioned the "lost" part behind the "flowers". I want to focus on reviewing these problems. Is it acceptable? Can it be avoided? If it is unavoidable, in the long run, are you willing to continue to maintain it? Is this really the way to go?
to dig recursively
See through the problem: When other theories are cited in other people's views, in order to better understand this part, it is necessary to look at its development process and study the laws behind it in depth.
Being good at discovering and communicating, maintaining sensitivity, being tolerant, and having your own main line are the key points to do a good job of connecting with external voices.
Ants transfer information and cooperate with each other
Judgment mode
Having the information to judge, we will make a judgment. When judging, according to the complexity of the matter, different judgment modes need to be adopted:
Based on experience: Some are relatively simple and clear, and you can think quickly and give answers based on "intuition" and "feeling".
Try the model: Some have multiple plans, each with its own advantages and disadvantages. It is necessary to determine the value formula of the target, calculate the input-output ratio, and choose the better result.
Simulation: There are so many influencing factors that cannot be measured in a short period of time, and there is also a lack of reference cases, so the method of simulation can be used to make subsequent judgments based on the simulation results.
▐Significant factors: based on experience
In the face of simple questions, the answer can be given at once, blurted out, and the judgment can be made immediately, which can be called "based on experience". The essence behind this is that there are very significant factors that can achieve the effect of "one-vote agreement system" and "one-vote no system":
Scenario 1: Synchronous call or asynchronous call
There must be results in real time: If consumers are waiting for results in real time, they need to adopt a synchronous method.
Synchronization can't handle it: If there are not enough machines and have to cut peaks and fill valleys, asynchronous methods are required.
Scenario 2: The way of unified integrated deployment is external transfer
Time-consuming requirements: If there are strict requirements on time-consuming, then it is necessary to integrate and deploy as uniformly as possible.
Boundaries such as organization: The system is very complex, such as logistics and payment systems, so they will choose the way of external transfer for connection.
Scenario 3: Self-built business or platform intervention
It cannot be done alone: the business cannot be realized in the extended area, and the platform code needs to be changed, so platform personnel need to intervene.
Willing to pay for resources: The business side can complete it, but there is no manpower. If it can pay the cost, the platform will be willing to support the development.
Judgment based on experience, the key is to find the core judgment points, find these points, you can quickly make a judgment. When it is difficult to judge, looking for "elderly" consultation is also to obtain these key points. How to find the judgment point better and faster, here are some suggestions:
Base-based: think based on the production materials in hand. With this data, you can determine what can and cannot be done.
Empathy: see who is more suitable to do this, can the business be done by yourself? From a long-term perspective, it generally depends on who understands and manages the core logic.
Consider the cost: judge the cost of doing something, if it can be done, but the cost is very high, and no one pays, then it can affect the investment decision.
Prefer to choose red apples, color is a significant factor
▐ More factors: try the model
Sometimes, there are many factors that need to be considered, and it is difficult to deny or agree with key factors. Moreover, various solutions have their own advantages and disadvantages. At this time, when making judgments and choices, the ratio of input to output (ROI) is often compared.
How to think about this input-output ratio more rationally, consider a more generalized abstraction:
Value X = Proportion ax Value of Factor A + Proportion bx Value of Factor B + Extra Value D
Cost Y = proportion cx cost of factor A + proportion dx cost of factor B + additional cost E
Benefit Z = Value X - Cost Y
When the "proportion" here is very large, it tends to be "based on experience" judgment:
One vote is in favor: when a certain proportion of value is very large, the income is basically positive, for example: it is the key strategy for this year.
One vote veto: When a certain proportion of cost is very large, the income is basically negative, for example: a self-developed team needs to be established to support the demand, and the market size behind this demand is small.
When these factors are very similar, the size of the "proportion" determines our choice, and this "size" depends on the values and styles of each team:
Speed priority: The proportion of supporting business development in the business department is very high, and sometimes a little "customization" of technology implementation is acceptable.
Quality priority: The horizontal department has relatively high requirements for efficiency, stability, and continuous efficiency improvement. It does not mean that it does not support the business, but it means that it will be more stable and general, and the construction period will be longer.
Such a trade-off, there are more classic scenes in life, for example, navigation:
If you are in a hurry, give priority to speed: sometimes you may go to high speed, and the cost will increase, but you can drive comfortably;
Don't worry, economic priority: don't take the high speed, when you encounter congestion, it is acceptable to slow down;
Take a comprehensive look: Avoid high-cost and particularly congested ones, and the remaining multiple schemes are viewed from the perspective of time priority by default.
It can be seen here that different development stages and different environments require different requirements. This is reflected in the fact that "extra value" and "extra cost" will be dynamically adjusted and have an impact.
Reveal objective laws based on models
▐Complicating factors: simulation simulation
When there are too many factors to consider and the cycle is long, any wrong grasp of any factor may cause a "butterfly effect" and affect the overall income. Such judgments are generally related to major directions, and particularly large strategies may not be accessible, but the direction choices of teams and departments are still closely related to us.
How should such a complex decision be made? The logic of the "micro-traffic simulation" that students came into contact with is actually similar:
The uncertainty is very large: a traffic accident of one car may lead to a series of rear-end collisions, causing the entire traffic jam to suddenly become serious.
No historical logic reference: Similarly, what will be the impact of adding a new ramp entrance and exit?
There is no experience to refer to these, and it is necessary to do simulation verification based on the traffic of the segment. There are actually many other simulation ideas: simulation of aircraft resistance models, rehearsal of various activities, everywhere.
In our work, we also participate in many "simulations":
Partial simulation: When doing large-scale system refactoring, first verify a certain function point and partially refactor it. If the result is good, then roll it out;
Grayscale simulation: To adopt a new architecture system, first take some business traffic for closed-loop verification, and look at the feedback encountered in the entire R&D ecosystem;
Offensive and defensive simulation: Focus on evolving a set of product capabilities: Material reporting is required, and a variety of daily bad case use cases are checked to see if the problem can be solved.
There are several points to pay attention to when doing a good simulation:
The basic data must be accurate: In traffic simulation, the models of various vehicles and the proportions of driving styles need to be sampled and investigated at various intersections. The same is true at work. You can't just pat your head and judge a few points, but you must be able to reflect the general public opinion.
The evaluation criteria should be recognized: in the traffic simulation, the final traffic operation status will be calculated comprehensively based on the national traffic congestion index calculation method based on the speed of the car, taking into account the speed differences of cars, trucks, etc. The same is true at work. In the early stage of launch, we may focus on investment, and various indicators are very good-looking, but we should pay more attention to long-term, common, and comprehensive evaluation indicators.
When the simulation results are not good: During the virtual simulation process, variables can be adjusted to test out good results, but in daily work, every attempt is costly, so it is necessary to deeply analyze the cause of the problem, whether to move forward? Continue to study? Timely stop loss? Judging requires objective and fair in-depth thinking.
Wind resistance simulation verification
judgment strategy
When making specific judgments, it is actually similar to "making test papers", with different types of questions. When answering different question types, there should also be corresponding question-making strategies.
▐High certainty: multiple choice questions
The multiple choice question is to choose the correct one among the existing options. In daily work, as long as you are good at analyzing the advantages and disadvantages of various solutions, choose an appropriate one, and focus with TL, you can basically get the answer right.
But when the question comes to you, you may not be given the complete options of A, B, C, and D, and will tell you: "There are routes such as A, B, etc., you can go and have a look." When we do not feel that there is a better answer, we are required to have the ability to supplement options. Measures to add more valuable options include:
Pay more attention to technical blogs: the highlights that others often mention are generally better answers;
Communicate more with products: understand the road ahead, and good options should also be able to pave the way for follow-up work, representing a long-term approach;
Duohe TL needs input: understand the environmental background and deployment of troops, take advantage of the situation and take advantage of the situation, and dig out better options for combination;
Pay more attention to your choices: some things have already happened and been solved on a daily basis, all that needs to be done is to abstract and define these solutions.
When making a choice, there is a saying that "I always choose C, and I can always be right." What is worth thinking behind is: no matter what you choose, you must establish your own judgment principles, otherwise you will follow the trend and often imitate others, which is counterproductive.
In addition, topics with high certainty are more suitable for the "judgment based on experience" mentioned above. The core is to find the obvious defects of each option and eliminate them.
Thinking when choosing
▐Deterministic : fill in the blanks
Filling in the blanks is to give a context and enter the answer to make the overall sentence more fluent. At work, a more suitable scenario is: the decomposition and alignment of the boss's OKR. The boss has not clearly stated what you want to do, but has given the general direction, and your OKR should also focus on these points.
From the input point of view, it is the context that has the analytical value of the fill-in-the-blank questions. To understand the context better, there are these strategies:
Understand the thinking of people around you: The question must require most people to perceive the correct answer, and it is also something that everyone can do and understand.
Asking more questions means: Unlike the exam, this part of our answer is "open book" and can be repeatedly focused with the upper level.
Multiple regression short-term focus: The topics at different stages are all investigations of knowledge points at this stage, which need to be rolled out in combination with annual and monthly rhythms.
Judging from the output, in order to narrow down the range of answers in the fill-in-the-blank questions, "key expressions" will appear repeatedly. These key expressions will correspond to some production elements such as applications, capabilities, and data in the work. This also reminds us that the answer is given based on the key knowledge points (the responsibilities of the team).
Filling in the blanks is similar to completing a jigsaw puzzle, you need to observe the surrounding plates
▐Low certainty: writing questions
Sometimes we will do some topics with a very large degree of freedom, similar to composition questions, which only limit the meaning of the questions for us. As long as the viewpoint (direction) is correct, there will be basic points. Whether you can get a high score depends on everyone's accumulation, and it is also where you can see the gap between everyone.
In the face of relatively broad propositions (such as: what should be done to improve R&D efficiency?), to write satisfactory work answers, you can learn from some principles of "writing":
Values must be correct: No matter what you do, you must promote positive energy. Corresponding to the work, it needs to conform to the development law of liberating productivity, not for doing, but for packaging.
There must be sufficient arguments: Three examples are often used in the article as arguments to prove that the point of view is correct. On the other hand, when we want to make a judgment, we need to find some deterministic results in advance to endorse the judgment and choice.
Creativity should also be pointed out: there are often novel forms of writing novels and classical Chinese, which can achieve better results. We can choose innovative methods. This direction is encouraged, with extra points or support, but the premise is that it needs to be just right and cannot be separated from the topic (goal).
Just like, it is generally done after multiple-choice questions and fill-in-the-blank questions before doing writing questions. Only by doing a good job of daily work accumulation can you further exert your personal judgment and execution ability on the "writing questions".
Strategy: Shoot for the bull's-eye, score high points repeatedly
Summarize
This article shares my personal judgment and selection experience. At the end, I ask another question: Do we need to make specific judgments every time?
No answer is also an answer: Territorial conflicts often remain deadlocked, based on artificial borders, fought over and over again, with no clear winner or winner.
Others will choose for you: there are many structural transformations, which may not achieve the desired effect. Do you want to make a decision and go back? When you are hesitant, you may receive a business decision: if you do not develop, or even close, the technical system will be nothing.
Use softness to overcome rigidity, don't rush to judge
Indeed, sometimes, only time can give results. But what can see the problem clearly in advance must be: not only understand the current situation of the bottom layer, but also understand the overall trend, and be able to make a very reasonable connection and thinking.
After making a judgment, the subsequent development is often difficult to be controlled by subjective will, and the cost of change is also very high. But make timely adjustments so that your judgments can change and your goals can be adjusted. I think this should be the only coup to get the final result: adjust your judgments and answers dynamically and in a timely manner~
team introduction
We are the Big Taobao Technology-Trading Platform Team. The team is mainly engaged in the delivery of transaction links. In the delivery work, abstraction and construction of horizontal product capabilities (such as: pre-sales, electronic vouchers, etc.), the team focuses on business architecture, DDD and other theories and practices, and is committed to efficient and stable business realization. Access and abstract empowerment.
¤ Extended reading ¤
3DXR Technology | Terminal Technology | Audio and Video Technology
Server Technology | Technical Quality | Data Algorithms