Reading Notes --> "Lean Data Analysis" Part One: Stop deceiving yourself

Chapter One We are all lying

1.1 Why are entrepreneurs "lying"?

• Able to identify entrepreneurial beliefs
• Can increase the courage to face the downturn in entrepreneurship

1.2 Why do we need data analysis?

• Verify entrepreneurial intuition or inspiration
• Achieve Lean Startup
• Questioning the reality distortion field

1.3 involved concepts

Minimum Viabel Product (Minimum Viabel Product): Refers to the minimum product that is sufficient to create the mechanism you advocate for the market

• Not for mass production
• Twenty In order to verify ideas in a short time and low cost

1.4 Book summary

Data analysis enlightenment: When you think you have found an idea worth trying, first think about how to quickly complete the test with minimal investment. Then define success in advance.

Chapter Two Entrepreneurship Scoreboard

Data analysis is required for key indicators tracking !

2.1 What is a good data indicator?

2.1.1 Measurement criteria

1. Comparative
2. easy to understand
3. Is a ratio
4. Have an idea to change behavior

2.1.2 Five-point measure of correctness of indicators

2.1.2.1 Qualitative and quantitative indicators

Qualitative indicators: data that are unstructured, empirical, revealing, and difficult to classify.

• Qualitative data incorporates subjective factors
• Data is difficult to quantify
• It is suitable for entrepreneurial exploration period, mainly to answer the "why" question.

Quantitative indicators: are indicators that involve numerical values ​​and statistical data for tracking and measurement.

• Quantitative data exclude subjective factors
• Scientific, easy to categorize, extrapolate and place data tables
• Suitable for daily normal operations, answering "what" and "how much" questions

2.1.2.2 Vanity indicators and actionable indicators

Vanity indicators: data indicators that cannot influence and change actions.

• Can't guide action, fail to achieve the purpose of data-driven decision-making
• For example, the total number of registered users increases monotonically over time, unable to convey user behavior information, and unable to tap useful value

Actionable indicators: indicators that can help determine action plans and guide business behavior

• Help reveal the value of information, point out the direction, and improve business models
• For example, the percentage of active users to the total number of users (proportion of active users) explains user engagement. For another example, the number of new users per unit time (or the growth rate of new users)

Vanity indicators to watch out for:

Vanity index	Replaceable indicators
Click volume	Clickers
Page views (PV value)	Number of visitors
Views	Number of visiting users
Number of unique visitors	With this indicator alone, it is impossible to know the reason for the customer's stagnation or leaving.
Number of fans, friends, likes	Fan activity is more meaningful
Time on site (time on site)/number of pages	It is impossible to see the behavior of customers staying, and cannot replace user engagement or activity.
Number of user email addresses collected	Unable to determine whether the mailbox is commonly used; you can test it by sending a test email.
Downloads	The amount of activation after download, the amount of account creation, etc.

2.1.2.3 Exploratory indicators and reporting indicators

Things in the world can be divided into the following categories: what we know we know, what we know we don’t know; in addition, there are things we don’t know we know, and things we don’t know we don’t know. ——Former US Secretary of Defense, Donald Rumsfeld

For these four types of things, the important actions are:

1. Question the first type of things, check the authenticity of the data, the reasonableness of the hypothesis, and ensure that you are not deceiving yourself
2. Find answers to the second category of things, optimize processes and behavior patterns, and improve efficiency
3. Evaluate the feasibility of the third type of things, verify ideas, learn new knowledge, make business predictions, and provide data support for decision-making
4. Explore the fourth category of things, broaden the business map, open up the blue ocean of the industry, and look for opportunities for future development

2.1.2.4 Foresight and hindsight indicators

These two indicators and indicators have greater practical significance for entrepreneurship, but the use period of indicators and the problems they solve are different.

Foresight indicators: can be used to predict the future.

• For example, the number of potential customers prompted by the "sales funnel" can predict customer acquisition ability and future development intensity.
• Applicable to the initial stage of entrepreneurship with insufficient data, indicators can be used as a benchmark for later development .
• Suitable for cohort analysis, group comparison, etc.

Hindsight indicators: can indicate the existence of problems.

• Indicators have the meaning of remedying the situation. For example, when users are lost, users have already lost.
• User loss is inevitable. The focus is on tracking data indicators such as account cancellation and product returns, accurately discovering problems, summing up experience, iterating and adjusting products.

2.1.2.5 Correlation indicators and causality indicators

Correlation indicators: If two indicators change together, they are related.

• Finding correlation helps predict the future
• Not all correlations have deep digging value

Causality indicators: If one indicator causes a change in another indicator, then they are mutually causal.

• Causality helps change the future
• There can be many factors that cause a certain result, so part of the cause and effect has high value
• Control variable testing can be performed on relevant variables
• When the sample size is large, the test is more reliable; when the sample size is small, the test design needs to be simplified as much as possible (the test method is provided in Section 4 of this chapter)

2.3 Moving target

The goal setting of the entrepreneurial process is a dynamic process. Under the premise of seeking truth from facts, it is feasible to adjust the goals and key indicators.
Early days, pay attention to uncover and examine assumptions and actual user behavior nuances, adjust goals and targets based on this.

2.3.1 The meaning of moving targets

1. Through direct communication with customers, it helps to truly understand customers.
2. Test the practicability of the goals and indicators set in the early stage of entrepreneurship.
3. It helps to adjust the "success" goal as soon as possible and avoid getting lost.
4. On the premise of seeking truth from facts, it is reasonable and necessary to reduce or increase the threshold of indicators.

2.4 Test analysis method

Testing is the soul of lean data analysis!

2.4.1 Market Segmentation

Segmented market: It is a group of people who have certain common characteristics.

• Look for differences in behavior of different groups
• Through testing and investigation to find the reason behind.
• Applicable to any industry, any form of marketing behavior analysis

2.4.2 Synchronous group analysis

Cohort analysis: Compare the changes of similar groups over time.

• It is a longitudinal study, data is collected according to the user life cycle
• Divided according to the time when the product was used and customer behavior
• Analyze the differences in the key indicators of users who have common characteristics but start to use the product at different times, so as to determine whether the indicator performance is getting better and better, and to determine whether product iteration is meaningful for increasing value.
• It helps to observe the behavior patterns of users at different stages of the life cycle.
• It is suitable for the analysis of data indicators such as revenue, customer churn rate, word-of-mouth virus transmission, and customer support costs.

2.4.3 A/B test (A/B test)

A/B test: Assuming other conditions remain unchanged, only consider the impact of a certain attribute that may affect the experience on the tested user.

• It is a horizontal study used to compare the experience differences of different tested groups at the same time.
• Can be used for detailed testing to help focus on key steps and assumptions
• The larger the sample size and the larger the flow, the faster and more effective the test feedback.

2.4.4 Multivariate analysis

Multivariate analysis: test multiple attributes at the same time

• For companies with small companies, small user traffic, and small sample sizes, the A/B test that tests a certain detail alone will slow down the company's progress.
• The factors that affect user behavior are often diverse, and multiple factors can be tested simultaneously.
  

2.5 Lean data analysis cycle

Chapter 3 To whom do you dedicate your life

3.1 Lean Canvas

A non-existent market will not care how smart you are. ——Mark Anderson

• This is a normative framework for choosing and controlling entrepreneurship

3.2 Looking for a career that can dedicate your life to it

Three criteria for finding a career that can dedicate your life: what you are good at doing, what you hope to do, and what you can make money. --Bad Kaderming

• Add these thoughts outside of the lean canvas to be more humane
• Remember to ask yourself, do you really want to do this?
• Pay attention to needs, abilities and desires!

3.2.1 Development planning inspiration

1. Can I do it the way I like?
   It is necessary to comprehensively evaluate various factors such as market demand, own ability, and desire for success.
   Never consider areas where you have no advantage, otherwise you will struggle!

2. Do you really like it?
   Are you convinced of the road to more struggle?
   Is there enough desire to succeed?
   Whether you can achieve the success you expect.

3. can earn money?
   This is the cost of broadening the road in the future.
   This is the actual need to persist.
   This is the essential question of life and survival.

In short, ask yourself , is it worth the effort?

Chapter 4 Data-oriented and information acquisition through data

Humans provide inspiration, machines are responsible for verification!

4.1 Intuition vs optimization

4.1.1 Criticism of "data-oriented"

1. Entrepreneurship should not be a data effort
2. Abuse of data to value partial optimization, and then ignore the overall situation
3. Seemingly irrelevant data ignored by data analysis algorithms may lead to missed growth opportunities
4. Algorithm blindly optimize data products, which may lead to bad results

4.1.2 Data optimization

The core of optimization is to find the maximum and minimum values ​​of a given function.

Progressive changes can achieve local optima, and innovation can lead to global shuffling.

• We must insist on using human judgment to reconcile the automatic optimization of the machine
• There is no denying the importance of data for hypothesis testing
• Data should be used as a tool, combined with human reflection, to have a new tree

4.2 The way of thinking of a data scientist (mode)

10 data traps that entrepreneurs need to avoid:-Monica Rogatti

Serial number	Data trap	How to avoid
1	Assuming no noise in the data	Check data validity and practicability; carefully conduct data preprocessing.
2	Forget about normalization	Find a reference point, unify the unit or classify
3	Eliminate abnormal points	For abnormal points, further investigation is needed, instead of simply deleting them
4	Including outliers	The abnormal points shall be further investigated, and the model adaptability shall be fully considered to determine whether to exclude or delete
5	Ignore seasonality	When looking for a pattern, fully consider the impact of different time periods on the data
6	Talk about growth regardless of base	Base and benchmark are both critical, this is the basis for comparison
7	Data vomiting	Clarify the goals of data analysis and filter useful data
8	Indicators of misrepresenting military intelligence	In order to effectively monitor the index setting threshold, this approach is correct. However, the sensitivity of indicators needs to be fully considered. Too frequent alarms will make people slow to alarm abnormalities.
9	"Not collected here" syndrome	You can try to analyze data from different sources together, and perhaps discover interesting new ideas, and through experiments, come up with decision-making plans that affect growth.
10	Pay attention to noise	Put the vanity indicator aside, step back, and widen your sight; stand taller to increase your viewing angle.

4.3 Lean Startup and Big Vision

4.3.1 Minimal feasibility and big vision

1. For entrepreneurship, it is easy to get stuck in a situation where data analysis is easy to go back and forth; ignoring data analysis is easy to slap your head and trust the horse.
2. The concept of lean entrepreneurship makes people feel that entrepreneurship is simple and the threshold is low.
3. Entrepreneurship without a big vision, the goal is not firm enough, and the steps are not strong enough.

4.3.2 How to make dreams compatible with entrepreneurship?

Take lean startup as the only way to achieve entrepreneurial vision

• Rather than building a product, it’s better to build a tool that can help people know "what to build"
• Cognition is at the top of the pyramid of lean entrepreneurship-helping divergent thinking, active exploration, experimentation and verification
• In addition, under the premise of understanding and innovation, the process of lean entrepreneurship is realized through the cycle of "development-testing-cognition"

4.3.3 Lean, but not small

• The essence of Lean is to look at the big picture and focus on details
• Encourage questioning everything you see while expanding your horizons
• Like an eagle, flying high, but with sharp eyes enough to see the rabbit hidden in the grass one kilometer away, in order to seize all the seemingly small but important development opportunities

Summary

1. Although this book regards data analysis as an excellent tool for lean entrepreneurship, it introduces data analysis for the purpose of guiding entrepreneurship. The views in the book are of great significance for guiding data analysis work and personal data analysis career planning.
2. The third chapter proposes the lean canvas method, which can be used by college students who are new to society or those who wish to change careers for reference to clarify their ideas and make career planning.
3. Work is a long-term entrepreneurship based on the development of life, so you must carefully consider the three points of what you are good at, the career you hope to engage in and whether you make money. (On this point, a special article will be written later)