What large numbers of people do companies need? How to interpret large data?

If you were a bakery boss, how would you solve the following problems:

Question 1. The capacity of the oven is limited, what kind of bread should be produced?

Question 2. What kind of advertising is more effective if you want to market a bakery online?

Question 3. How to judge from the bread, the baker has not cut corners?

In the language of management, these are the "estimated demand allocation capacity" and "product quality control" issues, which are often encountered in the workplace. These questions, at first glance, seem to have nothing to do with statistics, but in fact, every problem can be applied to the concept and method of statistics to conceive the solution.

If you have statistical thinking, consider the following questions:

1. List the most popular kinds of bread, giving priority to the production of star products.

The key of question 1 is to find out the star commodity, which needs to count the total turnover of the bread, and then calculate the relative proportion of each bread to the total turnover, and give priority to produce the product portfolio which can cover 70% turnover. This uses the statistical number allocation table and histogram, which is also called ABC analysis (see P.65).

2. Write two kinds of copy, advertising for a period of time to see how effective.

Question 2 to compare the effectiveness of advertising, the best way is to use statistical randomized controlled trials, so that two kinds of advertising randomly appear, after a period of time, to observe what kind of advertising effect is better, and then a large range of use effect better advertising (see p.85).

3. Check a few bread, scales to see if the weight gap is too large.

To solve the problem 3, you need to know the average weight of the bread, and then sample the bread to see whether the weight of the bread is the normal distribution of the bell-shaped curve? If you deviate from the curve, you may imply that there is a problem with the noodle. (See P.65)

As a collection, integration, analysis of data science, statistics are also commonly used in enterprise analysis tools.

E-commerce Web site Amazon can when you browse, it is recommended that the people who bought the book buy these books, too, with a correlation analysis; President Barack Obama's campaign team knows which pages will allow voters to increase their contributions, using randomised controlled trials Market research can deduce the view of the whole market with the opinions of a few people, based on the survey of the Yuan.

Dartmouth College professor Charles. Whalen (Charles Wheelan), in the book "13 and 1/2 lessons of smart statistics," enumerates the purposes of statistical studies, including:

Analyze the data and make a summary of the information;

make better decisions;

Identify patterns that enhance the effect of doing everything;

Evaluate the effectiveness of policies, programs, and other innovative issues.

Does that sound familiar? Is it not the manager's job to make decisions, find more effective ways to do things, and evaluate the usefulness of the paintings?

Sinei, author of statistics, the most powerful commercial weapon, points out that statistics can help business people think about the following 3 questions:

Which factor changes can increase profitability?

Is it possible to take action that triggers this change?

If the action that triggers the change is feasible, will it cost more than the gains?

Fongtingyu, associate professor of business information studies at Tama University, believes that business people learn statistics not only to develop the ability to inspire data, but also to use statistical techniques to assess data and to personally experience the process of "building hypotheses, testing, interpreting results, rethinking".

Toyota explained that in the marketplace, it is important to find the relationship between the "results" and the resulting "plateau cause", which is the process of repeating validation in a digital heap, which is the ability to build hypotheses that require constant practice. And once proficient at a certain stage, the person who is good at statistics, even if does not depend on the data analysis, can also want to draw the good hypothesis.

Thanks to internet popularization and technological advances, the era of big data is coming.

The big data means that companies can collect huge amounts of data and have the ability to analyze the data, and it is a popular trend to count the science that is the meaning of data collection. Therefore, in order to make good use of large data, managers need to have statistical literacy.

What big data can do can be told from a story. One day, the department store sent a baby-supplies catalogue to your unmarried daughter. You think this is an insult to your daughter, so angrily call the department store customer service to complain and ask them not to do it again. You complain, but see the daughter of fun to turn the record, she has been pregnant.

Department stores know better than dad that the news of their daughter's pregnancy is not black humor but the real events of target department stores. How did Target know the customer was pregnant? It's not magical, because the company has a mother's gift registry, which allows mothers to register their baby gifts. Since this list is equal to knowing the list of pregnant customers, target set up a shopping model for pregnant women (a list of items to be purchased during pregnancy) based on the list's consumption records, and then use the model to find customers with similar consumption patterns and marketing related products to them, to achieve "foresight" results.

Using this approach, target has found 30% of its marketing targets, and has been quite successful in terms of reducing marketing costs and improving marketing precision.

What target does is to remit a large amount of data (target all consumption records) to the information that decision-makers interpret, to enhance the ability of enterprises and organizations to predict, and the fundamental knowledge behind big data is statistics. In this case, target can find the relationship between "pregnant" and "shopping list" by statistical regression analysis (see p.88).

Carried, professor of science and technology research at the apex of the University of Tokyo, said that enterprises need to use large data, needs 3 kinds of talent: the first is the data of IT experts, the second is the analysis of data analysis personnel, which is the use of data managers.

The paddy field emphasizes that statistics are not the only way to analyze the data, but to speculate on how to influence customer behavior, and to formulate it as a specific business plan, is the key to doing so.

From the point of view of enterprise strategy, it is the management work to instruct the research data according to the management policy, and to translate the result of analysis into actual action, also belong to the domain of management. Managers have to decide how to analyze the data, but also have the ability to rely on the collation of information, change action.

After the change, the manager may change the way of data analysis again because of the action change, and form the positive loop which makes the data analysis more and more accurate, which is the utility of statistical management.