Analysis of two choices of enterprises on the way of large data nuggets

For most companies, there are two options for how to turn data into profit on the big data nuggets: data-driven processes and data-driven products.

Big data is now a word that has become a secret. A company like Amazon, down to a small startup, can have a few g of data a day. A photo-sharing site like Instagram can produce 500T of data a day easily. Many CEOs of companies ask a question: "Well, now that I have so much data, what do I do next?"

A man, if he is only standing on the land of gold and not digging, will not be rich. Similarly, having a lot of data doesn't mean your business can succeed. The industry is successful in companies like Amazon, Netflix, who can make better use of data than rivals. Otherwise, you can only stare at a bunch of hadoop clusters without knowing what to do. However, if you can make good use of your data, you can be a step ahead in the competition.

So how can we turn the data into profit? For most companies, there are two choices, data-driven processes and data-driven products.

Data-driven business processes:

Traditional data analysts use Excel or write SQL statements for specific queries. Now, these are far from enough. Today's data scientists need to understand the small data age and the large data age of various tools, including traditional business intelligence tools, query language, statistics, and even machine learning.

Good data scientists can help businesses from analyzing products, such as which products are popular, why, and what users dislike (as Zynga does), to build predictive models and analyze future trends to help current decisions (as Wal-Mart Labs do).

Here are some specific examples:

1 If you are a marketing software that is a service (SaaS) application, data scientists can help you analyze the characteristics of high-end customers, such as their transformation channels, their basic commonalities (age, gender, income levels, geography, etc.), and their use of your application special way. In this way, you can be more targeted to design your product features, the introduction of targeted advertising, optimize the marketing channels, so as to improve your profitability.

2 data scientists can help you optimize your entire price system by analyzing the impact of the price of a particular product on other categories of product sales.

3 Data scientists can build an accurate prediction model based on historical data. As a department store target, for example, it is possible to determine which customers are pregnant women or, like some insurance companies, to predict which potential customers to consult are most likely to be converted into customers.

4 data scientists can also make it easier for you to use existing data to analyze operational results. For example, data scientists will advise you to correlate your marketing data with Web site access logs and transaction data to measure the effectiveness of marketing campaigns.

Data-driven Products:

In addition to data-driven processes, you can also use the data to enrich the functionality of the product. Some companies, but also the data packaging into a special product to sell.

Twitter, for example, is not a data product, but by authorizing companies such as DataSift to use its data, companies like DataSift use Twitter's data to make data products for businesses to help companies make better use of social media. There are also media companies that package the data that viewers view and sell it to some channels or content-producing companies.

However, as opposed to selling data directly to revenue, more companies are using data to improve existing products, making them more efficient, more intelligent and more user-specific, thus directly or indirectly increasing revenue.

Here are some practical examples to illustrate how the data makes the product smarter and more in line with the user's needs:

1 in order to improve the click rate of advertising platform, advertising platform through the analysis of advertising broadcast media, advertising itself, as well as user behavior. Show ads to the most appropriate users.

2 e-commerce website, through the recommendation system of data analysis and machine learning, improve the user's recommendation of the purchase possibilities.

3 Media site through the analysis of user characteristics, to different users to show different content pages, improve the user's stay in the site time, so as to obtain more advertising revenue.

4 Video publishing platform through the analysis of the user's viewing and interactive behavior, to the video producers on the user preferences of various feedback, so as to create a more satisfying user preferences video. This is an example of an indirect increase in revenue. Through data analysis, to improve the popularity of the video platform.

How companies should begin to act

So as a business, how should we start to prepare to turn cold data into golden money? Here are some suggestions:

1 Save as many kinds of data as possible. Today, the cost of storage is not a factor to consider. Remember, no good analysis, no data is not. There are a lot of data, even if there is no way to analyze it, try to store them for future analysis. Many companies ignore this. In fact, a lot of data can be stored in the original format, including transaction data, user behavior, log files, user-generated content, sensor data, etc., in short, you can have the data, first save. Will always be useful in the future.

2 Find a data scientist: If you are a small company, you may need to find a data scientist to join, or one of the team needs to be a data scientist. If you manage a large company, then you may need a team of data scientists. Data scientists can be cultured from within. A good business analyst or anyone with a strong business intelligence or database background can become a data scientist. You need to equip data scientists with the right tools and access to different data from the company so that he can perform data analysis, data mining, business intelligence analysis, and data production. A good data scientist can help you improve your productivity and help you make better use of the various data generated within your company.

3 Data production: For any company that has specific data, it should consider the product of this data. In fact, any company with desktop, mobile, network or server applications has its own unique data. Companies in the advertising and retailing industries have increased their revenues by billions of of billions of dollars in data products.

For example, if you are a business-to-business software that serves as a service company, the service of providing your customers with self-service reporting is the simplest example of a data product. If you are an E-commerce site, using the data to provide users with recommendations can increase your income, if you have a mobile application, then consider how to make your application more intelligent will bring a better user experience and revenue. There is a data scientist to consider how data production is the first step, and ultimately, the enterprise still needs to put resources into real implementation.

4 Data-oriented leadership: Big data is not just about data, it's more about how to use data to drive workflows and optimize product functionality. All this requires the managers of the enterprise to use a data-oriented way to lead the enterprise and promote the large data of the enterprise. 21st century is the century of the large data. If the enterprise cannot transform smoothly under the trend of data-oriented, it is likely to be defeated by competitors.