Data product value: Exists to solve business problems.
No matter what kind of products, are to solve business problems and exist, data products are no exception. There are many kinds of data products. Different data products have different data function components.
Why different data products? Aren’t data products just reports?
Classification of data products: data products are made up of people, usage scenarios and usage methods
Different groups of people have different demands for data. Some of them are numbers, and some of them are analysis. SQL is needed for counting, Excel is needed for analysis, and the data demands of different groups of people constitute data products.
The audience can be divided into TOB/TOC and paid/free. The use methods of the crowd can be divided into the tool category/consulting category. For large enterprises, data products are mostly self-built, and for small enterprises, most of them are purchased in the market.
The above classification is too abstract. The following is a detailed analysis of data products based on actual cases.
TO B: External purchase – operational oriented
TOB refers to the operation and management of enterprises. For example, today I wanted to buy toothpaste on JINGdong. After a fierce operation, I screened the toothpaste, placed an order and waited for the delivery boy to come to my door.
Toothpaste manufacturers want to know their recent business status. TOB refers to the management of store operation information by merchants, that is, the demand of enterprise operators to see data.
The characteristics of TOB data include large amount of data, scattered data and complex and changeable data relations, which are determined by business scenarios. For example, I buy LG toothpaste, let’s talk about the LG company managers look at the data scene, first of all, let’s see what kind of business data, business data can also be called factual data, atomic data without any processing and processing.
In the face of changing analysis demands, huge amount of data, scattered business data, the old way, the operation personnel will import the most detailed business data into Excel, perspective, Vlookup to solve the above analysis problems. As the number of analysis points increases, excel needs more detailed data. It encounters data source updates and relationship changes, and eventually Excel crashes.
The above cases refer to the business operators who are unable to build their own data warehouse data products and conduct data analysis with Excel.
In this context, enterprises need to purchase data products from external markets. At present, well-known data products in the industry include QuickBI of Alibaba, FineBI of Fansoft, Tableau, etc. These data products provide help for users who cannot build data systems. From ODS to DWD to APP, users are responsible for the construction of data warehouse (data source intervention and data association in the figure below). The following figure shows the data functional components of FineBI.
Abstract part of excel capabilities, as long as people understand the business, they can start to analyze data, there is no need to set up a data team, and then to maintain data
TO B: Business built – operational/product oriented
The above example is an example of external acquisition of data products in TOB˙. The drawback of the product is that the original business data needs to be uploaded to the server of the other party. Although the platform guarantees the data security, large Internet enterprises still do not believe it. With enough human resources available, companies should play it by themselves for data security reasons.
Internal data products are done around the business, and the data needs to change with the changes of the business. Keeping up with and leading the business is the measurement standard of internal data products. This kind of data product form is mainly to report form, and divided into different business modules; Visualization is supplemented by orientation of some index observation data; Monitor the content and ensure the health of the business at any time.
Such data products pit lies in the business caliber, business forward-looking, forward-looking data, determines the stability of the data. Abstract ability and logical thinking ability of business determine the stability of product form. If you can’t do that, every day will be hit by business, fell into the pit of check number gu and big cousin.
TO B: Enterprise built – development oriented
Is there something missing from FineBI? Where is the data cleaning, where is the data processing? Of course we have our own data warehouse, processing and cleaning are also digested in our own data warehouse.
58’s slogan, “Everyone’s trusted life service platform”, 6 business lines and thousands of categories, all your needs can be met on 58 platform, is it possible that our own data warehouse is directly connected with 58 so many categories? Obviously not. There is a more detailed division of internal data in 58.
For example, if the department’s business goal is ROI>1, connect as many C-side users and merchants as possible. ROI is revenue divided by cost. We have data exposure, click and cost on the media side, the cost paid by merchants, and the information interested by c-side users. We need to detail some data and summarize some data. How are such huge data and such complicated data interaction realized?
This is the birth of development-oriented data tools, namely big data processing tools, in order to achieve data asset sharing, improve data computing performance, reduce resource consumption and the birth of data products.
Data extraction:
The above cases are all product cases of TOB industry, which can be divided into TOB enterprise purchase, TO enterprise self-construction, which can be divided into operation (business) and development oriented.
These TOB data products have different skill requirements for practitioners. For QuickBI, you need to be very clear about what Excel does. If some of excel’s features can be ported to the platform (I guess), you need to learn not only how to use Excel, how to store data, how to parse data, how to run data, but also a skill behind each function. I am not familiar with this aspect, so I won’t go into details.
The corresponding TOB side operation mainly tests the business ability and logical thinking ability of the product. Personally, it is relatively simple. The corresponding TOB is development-oriented, which needs to understand the development environment, the development data processing process, the development data processing tool, the use method of the tool, including scheduling memory occupation and so on. Guessing and FineBI getting data and creating data sets are the same practitioner skills.
Well, TOB for today, TOC for next time, to be continued.