The traditional process of collecting data is generally limited, conscious, structured for data collection, such as the form of questionnaire research. The data you can collect must be the situation you can envision. The data is better structured. A general database Mysql or even Excel will suffice for the data handling process.

2, business layer: modeling and analysis of data

Make

Data analysis models used, such as basic statistics, machine learning, for example, data mining algorithms such as classification, clustering, correlation, prediction, and so on, there is not much difference between the practice of traditional data and big data, for example, banks, communications carriers, retail

Commerce has long been mature use of consumer attributes and behavioral data to identify risks and payment possibilities. But as the volume of data expands dramatically, so do the algorithms.

3. Application layer: interpreting data

The most important aspect of data-guided marketing is interpretation.

Traditionally, after defining the marketing problem, the corresponding data is collected, and then according to the determined modeling or analysis framework, the data is analyzed, the assumptions are verified, and the interpretation is carried out. The scope for interpretation is limited.

And big data provides the possibility of both closed to mining corresponding data for verification based on marketing problems, and open to exploration, to come up with some conclusions that may be completely different from common sense or empirical judgment out. The points of interpretation become very rich.