Central analyses need central data
Value added from big data is often achieved for a business only when most or even all sources of enterprise data are integrated. Initially it is not the amount of data that is significant, but rather the informational content of the data in relation to concrete requirements.
At the beginning, a clear use case and business case must be written in order to define what the company expects to achieve through big data integration. Of course, data quality plays a key role when integrating many different sources of data. Thus data cleansing and data stewardship must be integral elements of the big data project from the very start.
Experience shows that big data technologies already deliver value added with smaller volumes of data – it doesn’t have to be in the terabyte range. Thanks to the special features of this technology, it is suitable for analyzing smaller data volumes at first, with plenty of headroom to cope with larger volumes in the future as data streams continue to grow.
The wrong question
Data science involves a complex mix of industry know-how, mathematics, statistics and programming skills. These aspects form the basis for the requirements of a big data solution if it is to deliver the results that a business enterprise expects. This is also the reason why the profession of the data scientist has come into being through the proliferation of big data technology. When selecting a qualified candidate for the position of data scientist, for a project or for employment in a business enterprise, it is essential that the candidate has both mathematical and technical background know-how related to big data. Another essential qualification is industry know-how. Specific industry knowledge is absolutely essential when developing a big data use case that matches the business model of an enterprise, and only then can value added be ensured because big data is fundamentally a universal technology that is adapted to specific individual business requirements.
At the same time, big data can only succeed if various experts within an enterprise collaborate effectively – management, specialist departments, sales, IT and data scientists, who often work in the specialist departments but in some companies are part of IT, since they are responsible for the data modeling required for the big data solution. Big data is still a “young” technology. Data scientists and other big data experts are not always readily available today because big data is still a new technology. Enterprises must be aware of this during project planning and when selecting partners.
Good change management is even more important for big data projects than for other IT projects. In fact, change management is frequently the key to success. This does not mean that technical issues have to arise in the project, or that the concept needs to be adapted constantly. But as soon as the first analyses from the big data system are available, new ideas arise quickly as to what can still be analyzed or which data should be integrated. This is where professional change management and release management are required to respond to such developments and cope with the complexity of realizing big data in an enterprise.
Risks can be mitigated by working with an experienced partner. Good change management, which evolves from technical or business aspects, can be provided by an experienced partner involved in the big data project.