Advantages And Disadvantages Of Big Data

DISADVANTAGES:

Even though Big Data has the potential to be a great opportunity, many are still doubtful about the real benefit big data bring to business, as results obtained from its usage are often limited.
A research from McKinsey Global Institute 21 asked to some companies making use of big data and data analytics how much their business improved in cost or revenue since they started using it and they found out that three-quarters of them replied it was less than one percent.
Also, Big Data may bring to other problems:
- Unstructured data; as said in the first chapter, a big part of big data includes also unstructured data, a kind of data which is hard to store and analyse;
- Misleading correlation patterns (as they are not causation

- Organization design changing due to different logistical approach; in fact, for an appropriate use of big data, companies need to modify their entire approach to how they previously used to analyse data; this is because big data usually comes constantly rather than just periodically as open data or other sources of data do. Also, the company will need new softwares which must be able to handle also real-time analysis and it will have to hire qualified data scientist to manage big data appropriately (and, in case the company is willing to protect customers privacy also qualified data protection employers are needed).
- Noises and variability; another problems in analysing big data is about understanding which datas are relevant data and which are not. In fact, having a large amount of datas makes it possible that some datas are incorrect or misleading. This is a problem for understanding their veracity and so their quality might be low.
Variability of data is another of the 3 V’s which can cause problems in the application of big data: data can change even after few time after its creation and companies might not to be updated about

Also, softwares and algorithms are set to segment customers and this is believed that it might bring to discriminate based on ethnic, background, health condition... and so on.
An example of this is 23 the study of Latanya Sweeney (former director of data privacy lab at Harvard University) about Google AdSense in which it was studied the probability of ad suggestion of words such as “arrested” (without any related arrest statistic database) through the use of names typically assigned at birth to different ethnics and it was found that names usually given to black people were more easily associated to ad suggestions of arrest than white’s ones, exposing that way to racial discriminations. Even though it’s not really clear the reason why this happen, it is possible that many tools are biased according to what the majority of people search the most with the associated word and so the algorithm can lead to involuntary discriminations (as it is known with Google autocomplete which can sometimes give inappropriate