Are you able to define big data? What do you understand about large data sets and big data analysis? And more importantly, how do you work with big data?
These are pertinent questions that most C-level executives are concerned with in an age of big data. The definition of big data aside, both big data applications and big data analytics are issues that many businesses are grappling with. Dan Bolland, Global Head of Retail Analytics, Barclays, and the first speaker for the Best Practices Track for Interact 2013, provided some key insights for consideration in his session that outlined a pragmatic, incremental approach to big data. Dan drilled down to the operational basics of big data, providing a starting point for businesses wishing to delve deeper into their big data sets.
The 4 Vs of Data – Value, Volume, Variety, Velocity
The value of big data is derived from being able to capture and manage the data. Data can come from different sources and there are different aspects of it to deal with:
- Volume: Records, transactions, multi-source queries
- Variety : Channel, internal and external data, structured and unstructured data
- Velocity: Real time, batch
There are multiple sources of data and they can be quantitative and structured, or qualitative and unstructured, such as in the case of social media. Here are some examples:
- Social Media
- Call Centre
- Direct Mail
- Mobile ATM
How Do You Use Big Data?
Big data can enrich current reports or improve management. Known metrics can be updated on a given frequency, or businesses can allow exploratory analysis to create insight. Here’s the simple three-step process:
- Collate multi-source data
- Find trends and outliers
- Create test to validate discovery
The results can be fed into regular reports or used as the basis to take management action, or insight-driven action. Think about the actions that enhance relationships with the client or customer… must they be the end result of information collected?
It All Goes Back To Manageability
The world is changing so quickly that what may be true this year, is no longer true a couple of years down the road. It is important to test and learn – consider data visualization. When it comes to testing the art of question creation is key. Clear questions allow you to ascertain customer needs as such it crucial to be clear on what you are trying to test, be it frequency, geography, channel, products or patterns. Specific exam questions are required to set up pragmatically achievable results.
While you can choose to work with big data inhouse or outsource it, both approaches comes with their own set of problems – when it comes to outsourcing, your vendor might not understand your business – it all goes back to manageability.
Working in the finance industry? Check out our webinar on Cross-Channel Banking presented by Forrester Research!