I believe every data-set has at least one story to tell, usually there are many. Big data analytics is the process of examining data to hear that story. Sometimes the simple frequency tables or cross-tabs are good enough to find the story. Other times one has to run slightly more complex processes like regression, time series analysis etc.
The story that a data-set tells is often referred to as an insight. Insights are used to achieve various goals e.g. increase revenue, reduce costs, etc. Depending on the aim, data should be collected for a long enough time period. At least a couple years data is necessary so that the same month could be compared to each other and one year can be compared to the next. The more the data, the better and more robust insights are achieved.
We use big data analytics to solve a variety for problems including (but not limited to):
- Predicting Attrition – Probability of closing account, probability of not renewing a contact, etc.
- Propensity modelling – Probability of a customer (or potential customer) to take up a certain product.
- Fraud identification.
- Determining which factors have the highest effect on revenue, cost or profit.
- Capacity planning.
Social media has opened new doors for analytics. There is a lot of data out there and thus, lots of insights can be pulled. Banks have even started using social media for customers’ credit ratings. Our tool “Tracx” allows us to determine the sentiment about a brand. The internal data can be used to find the company’s spend on product (e.g. innovation, R&D, packaging), marketing (channels), delivery, etc. External data (from Tracx) can then be used to see the effect on brand sentiment (image) if there is a change in any of the internal variables.
Once we have found the solution to the problem we need to implement it. Analytics without implementation is wasted effort. One of my mentors always said “don’t do analytics for the sake of analytics”. We ensure that our clients implement the solution the way it is supposed to.
The final piece in the cycle is monitoring to ensure that the solution is up to date over time.
Written by Adnan Barakzai – Head of Analytics