“Big Data makes us smarter, not wiser.” – Tim Leberecht.
The term ‘Big Data’ got introduced in 1940s. Companies around the world have put in ceaseless efforts to explore its potential. The global tech giants have massively increased their spending on leveraging big data technologies. This trend got quickly replicated among major industry players.
As a result, according to a forecast issues by research firm(IDC), big data technology and services will grow at a CAGR of 23 percent through 2019. The big data annual spending will reach $48.6 billion in 2019.
That’s how big data services are being accepted worldwide!
Big Data has given a ‘ray of hope’ to companies and enabled them to make use of data of any size and volume. The bits of data collected via our mobile phones, GPS, sensors devices is no longer useless. Every bit of data collected gets collected and processed to derive useful insights about us (customers).
Amidst, the increasing benefits of Big Data, people fail to see the things it “can’t” do. This was surprising for me too. But soon I realized, Big data always complements business intuition but can’t ever replace it.
In this article, I present you my research of last 7 days. My crazy curiosity led me here. I just couldn’t digest the fact that big data has all it takes for a company to succeed. Big Data is ‘capable’ of many things. But ‘incapable’ too.
Note: My thoughts are not exhaustive but an attempt to put a framework. Feel free to add your perspectives in the comments section below.
‘Tiny’ Exercise on ‘Big’ Data
This exercise will prepare us for the future. We must know about things which are yet to come. Hence, If you are reading this, I invite you to take a stab on this question. You just need to write(I’ve already shared the answer though):
” 5 things which Big Data can do” & “5 things big data can never do”.
For instance, if I conclude using a logic that X is not possible using any technology platform with Big Data. I will simply eliminate all the business problems which are related to X. Getting it?
Below is my list. If you disagree with any of the element in my list, please justify! I will love to modify this list with time. Let’s start with a short note on my ideology on using business intuition and business analytics
The 80:20 rule
The rule says,
“80% of time is spent in creating stories from past data, and 20% of time is spent in connecting those stories with current business”
Explanation: I believe no analytical insights are useful until they are in sync with business intuition. Agree ? Moreover, with time, the data driven component has grown exponentially. Companies are now flooding with data. But would that really make a difference? No!
Companies must realize, a right combination of successful business analytics over required business intuition is in ratio of 80:20.
If we can build a story using analytics which describes the past to predict future expectations 80% of time, we need to invest 20% of time thinking, how this information is useful to business. We must think of ways which can change our future and meet our broader business objectives. This requires a strong business understanding and sound knowledge of business rules.
The 20% component in this rule is non-replaceable. Hence, humans intervention is required for solving this 20% and possibly no machine can make up for it. Not even artificial intelligence. Because, humans think in a non-defined fashion which leads to creativity. Creativity is what I believe no machine can bring to the table. My list is inspired though this rule.
5 things Big Data ‘CAN’ do
- Diagnostic analysis : We do it every day. Machines are superb at this. Once an event occurs, we find interested in seeking its causes. For example, suppose, there is a sand storm in Desert A. We have all the parameters in different regions of Desert A: Temperature, Pressure, Camels , Roads, # Cars etc. If we can relate the parameters to the sand storm in that area, if we know a few causal relationship, we can possibly avoid sand storms. Imagine how powerful big data is!
- Predictive analysis : We do this often. Predictive analysis is in our DNA! For instance, we have a hotel chain all over the globe. Now we need to find which of these hotels will not meet their target sales. And if we know it, we can focus our efforts on these hotels. This becomes a classic problem of predictive analysis.
- Find relation between unknown elements/events : I love this part of the analysis. Let’s say, number of sales employees has literally no relation with sales. Then possibly you can reduce the number of sales employees if that does not do any other loss.
- Prescriptive analysis : This is the future of analytics. Let’s say we are trying to predict a terrorist attack in a popular destination and possible strategy to safely move people. To make this prediction, you need to make a series of prediction, which might involve predicting number of tourist then number of tourist in that location, then predicting area which can be affected by a blast etc. etc.
- Monitoring an event as it happens : Majority of people in the industry work on monitoring of events. For instance, you need to monitor the response of a campaign and find segments which responded most and least. These analysis become crucial for running a business.
5 things Big Data ‘CANNOT’ do
- Predict a definitive future : We can reach higher 90s in terms of accuracy using sophisticated machine learning tools. However, you never reach 100% in accuracy. If we could do that, I could have told you exactly whom to target and achieved a 100% response rate every time. But sadly, never going to happen!
- Imputation of new data source : Imputation takes most of the time in any analysis. And I believe this is where you bring in your creativity and business understanding. Possibly, one of the most boring piece in your analysis which you will never get rid of.
- Find a creative solution to a business problem : Creativity is one thing which will always be a patent of human race. No machine can ever find a creative solution to a problem. This is because even AI is coded by human and creativity is never learned through algorithms.
- Find solution to a not so well-defined problem : The biggest challenge in analytics is to shape an analytics problem from a business problem. If you can do this well, you are on the right track to become a analytics superstar. This piece of role is something machines can never take away from you. For instance, your business problem is to manage attrition. Now until unless you define responders, the time wondows etc. No predictive algorithm can help you.
- Data management/Simplify data for a new data source : With growing data, management of data is becoming difficult. We are progressing with different types of data structures for different types of data. For instance, graph data might be well suited for network analysis but is useless for campaign data. This piece of information is again which machine cannot analyze.
I believe this article will reach its true potential if people start trying out the exercise in this article. Try thinking in a more holistic view where you can see what machine cannot do ever. For instance, my starting point was the 80:20 rule that machine cannot bring creativity. This starting point helped me think of what are the pieces which needs creativity in the process of analytics.
What is your list of do/don’t? Did you like this post? Write your comments in the box below.