Analytics Is Only As Good As The Quality Of Data

Analytics and the insights it throws is only as good as the quality of the data, says Sandesh Bhat, Vice President, India Software Lab, IBM. In conversation with, he elaborates on this and the other best practices for enterprises initiating analytics projects or wanting to optimise their existing analytics investments. He also clears the air on some misconceptions surrounding analytics and its understanding among companies.

What are some of the misconceptions that surround the understanding of analytics among Indian companies?

One of the misconceptions that we keep hearing about is that you need to have a data warehouse or BI system for analytics to happen. That is not the case. You can actually have analytics even on an Excel sheet and you can scale it up to doing it on Hadoop. So, it scales to any kind of volume and can go to any kind of variety. Another misunderstanding is that you need to have something big to start on analytics. Data volumes will not be a problem. You don’t have to worry about putting big scale in place before starting on analytics. You can start small and grow in scale as you proceed.

Another misconception is around its complexity. The belief is that you need PhDs and specialised workforce in your organisation for analytics. However, that’s not true. Today, the analytics solutions being developed are simple and user friendly, so that the business users can use analytics rather than only a specific set of people. In fact, the analytics solutions that we are developing in our labs are applicable to the masses, and we are doing so by making it user-friendly and giving it a business context.

What best practices would you suggest for CIOs when initiating the analytics project within their organisation, or for optimising their existing investments into analytics?

Quality of data is the first thing that needs to be ensured before embarking on the analytics journey. Clean and quality data is important so that you can have veracity of data. Always remember that analytics and the insights you get out of it is as good as the quality of data. This is the fundamental starting point. Also, you have to understand what is relevant for you and what to tap into. Once that is established, everything else will fall into place.

Open standards is another aspect that a CIO should look into. Go with open standards in such a way that your analytics technology can be easily integrated into your other systems, such as CRM, for instance.

Most importantly, the business user needs to have a keen interest in seeing the value of analytics in the organisation because the key component of applying analytics is the business knowledge. So, the business actually needs to buy into the usage and value of analytics and its application.

What are some of the primary concerns that CIOs today come up with when they approach you for analytics solutions?

What CIOs are increasingly asking us today is how the power of predictive analytics and sentimental analysis can be used to leverage social media to get insights and intelligence about the customer. They want analytics to help with answers to these questions - What is the customer’s touch point with the company, perception he/she has, what is the brand loyalty emerging, what to do if the customer likes the product. The answers, and those too in real-time, will help take the right decisions, prepare the apt go-to-market strategy, or help plan future investment in specific products, and tailor the products and tailor the campaigns. Speed is going to be critical here – the speed at which you can make decisions before competition, the speed at which you can tailor or change your campaigns or products to be successful in the market.

Till about five years back we didn’t hear these concerns from our customers, but today these are coming up more often. These questions are very much on every C-Suite executive’s mind.

How are the smaller and mid-size companies approaching adoption of analytics?

We have seen different phases with different sets of customers in different categories, wherein the smallest set of customers have started with statistics. SMEs are getting their feet wet by starting off with statistics, wherein they don’t need to put large infrastructure in place. Once they understand the depth of how analytics can be applied and actually start getting the business value out of it, that’s when the business scenarios occur. And, that is when they actually start moving to the next level and apply predictive analytics.

For a lot of our portfolio products we have express versions for small and medium customers. For instance, specifically in analytics, we have our SPSS portfolio which has traditionally been offered with one version called SPSS Statistics. Often we have seen customers downloading and install it, playing and experimenting with it. They come back to the labs with the realisation and understanding of the depth to which analytics can be applied, to then go on to the predictive phase.