Indicus Analytics Pvt. Ltd., widely regarded as India’s premier economics research firm, has been providing analytics and economic research services for a broad range of clients. Analytics India Magazine talks with Laveesh Bhandari on how analytics is being used at Indicus as well as the current trend in this space.
[dropcap style=”1″ size=”2″]AIM[/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization?
[dropcap style=”1″ size=”2″]LB[/dropcap]Laveesh Bhandari: As is true of anything, the better the understanding of the domain, the more analytics can get you by way of insights. Good quality analytics requires
(a) very disciplined thought process
(b) frequent interactions with clients/stakeholders
(c) a carefully thought through framing of the problem statement.
Once these are done, analytics requires a good combination of vertical and horizontal thought process – the former to delve deep into the data, and the latter to enable unusual insights.
As yet we have focused on the Indian market and have only recently started to work on cross-border clients. Moreover, for us research and analytics go hand in hand. Without research good modeling skills are difficult to develop and sustain. We therefore feel that pure analytics firms will eventually be reduced to body-shopping sheds if they shy away from developing a strong research foundation.
AIM: Please brief us about some business solutions you work on and how you derive value out of it.
LB: We tend to specialize in problems where the environment is data poor or where the data are highly flawed. Take for instance data on Indian demography where we simply do not know adequately the economic profile of Indian households – all data sources including those of the government are flawed, most sources are also poorly or inadequately sampled, there are serious issues with the quality of data collected, and there is also the problem of dated-ness. But if you have enough data from many different sources – then using certain data fusion techniques we can use the positives of each data source to get the right picture.
Indicus for instance has created some of the most granular databases on Indian demography and economy that even the Government of India is using. What came out of an analytics problem was later made into an ongoing process and productized. We have a new high growth data products and solutions stream that initially was a by-product of our analytics business.
Since we focus on the Indian market, we frequently help firms undertake surveys to fill in the data gaps. The combination of firms’ internal data, data out there in the public domain, and sometimes specific data from a survey can create a powerful triad of information sources. Together this allows a far superior set of analytics driven insights for businesses.
AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth.
LB: More than half of our 35 post graduates of which 10 are PhDs are involved in analytics at any given point in time. Unlike most other firms we rotate people between research and analytics – we have found that improves quality of analytics. It’s a simple 2 tier structure, and most problems/projects have a 2-3 person team. We also outsource a large part of our business – at-least those elements where repeat processes are involved. We have an associate IT firm and prefer to automate as many processes as is possible with their help.
AIM: Would you like to share any example of an Insight that generated a huge positive impact for your clients?
LB: A large retail store chain was about to open a large property in one of the mega cities. Analyzing their product profile and using some other data converting that into income segments of their customers allowed us to identify various potential locations that would be most appropriate. We matched this data with Indicus own estimates of income distribution database. The top location from this analysis was in a city that had less than a third of the population in the first location. The client opened a store in his preferred location as well as that from Indicus’s analytics driven first choice. Client feedback is that the store in the smaller city turned profitable within the first two years, whereas the original choice has not turned around even after 4 years.
AIM: Do you think it’s possible to become too married to the data that comes out of analytics? Where do you draw the line?
But even after all this we use the three way test:
(1) Does this result make sense?
(2) What is a good cross-check?
(3) Are there early benchmarks in the implementation process?
The first is the common sense criterion, and is well known. The second is less utilized in the profession but is more important – if you cannot cross-check something it is better not to do it. Analytics can throw up a lot of crap – think of it like a manthan, one will get a range of results and one needs to have some way of separating the good from the bad. If we don’t have a well defined cross-check mechanism, then we tend to depend solely upon the gut feel of the researcher or the manager on the client side. This negates the whole point of analytics.
The third and last is perhaps the most critical, analytics is not only about insights, but also about monitoring. Good analytics can also help in devising early checks in the implementation process.
In sum therefore, if the analytics problem is framed properly – (a) domain specific research (b) cross-checks and benchmarking and (c) early monitoring – then the problem of being married to the data does not arise.
AIM: What are the most significant challenges you face being in the forefront of analytics space?
LB: Two critical points on the client side – One: Convincing the clients that analytics is an ongoing process and should not be a one-time task; the great returns at both tactical and strategic level tend to occur after a few quarters of ongoing analytics.
Two: The orientation of the clients that once the task has been outsourced insights will start flowing in automatically. The client needs to put in a lot of time working with the analytics teams especially in the first 2-3 quarters. On the staff side: Many staff are looking for a sort of a check-list, and they would like to be told specific steps to be done. But analytics requires a lot of lateral thinking.
AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe?
LB: India is about 10-15 years behind in the corporate analytics space. Most work in India is servicing international requirements. The great growth will come when analytics becomes part and parcel of Indian corporates and governments to solve India problems. This is going to occur in India over the next few years. Internationally, analytics will become less quantitative and more thought and logic based on text analysis.
There will also be many analytics solutions that will marry text and quant in ways we have not thought of now and the currently fashionable uni-dimensional quantitative analytics will become bread and potatoes. [quote style=”1″]Eventually, worldwide, it will be those with strong foundations in literature, psychology, anthropology/sociology, history and philosophy who will rule analytics.[/quote]
AIM: Anything else you wish to add?
LB: The IT department is the enemy of innovative analytics, and will always get in the way of implementing Analytics based processes because its orientation is different, which is to maintain status quo. Whereas the Analytics department has anti-status-quo forces built in its very DNA. Since Analytics has the potential to provide firm level benefits across all departments; large firms need to subsume the IT department into a new larger Analytics department.[divider top=”1″] [spoiler title=”Biography of Laveesh Bhandari” open=”0″ style=”2″]
Laveesh Bhandari is an economist by training and has authored several publications and articles. Prior to Indicus Analytics, Laveesh served as Senior Economist at National Council of Applied Economic Research (NCAER).
Laveesh has a Ph.D. in Economics from Boston University.[/spoiler]
Try deep learning using MATLAB