Data saturation is everywhere. We’ve often had the belief that more is better; however, that actually isn’t true in the case of data.
Why do hypermarkets keep chewing gums and candies at the cash counter? How do coffee chains manage to have two cafes on the same lane and still be profitable? How do online searches draw advertisements of the same products on your devices? Most of this is not conventional wisdom, it’s use of data, which companies across the world are obsessing about. Data analytics, for many, is the holy grail to drive user demand and revenue.
To put the data deluge in perspective, Google processes 3.5 billion searches per day, Snapchat users share 527,760 photos, 41.46 million people watch YouTube videos, Instagram users post 46,740 photos, and 456,000 tweets uploaded on Twitter. India generates a subset of this consumption, but the numbers are bound to be massive.
India is the second largest internet nation with close to 400 million internet users, many of them on smartphones, only second to China. Access to social media, Google search, entertainment, among other things, on the palm generates data that is dissected to get desired outcome.
Add to that, hyper-focused targeting and segmentation to define niche audience segments. But is the entire stack of data important and efficient?
Data mining became an important aspect in the late 1990s, but as new concepts like big data, and technologies like artificial intelligence and machine learning surfaced, it opened-up the Pandora’s box. “There isn’t a thing as too much or too less,” says Pranay Agrawal, CEO of Fractal Analytics, a Mumbai- headquartered analytics service provider. Citing the example of a leading diagnostics chain for medical data, he says, “If the prediction accuracy improves by even 0.5 per cent, millions of additional data is worth it.” Size doesn’t equate usefulness. Many believe that data isn’t about size but relevant attributions and metrics — so, more the merrier. But there are different approaches.
Abhishek Ganguly, Managing Director of multi-channel sportswear brand Puma, believes in a “business objective first” approach. “Instead of looking at ways to collect and mine data, we start with business objectives. Then we create platforms to collect data,” he says. There are many, like Ganguly, who believe in the business- first strategy. “Business-first framework should take into account use cases, data sets, data collection, data preparation, learning and intelligent actions,” says Pramad Jandhyala, Director of Finance and Human Capital at digital analytics firm LatentView.
But there are segmentations to this approach, one them being the size of the organisation. “For large, established enterprises, data-first strategy drives innovation. But for start-ups and SMEs, setting up and aligning business objectives first should be the priority,” says Arijit Lahiri who runs QuoDeck’s, a game-based mobile learning management system.
Hero MotoCorp, Mahindra Group, Reliance Industries, and Baja Auto are spending millions to build data-driven enterprises. Others like Amazon, which has the customer at the center of its business model, has constituted a team of data scientist, A1 experts, and data analytics models. For them India is no different.
Amazon uses AI and ML to analyse huge chunks of data in various fields — from improving address quality to ranking of deals, and from improving catalogue quality by finding missing descriptions in titles to weeding out inappropriate images. The good part is that Amazon has a lot of user-generated content from product searches, buying patterns, and search. All this data is analysed to figure out what customers want. It doesn’t stop there, Amazon takes its findings forward to its entertainment platforms: Prime Videos and Music. It has been at it for ages. But every firm is not Amazon and every CEO is not Jeff Bezos. Even with the right approach, money-inflow and business alignments, a digital-data strategy can fail. Making sense of data is the biggest challenge. With colossal amounts of data being generated every day — 2.5 quintillion bytes — we are bound to get lost in the web jungle.
QuoDeck’s Lahiri says that data without a hypothesis is pointless. “Before collecting data, one needs to have a hypothesis in place. You might know the answer but if you don’t know the question, it won’t make sense,” he says, alluding to the bottleneck in the analytics industry.
“Forward-looking approach that relies on data-based intelligence, trend analysis, forecast modelling, and predictive analysis is what people need to obtain success in today’s dynamic market. And only accurately collated, actionable insights can give them that, not massive amounts of raw, unfiltered information,” says James Giancotti, Co-founder and CEO, Oddup, a data-driven research and insights provider.
Oddup’s chief operating officer and Co-founder, Jackie Lam says that just analysing past trends alone won’t help, because the market is constantly evolving. But all of this cannot happen without making data-analysis any company’s culture, like in the case of Amazon.
“Creating an organizational cultural can solve the issue of failure. And the culture should follow the top-down approach,” says analytics academician Dakshinamurthy Kolloru, Founder President & Chief Mentor of International School of Engineering.
This story has been written by Nishita Chandak from Times Group
This article was first published in, The Times of India