Data: Technology costs falling, human costs rising
Data that is incomplete, incorrect, poorly segmented or miscategorized is disaster. It foils timely operational action, good customer service and i mplementing strategies. A rough estimate is that bad data costs companies as much as 15% to 25% of their revenue. Getting high-quality data, running it through appropriate analytics tools and applying the resultant insights forms the core of the digital transformation across enterprises. And making sense of the raw data needs technology and costs money.
Rahul Chopra of Clix Capital points out that technologies for efficiently leveraging data is getting quite common day by day largely due to the rise of public clouds, which offers numerous services that can be leveraged for collecting, storing and converting data into meaningful information. “Earlier one needed to invest huge capex for such data-driven technologies, but now one can try with different options/services in the public cloud and see what suits them best. Moreover, instead of investing in capex, it is shifted to monthly usage cost, which makes it affordable for even small business to consider and invest,” he says.
NOT AN UPWARD TREND
Shreeraj Deshpande of Future Generali India Insurance do not see any increasing trend in the cost in view of advancements in technology and availability of granular data from various touchpoints. “Only ETL, or ‘extract, transform and load’, doesn’t suffice to make the best use of the data for data preparation process and managing it in different formats. Hence data wrangling (ie. to manage the newly generated data from various sources for analysis) is necessary. Data wrangling is used for improving the analysis process of complex problems during data preparation and tools devised to perform data wrangling are very much available these days,” he explains.
For Rahul Bhargava of InCred, costs in this case are either flat, or on the verge of falling. Quite a few vendors and service providers, he says, have mushroomed in the past few years, who are providing software as a service, including open source technologies. Such options are being pursued ubiquitously.
“They could take a physical copy of a document with all kinds of data and return them into digital format after deciphering all the elements in it. For instance, based on the data provided, they can detect salaries from a page and check if they are within the salary range and check for address proofs for validating identity etc,” he adds.
MEASURES TO CUT COSTS
As the cost of storing data is high, says Prashant Deshpande of Shriram Transport, the company has enabled de-duplication with various compression ratios. This reduces the space required for data storage, cooling and power cost of the operations. “We have designed a hyper convergence infrastructure for our storage environment with software designed network in a virtualized environment. This ensures intelligence is built in for managing technology cost and resources at every level,” he elaborates.
MANY PLAYERS
Rachit Chawla of Finway Capital finds the