Almost all organizations equipped with differential competitive advantage in Big Data era are data-driven organizations
. An MIT report released in 2011 (Strength in Numbers : How Does Data-Driven Decision-making Affect Firm Performance?) showed that general productivity of data-driven organization is 5-6% higher than industry average.
One survey from The Economist revealed that 78% of companies that regarded themselves as data leader achieved beyond average performance in respective industry.
What is data-driven organization then? An organization operating on a lot of reports? An organization which invested a hefty sum of budget to set up Hadoop or ML platforms?
No. Data-driven organization sees its past and present objectively based on quality data, learns from extracted patterns residing in data, and has cultural basis to make decisions data-based.
Why does data-driven decision making matter? Because human intuition is faulty especially in objectivity.
Data-driven organization relies on data so as to minimize negative impact of these biases by moving the center of decision making from intuition to inductive logical process based on evidence, or data. In order not to be swayed by HiPPO (Highest Paid Person’s Opinion), it creates effective metrics, collects unbiased data for analysis, and extract knowledge out of data using statistical methods. A data-driven organization must not cease this line of activities, and being a such organization must be the top priority for all.
There are three pillars that support data-driven organization.
1. Data-Driven Culture
In data-driven organizations, there is a wide belief that data is more trustworthy than intuition. The more members share this perspective, the closer an organization approaches to be data-driven. This is not an easy nor a short-haul task. It requires a stern internal organizational commitment to drive constant change management. No outsiders can do that for you.
2. Data Engineering
Without proper infrastructure, digital data cannot be utilized. To have that infra, you need data engineering. It is a series of technological activities to collect, unify, cleanse, distribute, and safely guard data asset in a given organization. Data engineering is a pre-requisite for down stream analysis where fancy things like machine learning and AI are happening.
Softline has been in data engineering business since its origin and accumulated know-how about design, implementation, and operation of data infrastructure.
3. Data Science
Oftentimes, data science, machine learning, and AI are synonymized or used interchangeably. But these three are not the same (Refer to ‘Data Science’ section for more detail). AI is an outcome of data science, and machine learning is an analysis methodology for data scientists. Even though there are diverse interpretation of data science, here we define it as a line of activities to analyze data, find statistically meaningful patterns, draw actionable insights, persuade others to buy in its findings, and help others to deploy its findings within business process.
Softline has implemented and operated a number of analytics systems using market-proven solutions such as MSTR and Vertica.