Understanding your company’s data management needs and competing on Analytics is crucial for adequate use of your data to achieve organizational goals. Evaluate these questions to get a holistic view of how ready your company is to compete on analytics.
- What data is needed to compete on Analytics?
- Where can this data be obtained?
- How much data is needed?
- How to make data more valuable?
- What rules and processes are needed to manage data from its creation through its retirement?
The prerequisite to compete on analytics is to get access to the right kind of data that can be accessed by a collaborative effort of business(concerned) and IT executives. For the optimum utilization of data the IT executives require business insights accumulated from quantitative analysis.
The basic idea is to avoid “dirty data” (inconsistent, fragmented and out of context) and “data overload”.
Where can this data be obtained?
Data for analytics is gathered from across the organization’s enterprise system. These data are scalable, consistent and well-framed which can be used for management decisions. However, these kind of data is exclusive to a particular firm, so some other types of data that align with the needs of the firm has to be combined for competing on analytics.
Also, an organisation’s personal computers and servers contains enormous data, these are internal data mostly in the form of Databases, spreadsheets, presentations, reports and IoT sensors and devices. To cite an example of oilfield drilling equipment, with the advancement in IoT and other edge devices, operational data from drilling equipment like vibration, temperature, oil-water flows etc. are used to change drilling strategies in real time.
The internet, social media, email, voice applications, photographs and biometrics are some external data that are less- structured or unstructured. And of course, there is option of purchasing data from firms and gather information through company websites. Progressive Corporation an American Car insurance company has collected 10 billion miles of customer driving data through a device plugged in the car providing the company insights about what insurance will cost the company.
Since “Open data” movement even governments play a vital role as information provider and considered a powerful source.
How much data is needed?
Though companies need a lot of data to process and predict customer behaviour but it’s always quality over quantity. Unless the company in question is in the data business, it’s irrelevant to collect all possible data because the cost of purchase or gathering the data is going to outweigh the revenue of the company. Companies tend to collect data that is easy to capture but not necessarily important which makes it difficult to separate the wheat from chaff.
In 2007 Walmart’s warehouse had 600 terabytes of data which grew to 60 petabytes in 2017.
How to make data more valuable?
Generally data is inconsistent, fragmented and out of context. To increase value of data and align that to the company’s analytical needs the prime considerations are –
- Authenticity of data.
- Completeness and distinctive capability.
- Consistency of data to chances of using outdated data.
- Analytics of data adds value to raw data.
- Updated data (periodically, daily, weekly etc.)
What rules and processes are needed to manage data from its creation through its retirement?
The life cycle of data passes through technical and management challenges to compete on analytics. The process needed to manage data from its creation through its retirement consist of the following steps –
- Acquisition of data through various sources.
- Data cleansing, integration and curation.
- Data should be systematically extracted and stored in the right repositories.
- Maintenance of data along with timely updation and ensuring privacy and security.
Some analytical competitors have estimated that they spend $500,000 in ongoing maintenance for every $1 million spent on developing new analytics-oriented technical capabilities.
Once an organization deals with the data management issues the next step is to determine the technologies and processes needed to capture, transform and load data.