Enterprises, machines and consumers are generating an almost unfathomable amount of data. IDC estimates that by 2020, digital universe or total data will reach 44 zettabytes.
Source: Visual Capitalist
According to Gartner, data and analytics are now central to how companies operate, innovate, compete and drive productivity improvements. Many are hoping new technologies like blockchain, artificial intelligence (AI), advanced analytics and machine learning algorithms, and the cloud can help them derive business value from all this data.
You are in trouble when
But is having a lot of data good for organisations? It depends, according to Pierre Bonnet, vice president of product and engineering at TIBCO Software. He cautioned that “When you reach a point where the data knows more about you than you know about the data, it means are losing control of our information systems.”
What C-suites don’t get about MDM
Gartner estimates that up to 40% of enterprise data is “either inaccurate, incomplete, or unavailable” and estimates annual loss as a result at about US$14 million.
Still C-suites understand they need to invest in data solutions, including data warehouses or data lakes, analytics and AI tools, etc. What should not be left out in all this is data governance as well as data management strategies.
However, in the race to deploy technologies and approaches to managing the 44 zettabytes of data being generated, executives would do well to remember that data management, including data warehouses and analytics tools, is not a point solution.
Bonnet said leaders must recognise that the data, market conditions, business models and regulations are constantly evolving.
Advice to CIOs
As a living strategy, master data management (MDM) is subject to the evolving nature of the business, customers and data suggests. MDM systems need to be as agile as the businesses they support.
Bonnet said MDM systems need to be able to change as and when the circumstances demand it. They cannot be the bottleneck to an organisation trying to stay ahead of the competitive landscape.
Must have skills for MDM success
Got data, need data analysts. Got data management tools, need data modelling experts. However, having someone familiar with the principals and strategies being data modelling does not necessarily mean the skills problem is over.
Bonnet said what most organisations will need, and most find very difficult to hire, are people with semantic data modelling skills.
“it is important to understand the meaning behind a data – hence semantic. Semantic data model is not a technical view of the data but is meaningful for the business, with a vocabulary that is fully aligned with all the stakeholders of the business,” commented Bonnet.
Data warehouse and data lakes are repositories for dead data
The “Data Lakes: Worldwide Market Growth, Trends, and Forecast (2019-2024)” report by ResearchAndMarkets.com forecasts the data lake market to grow at 27.4% over the forecast period 2019-2024.
When compared against investments in data warehouses, the report claims that data lakes are an economical option for many companies. The report said cost of maintaining a data lake is lower than a data lake owing to the number of operations and space involved in building the database for warehouses.
Bonnet alluded to information stored in data warehouses and data lakes as dead data. In his view, data stored in these repositories are data created in the past. Data stored in data management systems, on the other hand, is living data.
“Data stored in MDM systems are directly connected to the operational system. Any bad data that finds its way into an MDM system exposes a company’s operational processes, and therefore will directly impact the business of the company. In contrast data warehouses, while also susceptible to bad data only provide historical information. The use of advance analytics or AI on data warehouses will not change the outcome – garbage in, garbage out,” concluded Bonnet.
