Mastering data, promoting its exploitation and dissemination have become essential to today’s business. But the goal is not easy to achieve because, in most cases, data sources are scattered in multiple databases. Not necessarily compatible with each other, focused on critical functions and poorly structured for analysis, these bases are not adapted to the long-term vision and decision-making. The solution lies in the Data Warehouse.
A Data Warehouse, what for?
This Data Warehouse collects, makes reliable, transforms, synthesizes and allows the dissemination of all information from the company’s IT systems or external systems. In industry, for example, it enables information on the various machines to be cross-referenced, data to be aggregated and predictive maintenance to be performed: What is the lifespan of the machines and how often is it maintained to avoid breakdowns? Focused on business intelligence, it helps to improve customer relations, monitor, manage and improve the company’s performance and forecasts growth opportunities.
Setting up a Data Warehouse cannot be improvised and requires specialists’ assistance. Still, the company has to choose them well if it does not want to make mistakes that it will pay a lot later. Very often, costs are underestimated (by a factor of 3 to 10!), the chosen technologies are put in question after a few months and if the project is exclusively driven by the IT managers, the risk of being cut off from the users and their needs is significant.
A concrete example of bad practice
This kind of misfortune happened to one of our customers, a transport company. The Data Warehouse, which it wanted to set up, was intended to meet the needs of three departments: ensuring compliance with regulations by drivers for human resources, knowing the operating time and the journeys traveled by the different vehicles for maintenance and optimizing production planning. First mistake: the IT department, in charge of the project, has insufficiently consulted human resources. New fields had to be integrated (+ 5% of the total cost). Second error: the database engine was wrongly selected. The market leader had such exorbitant licensing costs that it was necessary to change it after two years. The purchase of equipment and the rewriting of a large part of the flows resulted in an additional cost of 30%. Third mistake: real-time features have been developed in the data warehouse – whereas its main utility is reporting and analysis – instead of being developed at ERP level. As a result, the data warehouse has become more critical than the rest of the environment and has resulted in many customers leaving due to its unavailability. If the client had consulted us from the outset for the implementation of his project, we would have avoided him an additional 35% of the initial budget as well as a consequent loss of his turnover.
Call on specialists
Knowing the tools, understanding the customer’s needs, applying best practices by involving all end users and being able to manage the project from A to Z are the key factor success of any Data Warehouse project. Not only must customers be given robust and stable access to their data, but also solutions such as the data hub, a data lock that prevents production databases from slowing down and facilitates real-time data implementation, the data mart, which allows optimal use of data by structuring them for specific applications or needs, or the data lake, which provides global unstructured data storage for later use.
It’s all about flexibility and budget. Contrary to what one might think, a Data Warehouse project is not only within the reach of large groups. Small and medium-sized enterprises can also benefit. Provided, once again, to call on specialists in data environments who know all facets of the business.