MARKET | Aug 24, 2017

The 5 pillars of Data Strategy

What are the key elements for implementing a data strategy and turning it into a business asset?

Too often when discussing data, importance is only given to the technological aspect, without considering all the management, organisational and analytic aspects. Data management strategy, where companies have designed and formalised it, also takes account of the technological aspect, but it is certainly not the starting point.

The data must be collected and analysed so it can be governed in a structured manner and represent a valuable support to extracting that valuable information needed to make quick but reasoned decisions. This requires a data strategy that transforms the data into a business asset.

What is Data Strategy?

In the SAS report, The 5 Essential Components of a Data Strategy, it is defined as the planning needed to improve how data is acquired, retained, managed, shared and used. Planning that requires common methods and formalised processes, which are defined as indispensable for shared management of company information.

While many companies have set up data management activities (metadata, data governance, master data management, data migration, data integration, ecc.), they must also however be integrated in order to be useful. In addition data strategy serves as a road map that allows all these activities to be interlinked, so that more benefit can be gained from each of them.

What are the 5 pillars of Data Strategy?

Historically, the same SAS report states that IT organisations focused on defining a strategy that was only connected to data retention. This is certainly important but is of no use unless it is integrated with the other pillars of the strategy.


Identifying data and recognising its meaning and usefulness, regardless of location, typology, and source is one of the indispensable prerequisites. Establishing shared rules within the business for identifying and representing information regardless of how it will be stored means that an important first step has already been taken. If a business does not describe the data and metadata, even by means of a data glossary (such as the data card catalogue), it is a bit like ignoring information because you prefer not to know it. If the data is an asset, the data strategy must guarantee that all data can be identified and described.


Securely archiving data is undoubtedly a crucial activity. Often, technological archiving choices are aimed more at storing the data as an end in itself, rather than at the possibility of its reuse via sharing. This occurs because data storage plans rarely highlight the need to share and move data from one system to another. If, when collecting Big Data, there is no proper data strategy plan, the issue becomes even more thorny, as this amount of data, collected both internally and externally, is often copied into the different systems rather than stored in a single lake from which it can then be fished. The key to solving this complexity is not so much in storing it in one place, but storing it just once, thus providing access to those who need it without having to recreate their own copy.


Data distribution must be effective as well as complying with the policies for its reuse and consultation. Unlike the past when “solitary” systems were conceived, i.e. not necessarily in communication with others and therefore having export-import data-exchange mechanisms, today, exchange has become a recurring requirement. This is why it must be planned and not based, as is sometimes the case, on courtesy requests for sharing between colleagues in the same company. Distribution is so important that it cannot be thought of as a one-off, but must be planned as if it were to be shared at any given moment.


The key to one-time storage and sharing for re-use lies in combining data taken from different places and systems, to convey it in one view. If we consider that 40% of the costs of new developments are absorbed by data integration activities, it is clear that this is not easy. To ensure effective integration, data development must be considered as a rigorous and methodological discipline. If it is judged that data accuracy and consistency are critical and if data is considered as a business asset, then the data strategy must include integration procedures.


Establishing, applying and communicating data-usage rules ensures better management. This is especially so when the amount of information to manage and share grows and policies must be formalised, roles identified and methods implemented that ensure regulated and monitored (but possible) use of the available data.