An enterprise data architecture strategy is the first step of building an enterprise data architecture. It has to be aligned with the overall enterprise data strategy and should detail out aspects of the solution mainly from the end users perspective.

The key aspects to consider while building an enterprise data architecture strategy are as follows:

Data Acquisition Strategy

Data acquisition strategy details out the way in which data is acquired from various data sources, both internal as well as external, so that it can be put to use. This strategy details out how data sources are identified, how data quality is monitored, how data integrity is ensured, how data is secured during the transfer of environments etc.

Data Cleansing Strategy

Master Data acquired from various sources may not necessarily have the right names, addresses etc and require cleansing for effective use. A cleansing strategy elaborates on the tools, technologies, external libraries and processes associated with cleansing the data.

Data Processing and Transformation Strategy

Data acquired from various data sources is further processed and transformed for deriving business value from data. It is necessary to decide the data processing and transformation strategies and tools that will be used. The strategy details out whether data is processed on premise or on cloud, and the way in which data transformation is carried out.

Data Standardization Strategy

Data is stored in various systems across the environment, and each system has a unique way to refer data. It is necessary to get a standard view of data that is consistent across the organization to ensure consistency in interpretation of data. Data standardization strategy details out the way in which data is standardized across the enterprise.

Data Consolidation Strategy

Data is stored in various disparate systems across the enterprise and it is necessary to consolidate it to get a complete view of enterprise data. Data consolidation strategy details out the manner in which data is consolidated across the enterprise.

Data Deduplication Strategy

Customer data coming from multiple sources can have multiple instances of the same information. The ability to deduplicate this information and to create an enriched golden master record allows enterprises to not only avoid embarrassment when dealing with their customers, but also to understand their customers better.

Data Integration Strategy

In an environment where multiple systems hold same data, it is necessary to ensure that the data across all the systems are kept in synchronization. Data integration strategy elaborates the way in which data in multiple systems are synchronised for a consistent view of enterprise data. It also details out the strategy for designing, sequencing, scheduling, optimizing and monitoring of data integration jobs. Data integration strategy also ensures measuring and optimizing data latency and throughput, in both real time as well as batch modes.

Data Storage Strategy

Data processed data needs to be stored for various business applications, and data storage strategy details out how data is stored. It defines the architecture and technology used to create persistent data stores, which are then used for the intended business applications.

Data Archiving Strategy

It is necessary to manage the lifecycle of data across the enterprise effectively, such that data that is needed is easily available, while data that is obsolete is archived to avoid the costs of maintaining unnecessary data. Data archiving strategy defines the policies and the processes for archiving data, and also details out the mechanism for retrieving the archived data, if needed.

Data Modelling Strategy

Data within the data stores is stored with an intention to use it for specific business applications, and it is necessary to organize this data such that the identification and retrieval of data becomes easy and efficient. Data modelling strategy details out how data within the persistent stores is organized to provide optimized storage and retrieval for various business applications.

Data Virtualization Strategy

Sometimes, it is convenient to create virtual data hubs that consolidate and standardize data, while data continues to persist back in the source systems. The strategy for virtualizing data details out the specific instances when virtualization is permitted and preferred, and the way it is done.

Data Reporting & Visualization Strategy

Data reporting and visualizations are key to interpreting data insights for business users. It is necessary to build a strong distribution channel for data to reach the right users when they need it, and in the manner that they need it. Various data reporting channels and methods are elaborated in the data reporting and visualization strategy.

Data Quality Strategy

Data quality is an enterprise function, and involves checking of data across the enterprise for accuracy, completeness and appropriateness. There needs to be a centrally managed data quality rules repository, and all data quality checks should be traced back to this central repository, to be in control of enterprise data quality. An enterprise data quality strategy elaborates on how data quality checks are implemented, how data quality rules repository is built, and how data quality dashboards are built and managed.

Data Cataloging and Data Traceability Strategy

Enterprise data is fragmented in various data sources, and it is very difficult to identify the exact meaning of data stored in various source systems. We need to build a single instance of an enterprise data definition repository where individual data definitions have the ability to be traced back to various physical instances of that data. This leads to a common interpretation of extremely crucial data for insights, reporting and compliances.

Data Security Strategy

Enterprise data is as asset, which needs to be secured. A loss of critical data can result into permanent damages to the reputation of the organization and to the trust of its customers. It can also result in business loss. Data security is implemented through a strong data access layer, which ensures authorised access of data to business users, and has the ability to predict instances of data loss or data theft before it actually occurs.

Data Audit Strategy

Data audits enable tracing data back in time from its creation to its current state, and ensuring that the changes are legitimate and reflect the reality. Data processing and data storage strategies should ensure that data audit strategies are incorporated within their design. Systems that enable data auditing ensure data integrity within the organization, and the data audit strategy elaborate on how these systems are designed, built and integrated within the overall enterprise data architecture applications.

Data Reconciliation Strategy

Enterprise data is replicated across various systems, and to ensure the accuracy of data, it should be reconciled with its source in an automated manner, such that any discrepancies are identified much earlier by the technical teams, before the business users access their data. A reconciliation strategy elaborates the way in which systems and data points are reconciled on an ongoing basis.

Big Data Strategy

There are several use cases of enterprise data that involve processing very large volumes of data or processing data that is unstructured. Data Warehousing tools deal with only structured data and get very expensive when data processing volumes exceed their limits. Also, there are use case to process data real time. Big data platforms provide the flexibility to process complex unstructured data, it can also process huge volumes of data at significantly lower costs and have the potential to process data real time. Big Data strategy is an important component of any enterprise data strategy.

Data Analytics Strategy

Analytical models are complex, and building an analytical model requires high level of subject matter expertise and precise judgements. However, modern methods of building self learning algorithms take out the complexity of building sophisticated analytical models. They also have the capability to self align the logic to suit the changing circumstances. Machine learning models, which have been largely a research area until now, have started finding real applications when combined with big data technologies, and have become an important component of modern enterprise data architecture.

Data Exploration Strategy

Modern enterprise data management strategy focesses innovation through data, but that requires extensive exploration of data to build innovative use-cases of data, which include improvements to existing businesses as well as building new business models. To enable innovation, enterprise data architecture should provision data labs for data scientists and data analysts. Data labs potentially use open source big data platforms to keep costs to a minimum, and enable data exploration within the organization to add value to businesses.

Data Communication Strategy

The insights derived from enterprise data are useful only when they reach the right users, in time when they need them, and in the format that they need them. This requires translating data insights into business guidelines that get delivered to users through modern digital channels.

Data Capability Assessment Strategy

An enterprise data strategy should also enable an organization to measure the level of maturity that the organization has achieved in its journey towards building a data driven organization. A data capability assessment framework leverages organizational data to access the reach and incremental improvements to enterprise data capabilities within the organization.