Data is a vital asset for all businesses. Not only can it improve existing business functions but can also reduce overall risk and fuel smarter decision-making. It is critical – particularly in the current climate – businesses have a solid data strategy ingrained into the fabric of its organization.
With digitalization accelerating across the globe, it is the organizations adopting a data-driven approach, implemented with an organization-wide data strategy (and leveraging this for actionable insight) which are best placed to survive (and importantly prosper) past the pandemic and into the future. But for organizations unsure of where to start or at the beginning of their data strategy journey, what are the considerations?
- How to form a data-driven digital transformation strategy (opens in new tab)
Identify your objectives and data sources
Before implementing a data strategy, data strategists and leaders must first ensure they have a clear understanding of the business’s short and long-term objectives. By doing so, they will be able to tailor the strategy to meet a business’s individual needs, while also determining any questions that need to be answered. This will provide a clear vision of how best to achieve those objectives.
Another critical consideration involves identifying relevant data sources. This includes determining the lifecycle and origin of potential data to avoid inaccurate, out of date or non-maintainable information as this can lead to inaccurate findings. For example, a data strategist may assume they are consuming data from a source of truth, when in actual fact, it is several layers out from the original source, leading to low data accuracy.
Consider governance and monetary requirements
Transparency around costs and governance requirements are fundamental areas to focus on in the initial stages of building a data strategy. To avoid issues, decision makers should be committed and ‘bought in’ to the potential costs and changes that may be required, including process changes and assigning of responsibilities i.e. data owners.
With legislation such as GDPR in place; how data will be processed, stored, and accessed are fundamental factors to think about. Identifying who owns the data and who within the organization has legitimate reasons for accessing it, plays a fundamental role in the governance requirements. In addition, factors such as retention and archiving must also be considered to comply with requirements.
Build a team with multiple skillsets
Too often data scientists execute tasks that do not fall under their dominant area of expertise. A key part of building an effective team is enabling each member to perform in their role and maximize the use of their key skillsets. By appointing team members to appropriate roles, data strategists will reduce uncertainty and ease overall operations. For example, choosing team members who already have a significant understanding of data processes can avoid future issues when allocating data resources, establishing and improving policies and dealing with data-related challenges which arise.
Domain knowledge should also be a consideration, such as HR, finance or operations experience and knowledge. Members with suitable domain knowledge can bring this into the project along with their core skillsets. Building a team with an appropriate mixture of skillsets – including analysts, data scientists and data engineers – will help to execute the overall data strategy.
- Why data literacy is key to an optimised data strategy (opens in new tab)
Consider your roadmap for success
Once business objectives, data sources, governance requirements and skills are identified and established, data strategists should then build a roadmap showing how the strategy will be deployed across the business.
By focusing on the quick wins, data strategists are likely to gain greater support across the business and perhaps even greater budgets. As an example, a data strategy might first look at customer churn analysis, to help manage or reduce customer turnover. Once these quick wins are achieved, organizations can then focus on bigger data priorities and publish the final roadmap across all departments to ensure total visibility and buy-in amongst all staff.
Define architecture and design requirements
Architecture and design are fundamental elements to consider when building a data strategy. Responsibility typically falls to the CTO but it’s important they do not get bogged down in unwieldy documentation sets but instead focus on documenting through architecture diagrams and process flows.
Engaging in proof of concept (POC) and fast fail design activities can be an effective way to trial multiple approaches with small subsets of data, refining through iterations and ultimately informing the target technical design. The cloud maturity of a business should be considered at this point as well. For a business that hasn’t started migrating to the cloud but has been considering it, conducting POC’s and fast fail design activities in the cloud, rather than on premise, can be an effective way of reinforcing the benefits.
The approach to solution design is dependent on a multitude of different factors. These include levels of legacy business technology, where workloads reside (internally or externally – on premise, cloud platform or partner /vendor), pre-existing technology agreements, the budget, the regulatory restrictions on your operating environment and the cloud maturity of a business.
Automating wherever possible through modern development practices such as DevOps is an essential point that should be considered during architecture and design activities. Done right this can result in highly effective deployment procedures as it ensures the entire architecture is reproducible in code and automated pipelines. Consequently, change control and team collaboration is far more effective when building and deploying preproduction environments for demo, development, QA and training.
Implement your strategy across the business
The final step in building a solid data strategy is the actual implementation, utilization and adoption of a data strategy across the business. The standardization of data collection, transformation and publication is essential. By doing so, different business units have the ability to carry out analysis using well governed, approved datasets for the benefit of their own departments. Having properly governed data sets and processes in place across the business can not only improve meeting data lineage requirements, but also a consistent approach to handling, managing and analyzing data and be implemented.
Ultimately, building a solid data strategy first requires the vision and then, efficient implementation and continuous review. If executed correctly, data strategists can empower business leaders with truth and certainty from their data, through a solid, well-rounded data strategy. Data should be considered as more than just facts and statistics; it should be considered as a corporate asset.
- Barely any businesses have a solid data and analytics strategy (opens in new tab)
Nicholas Finch, CTO, TrueCue (opens in new tab)