How to design your data strategy aligned with the business results you are targeting

In today’s fast-moving and competitive business environment, a data strategy is a ticket to play, no matter what the size of your organisation is or in which industry sector it operates. The good news is that it can be faster and easier to create a data strategy than many businesses imagine, especially when you weigh the effort against the value it delivers.

Nonetheless, most enterprises have some work to do. One global study from KPMG reveals that 57% of enterprises do not have a data strategy and only 32% fully utilise their customer data. One potential stumbling block, in my view, is not knowing where to begin. The right starting point for a data strategy is the business requirements.

The use cases

By starting with a robust business case, an organisation can ensure that the data strategy and the business strategy are aligned and that the investments into data will deliver distinct and trackable outcomes for the business.

Creating a data strategy is about understanding the requirements for individual projects and use cases, and building out a multidimensional data solution that accommodates each of them.

Some potential use cases include:

  • Analytics through reporting
  • Data democratisation through self-service reporting tools
  • Knowing your customer through data augmentation and segmentation
  • Next best actions for revenue generation
  • Forecasting through machine learning
  • Commercialisation of data (data monetisation)

Knowing where the data is

Understanding which data you have access to and where it might be found is the next step in fleshing out a data strategy. Most structured data will be stored in databases across the organisation either on-premise or in the cloud — this transactional data can be easily pulled into a centralised repository for analytics and processing.

However, some of the richest data organisations can access today includes flat files, data accessed through application programming interfaces and unstructured data such as images, audio and video. There are also many third-party data sources, ranging from public government databases and social media platforms to business partner systems.

Full data discovery can take some time, however, investing in discovering, data cataloguing and meta data management will pay significant dividends. Don’t underestimate the importance of understanding the quality of your data. Data wrangling accounts for up to 70% of the time expended in data projects, but it’s necessary since improving the quality and availability of data can dramatically reduce the time it takes to implement future data projects.

Putting the data infrastructure in place

The technical infrastructure that underpins your data strategy is not just about what data you will leverage, but also what the underlying technology is and who the vendors are. You might require extensive vendor support in implementing projects using vendor provided environments for your data pipelines (extraction, transformation and loading), data lakes and data warehouses, machine learning and visualisation.

Your data strategy is a long-term programme that affects multiple business cases and departments. The accuracy of the data it collates will be instrumental in many of your critical business decisions. As such, the vendors and service providers you choose will have significant bearing on the success of your data strategy.

Your vendor of choice needs to earn the trust of your organisation and provide a solution that is tried, tested and certified. Technology is only as good as the people behind it, and understanding the technology administrative, maintenance and support requirements is critical in terms of choosing the right solutions. As such, your vendor should offer a roadmap that makes the people requirement clear from the start.

This should include a clear view of which training and support it will offer to your internal teams and the associated costs. Skills for some data competencies, technologies and products are scarcer and more expensive than for others, so it’s important to understand what your requirements will be. You can outsource some aspects of data strategy and operations to a systems integrator.

Govern data across the entire organisation

Data sovereignty, governance and compliance are becoming ever more critical. Your data strategy should enable you to comply with the organisation’s data privacy regulations and security best practices in terms of access. The ability to identify, tag and report on data use is essential for auditing as is the introduction of anomaly detection on data movement. The physical location of the stored data is also a key legal consideration.

Embracing the change

Some data projects, like improving operations through automation, will demand wide-ranging changes in the organisational culture and how people work. Successful change management depends on high-level sponsorship, diversified stakeholder involvement and early adopter champions. However, other data projects like dynamic reporting and data-driven decision making have clearer non impacting business value that the whole organisation can easily adopt.

Keep it going

Unfortunately, implementing a data strategy isn’t a once off event and data operations is becoming more complex, especially with the introduction of AI into the environment. Having the right resources to maintain the environment, the right processes including both DevOps and MLOps as well as the right technologies like code repositories and monitoring tools are all critical components to take into account.

Designing your data strategy using the steps outlined above will help you to take into account both strategy and execution. This significantly increases your odds of success in delivering the desired outcomes for each of your use cases. Your use cases should, however, each have their own execution timelines with quick wins, priority or significant value projects done first.

If you follow agile and new product methodologies, you will be able to map out your projects and aim to get some minimum viable projects out of development and into production as soon as possible. This will demonstrate the business value of your data strategy in the shortest amount of time. “Start small and grow successfully” should be your mantra.