By Atul Sharma, CTO, Peak
A data strategy is a roadmap, it determines how a business will collect, store, analyse and use data to achieve its goals. In the current climate, a good data strategy can make the difference between boom and bust.
That’s because every business is unique. It has its own logic, customers, products, and learnings – and its data capture all of that. A good data strategy means a business can leverage what makes it individual, optimising every process to create efficiencies and deliver on organisational goals.
But how do you build a data strategy? Peak’s CTO and co-founder, Atul Sharma, shares five core steps:
1. Make sure your data strategy represents your business strategy
Your data strategy should be built to deliver on business goals, which means it needs to be aligned to both long and short-term objectives. You need to understand both the goals you want to drive and how data supports and complements that.
For example, if your business is working towards reducing data storage costs in the short term, then your data strategy would need to outline storage solutions and services that meet that need while considering longer-term goals that might be affected by those choices, such as AI adoption.
2. Understand your organisational structure
To enable collaboration and avoid duplication, a data strategy should incorporate organisational responsibilities. It needs to be clear which teams and individuals have responsibility for collecting data, which for structuring it, and which for storing it.
Each team uses data differently, they all have different objectives and goals for their data use, and most need it in different formats. In larger businesses, there may be one central team that owns data, but in smaller organisations teams need to be empowered to take responsibility for their data and nominate functional owners to oversee that. Even the best, most carefully thought out data strategy will fail when faced with missing or incomplete records.
3. Combine all your data in one place
Most businesses collect data in silos, often in different systems of record. The Marketing team might report in excel, while the Sales team stores data in a CRM, and Finance will have a different reporting tool altogether. The result is multiple functions each with a narrow, incomplete view of what is going on across the organisation.
A data strategy should aim to unite all of an organisation’s data in one place. It’s not an easy task – these teams all have different labels for their data, they’ll use varying metrics, and they’ll often have different values for the same data point. Getting all of your data into one place is a mammoth task, agreeing on which measure is correct, and ensuring it adds value for every team that relies on it can take months.
But once complete, this is often the first point at which a data strategy can truly prove its value. Having a holistic view of all your data for the first time allows for the identification of patterns and trends that would previously have been missed. There’s a huge amount of insight and value to be gained by combining an organisation’s data in one place.
4. Build an architecture that can stand the test of time
Every data strategy should include an architecture, this is a framework of technology and services that will support its deployment.
Traditionally, data architectures focused on optimising systems; ensuring information flowed and was stored appropriately. It’s a tried and tested approach that worked well for Business Intelligence, showcasing historical data in easy-to-interpret dashboards to aid decision making. But technology has moved on, and technologies like Artificial Intelligence (AI) mean more and more businesses are starting to model future scenarios.
And tech teams in turn need to rethink how they approach data architecture. If technology like AI is likely to be employed by your business in the future, then your strategy needs to include an architecture that can ultimately support that.
5. Future proof your data strategy
AI deploys machine learning algorithms to identify patterns and trends in data. For businesses with huge, transactional datasets, this software is a gamechanger. It enables fast, real-time analysis at a speed and scale that would previously have been unthinkable.
AI isn’t the inevitable outcome of a data strategy – but it should be. This technology will change the way we work in much the same way that the internet did, and devising a data strategy now that isn’t designed to facilitate AI adoption either in the mid or long-term is regressive.
