Without a strategy, data becomes aimless. Experts in data, artificial intelligence (AI), government and business say it’s vital for organisations to develop a data strategy in order to optimise their businesses, meet new regulatory demands, deal with economic challenges and benefit from AI. Recently at Big Data and AI World 2023 in London, practitioners across sectors discussed the importance of developing a proper and effective strategy.
Organisations are analysing data as they think that is what they should be doing, but they are analysing without purpose. Technology analysts Gartner has found in repeated studies since 2015 that between 60 and 85 percent of data projects are failing. The data strategy has to be aligned with the business and its goals.
Without a data strategy in place, organisations face increased risks. These risks range from data project failure to an inability to respond to changes in the marketplace. “Too many people think if I get more data, it will become clear,” explained Eliza Manningham Buller, the former Director General of MI5, the UK security service. In her Big Data and AI World conference keynote speech, the counter-terrorism expert added that organisations need a strategic approach towards data to deal with major challenges, such as “the need to move to Net Zero carbon emissions, the War in Ukraine, pressure on the NHS and future wars about water and food security.”
Shell Chief Data Architect Carol Parillon agreed with Manningham Buller on the risks of too much and poorly considered data: “Organisations have not thought about the data in terms of cleaning it and the data sources. Architect up front, and you will have a lot more success in your project delivery.” As part of that architecting of the data strategy, Andrew Brown of IBM added: “Governance is often talked about as a process that might slow you down, but governance is a way to accelerate, as you are bringing together groups of capabilities that can create and drive change.”
Brown said good governance and strategic planning would prepare the business for the risks of AI and also enable the adoption of this important technology. I would add that organisations must make sure they include as many departments as possible in the process. In addition, businesses often have a lack of well-defined success criteria. The data team and all of its partners will agree on the outcomes if the data strategy is created collaboratively.
Key elements for a data strategy
A successful data strategy has six key elements, these include business alignment, access maturity, a data framework, governance, integrated solutions and scalability. A data strategy starts with business goal alignment in order to secure the data programme budget. Organisations then need to analyse and understand the data access abilities of the business, which will uncover data silos that are preventing a data-led operating model.
I also advise organisations to create a data framework to set the objectives and outcomes the data strategy will deliver. Like Brown at IBM, I, too, believe governance is an important consideration from the outset — as it will shape behaviour. The final steps are planning for data integration and scalability so that the data strategy continues to meet the needs of the business as it changes. All too often organisations underestimate the time and requirements to get a data strategy off the ground.
Parillon at Shell says alongside the need for a data strategy, organisations must ensure the data team is skilled and diverse. “You have to think about the diversity in the team. I have male, female, and different backgrounds in the team, and that gives me a rich mix when it comes to developing strategic plans,” she said. “Make space for diversity as our markets are changing,” she added the need for data teams to reflect the customer base of a business.
Embracing artificial intelligence
Organisations with an effective data strategy will have the capability to embrace AI and develop business benefits. “AI is very important for analysing data,” Manningham Buller said of why this technology is vital to organisations rich in data. Brown at IBM says data leaders need to develop data strategies that already contain the “guardrails” to benefit from AI but also protect the business.
“AI guardrails are about how do I make it safe, how do I make AI usable and effective and how do I accelerate my data design?,” Brown said. A number of AI regulations are on the horizon, in particular from the European Union. Complying with regulations need not inhibit AI innovation, and Brown said organisations should prepare their data strategy with these regulations in mind so that regulatory compliance doesn’t become a bolt-on to an existing strategy. In addition, Brown says data strategies must consider the entire lifecycle of an AI deployment, as customers and regulators will demand answers on how an AI was deployed, tested, developed and decommissioned.
Data provides businesses with the insight to achieve their aims, therefore, data has to be strategic and closely tied to the objectives of the organisation. As economic challenges continue to impact all sectors, new technologies provide opportunities and disruption —and a well-defined data strategy has never been more important.