Most tech projects inevitably involve the innovative use of data, whether that's through analytics, IoT, artificial intelligence or machine learning. In fact, most digital transformation projects are more about data-led change than anything else. So, how can professionals create data-led strategies that make the most of the information their organisations collect?
1. Build the right foundations
Brandon Hootman, director of digital data at Caterpillar, believes there's been a pivot when it comes to exploiting data during the past couple of years. While data wrangling – the process that transforms raw data into more readily used formats – is still important, some executives are focusing more on business requirements to drive experimentation quicker.
"I think the companies that you're going to see that are really successful in this space have taken a more mature approach to managing and building data capability. So, when it's time to do that experimentation, you bring the business to the data instead of taking the data to the business," he says.
At Caterpillar, Hootman has built a foundation for analysis – using data pipelines, management functions and data lake technology from Snowflake – that allows his company to use consolidated information sources to support new business use cases. He advises other digital leaders to take a similar approach, but recognises that isn't always straightforward.
"That's a shift – and unfortunately, there is no easy button," he says. "I feel very fortunate that we've made the investment that we have to be able to do this work and I feel that we're doing that in a way that means we're starting to see the benefits."
2. Understand where better data can make a difference
Access to the right data is important across the business – from analysing customer sentiment right through to security. Prabhath Karanth, director of security compliance and assurance at travel management company TripActions, encourages everyone to see security from the point of view of data that can be shared with management.
"If you set your environment up in a way where you are building your security practices and programmes on top of the data layer, it becomes much easier to scale and it becomes much easier to produce the metrics that your management is expecting," he says.
"It also becomes much easier to provide deep, data-driven insights to your executive management to get further security investments for your programme."
Karanth also encourages professionals to consider carefully the technology solutions and vendor partners that will help them to think about security in a data-driven way.
"This movement is here to stay," he says. "The teams and programmes that embrace a data-driven mindset for security and compliance are going to be hugely successful."
3. Give control to business users
Salim Syed, vice president of slingshot engineering at Capital One Software, says companies must focus on democratising data expertise and pass more control to line-of-business users. The faster the business can more, the more likely its success.
"Data needs to be democratised – data engineering needs to become democratised," he says. "What business wants is to get to the insights it needs, the models it wants to run, and to move at the speed that modern work demands."
Syed says organisations that get stuck on engineering and backend integration waste a lot of time. Instead, create a platform that works, put the right policies and processes in place, and then give business users the opportunity to experiment with data.
"Treat data as a product and have an owner and manage everything effectively," he says. "Something I say is: 'central policy, central tooling, federated ownership' – that's the model. Think of the risk, put the guardrails in place, and then go and innovate."
4. Be prepared to tweak systems
Daniel Smith, head of analytics at fashion house PANGAIA, says it's crucial that professionals who are looking to get the most from data put the right source systems in place. It's then much easier to think about how to exploit insight for decision-making processes.
"You've got to have a 'fail fast, learn and iterate' culture. You're never going to build the perfect solution, version one. You have to iterate and, even when you think you've got it right, the important changes you'll have to do are constantly moving," he says.
Smith is working with Board International to transform the sales-reporting process, to consolidate multiple data sources and to enhance the company's analytical capabilities.
The key message, says Smith, is that any attempt to make the most of data is a constant work in progress: "I wouldn't say that we've ever had a dashboard that hasn't been tweaked. The longest a dashboard has gone without being tweaked has probably been two months."
5. Aim to build an information ecosystem
Milena Nikolic, CTO at Trainline, says professionals shouldn't just confine their data-led transformations to internal knowledge. Her company uses data – some of which it collects internally and some of which it draws from across the industry – to power features that improve the experiences of its customers.
"We use data a lot for decision making – that's very important," she says. "Certainly, the data that we acquire from how users are using our products, which we obviously acquire in a consented and privacy-conscious way. We use that a lot internally to make sure we count and measure the right things, we set the right goals, we set the right level of ambition, and we detect when things go wrong."
She says the long-term approach at Trainline is to build application programming interfaces with external parties. The aim is to create benefits for the business, its customers and other organisations across the sector.
"Rail is an ecosystem, and we understand that we are part of the ecosystem and that the only way for us to succeed is to work with the rest of the ecosystem, which is train-operating companies as well as everyone else. So, it's all about partnerships. It's very much about doing the right thing to bring more and more customers to rail."