Chief data officers need to focus on fundamental information quality and management, before attempting complex analytics or machine-learning
Many businesses are running before they can walk with data. As a result, chief data officers (CDOs) often realise too late that they need to go back to basics and get their house in order before progressing with complicated projects.
“There tends to be too much quick use of buzzwords like artificial intelligence, data science, machine-learning and robotics, before a company is ready with the right data to feed into these projects,” says Michael Gilmartin, chief executive at consultancy Dufrain. “Getting the fundamentals right first is the only way to succeed.”
These fundamental factors include basic data management, including checking quality and knowing whether information is accurate or duplicated. It can be an understanding of whether a business’s employees can access the necessary information and derive analysis quickly. Getting such steps right early on can also help smaller firms avoid serious problems as they grow.
“Everyone is looking for an amazing dashboard and all sorts of cutting-edge systems, but they often overlook the quality of data. Unless they get their data in good shape, the advanced projects won’t help them very much,” says Mr Gilmartin.
Accuracy of data is a major problem for many organisations. “There are financial services firms we see where only 60 per cent of data is accurate,” says Joseph George, a director at Dufrain. “They may lack accuracy over transactional information or they may have a large amount of duplicated data, such as when a customer’s name is spelt different ways in their records.”
Cutting edge analytics, artificial intelligence and machine learning are the future; but you still must get the data right first.
Many businesses draw in information from multiple locations, including customer, transactional and payment systems. Companies struggle with the different data sources and whether the information is well managed, and they also often lack real control over it. Organising this data, cleaning it and integrating into a central warehouse can enable them to draw analysis from the different elements, and to manage it properly.
Then there is the question of the skills of CDOs and their teams. The CDO has a responsibility for data accuracy and management, and using information to drive the business forward. However, that individual can be very different in disparate organisations. Sometimes the CDO reports to the chief technology officer, sometimes to the chief information officer and sometimes to the chief operating officer. This can lead to a wide variety of outcomes.
To achieve a better and organisation-wide strategic gain, CDOs must be able to establish a department with a variety of skills. “This can include ensuring that staff understand data, can play with it and bring information together to answer important questions quickly,” says Mr George. “While organisations used to recruit data personnel based on their skills around specific systems, there is now more value in hiring people who are vendor agnostic and have good core data skills that can be applied to different technology.”
CDOs also need to make sure staff have softer skills, including how to challenge ways of doing things and how to manage stakeholder expectations. It is important they are able to garner support from across the workforce and this is most easily achieved when businesses go through data by department, collaborating with managers to improve the information. Project management skills are essential so any work around data succeeds on time and on budget, and achieves a clearly defined result.
Businesses with an influential CDO who can drive change stand to perform much better. But Mr Gilmartin adds that it can be hard for CDOs to establish their influence. “There is a real question around how much clout the CDO has in some organisations,” he says. “It’s a relatively new role and there is great pressure to drive change. Some of them may be pure technologists, while others are born-strategists and influential leaders.”
As CDOs look to improve how they work with data, they often seek support. Dufrain works with companies throughout their journey, including from the initial analysis of information accuracy, to improving and making sure data is not duplicated, storing it properly, and right up to advanced projects such as big data analytics and machine-learning. The company works with a number of businesses ranging from banks and retailers to universities and management advisers.
At many firms, one of the triggers for addressing data properly is when the business puts increased value on data because it can help make much more informed decisions. A large retail group, which lacked a single view of performance and struggled to manage £100 million of payments due to it from storecard holders, brought in Dufrain. The consultancy helped the firm clean the data, then improve its analytics and methodology, so it could see all the information and take steps to recover money due much more easily.
Other firms look to harness data to make sure they are able to serve customers in a more personalised way. A banking company recently decided it needed to improve its handling of calls to serve clients in a bespoke fashion. Dufrain worked with the company to develop a “single source of truth” that incorporated internal and external data, enabling detailed client information to be supplied to customer service and sales staff. It could then offer personalised service and more easily cross-sell, while generating automated activity reports for regulators and managers.
CDOs increasingly realise they must get data right to make well-informed decisions more quickly than their competitors. They also have a regulatory imperative to get data protection right and this is motivating improved governance. Only by starting with the basics can they succeed with advanced, real-time data analysis, machine-learning and complex projects.
To get the basics right around data in your organisation, before progressing with advanced data analysis, please visit dufrain.co.uk