In an era of data-driven enterprise, data silos are a major obstacle. When departmental data is isolated in a separate repository, it’s inaccessible to teams and applications that could process it for valuable insights. What’s more, the data becomes harder to secure, audit and manage. As a result, silos restrict and constrain data strategies. What can organisations do to find and eliminate them?

Silos are a natural occurrence, but become an issue where wider business goals have to be achieved. To remove dangerous silos is difficult and time-consuming and has to come hand-in-hand with the data strategy in terms of oversight or governance.

Andrew Robson, CISO for Bentley Motors

The secret, Robson feels, lies in using people, process and technology to find silos and allocate data responsibility, defining roles, evaluating data in terms of desirable and undesirable outcomes, then making sure it’s correctly managed and secured. One key task here is establishing a comprehensive data inventory.

Removing data silos requires a top-down programme to establish a data catalogue, data dictionaries, and reference data models for the most crucial enterprise data assets. It also requires developing a bottom-up inventory of siloed data assets in spreadsheets, isolated databases, applications, SaaS tools, data feeds, reports, and dashboards.

Isaac Sacolick, president of StarCIO

Stéphane Nappo, cyber-security expert for Groupe SEB, suggests another approach to silos, to ‘consider their root cause.’ ‘Potentially, these could be different kinds,’ he adds, ranging from technical, organisational and architectural issues to a split between operational and decisional systems to a reliance on legacy solutions.

By identifying and addressing these root causes, CIOs can begin breaking down the silos they create. What’s more, by streamlining and consolidating data sources, firms can not only improve their data accuracy, but bring corporate, supplier and third-party data into a more interconnected vision of the business.

Ade McCormack, founder of the Digital Readiness Institute, has a different take.

I think it is less about the CIO needing to know all the data the organisation has and more about defining what data is needed. With this knowledge you can approach the various business outposts to request that they provide that data in a defined format.

Ade McCormack, founder of the Digital Readiness Institute

At times, this can meet resistance.

In a perfect scenario the CIO would own the complete IT estate and the associated data. However, with acquisitions there is often in reality a loss of control. When the CIO probes for data, they may face pushback. One complaint to the CEO later and the CIO is told to back off

Ade McCormack, founder of the Digital Readiness Institute

If this is a constraint, McCormack suggests that CIOs ‘define the data needed and let the subsidiary worry about how it compiles it’. Where necessary, CIOs may need to work harder to bring the senior leadership onboard.

‘Data strategy should come from the top,’ says Bentley’s Robson. ‘As with all things that require control, the Board sets the tone, and their respect for and support of data management provides an excellent platform for the wider business.’

StarCIO’s Sacolick concurs, noting that it’s not enough to build a picture of a fractured data landscape; that CIOs need ‘to present this landscape to the leaders and employees that are the most active building and extending data silos, along with the stakeholders most impacted by the lack of data integration.’ ‘It’s only when people connect the pain points with the exposed underlying causes,’ he suggests, ‘that CIOs can drive behaviour change.’

Learn More: https://www.hpe.com/uk/en/solutions/artificial-intelligence.html