The task of breaking down data silos may seem time-consuming, complex and expensive, but letting disparate legacy systems atrophy in isolation is likely to prove costlier in the longer term
The widespread adoption of powerful analytical tools and AI-based systems that can offer insights into everything from operational efficiency to consumer behaviour has had a clear positive impact on businesses worldwide. Yet, despite such advances, even the most forward-thinking companies are still likely to have data silos that are preventing them from performing to their full potential.
“It doesn’t matter whether you’re a startup or a large corporation – silos will exist if different departments inside your business store their data in separate locations,” says Mihai Cernei, CTO at Amdaris, a specialist in digital transformation. “As these disparate assets grow, so do your silos.”
The existence of silos makes data analysis a far more time-consuming process than it needs to be. And, if similar material is being stored in numerous places, inconsistencies can easily creep in, jeopardising the quality of the information extracted.
IT specialists faced with a lack of joined-up platforms tend to devote more effort to data management than they might wish to. A 2020 Gartner survey of data professionals found that on average they were spending 56% of their working time on routine data management work. That meant they were spending only 22% of their time on more value-adding tasks such as data monetisation and the extraction of valuable insights.
More data more problems?
Every company will have its own unique set of data consolidation problems to overcome. In Cernei’s experience, nearly every obstacle a business will face in its efforts to bring together fragmented data falls into one of three categories. First, it’s becoming increasingly difficult to find high-quality technical experts who can do this work. Second, the task of even accessing many legacy systems requires either bespoke software or new data connectors to be built. Third, and perhaps the most significant factor for companies looking to eliminate data silos, is that the process is likely to be expensive.
“Some integration is costly and requires a lot of computer power,” says Cernei, although he adds that it’s “important to note that this isn’t true of every solution. Some can save you money.”
Chris Gorton is a senior vice-president at Syniti, a specialist in enterprise data management. He recommends that the first step any organisation should take when attempting to break down silos is to obtain a comprehensive understanding of exactly what data it wants to gain control over.
“Companies need to develop a plan to consolidate its information, harmonise duplicated material and ensure that it is of a high quality, so that can be trusted and used throughout the business,” Gorton says.
As data-synchronisation work can affect all parts of an enterprise, it’s vital that this plan factors in how all operations can keep running seamlessly as the process gets under way.
One practical step towards making progress on data, while also keeping an eye on cost, is to align data towards business strategies
Many business leaders are facing budget constraints that will require them to make hard choices about which projects need to be prioritised and which ones can be shelved until more funds are available. If they are to approve a data integration project, members of the C-suite will therefore need to see a strong business case for it.
“Are there any quick wins that would increase overall acceptance for making the change? If so, this could be an opportunity to save money, reduce risk and create efficiencies that will drive greater acceptance from the entire business on why it should be done,” Gorton suggests.
Replacing outdated software and hardware is a key element of any plan to remove silos. The longer that legacy systems are left unmodernised, the more work will be required further down the line when they cause compliance and security problems, hinder innovation and restrict growth.
Effective governance
The emergence of cloud-based integration solutions has made it easier than it’s ever been to eliminate silos caused by the continued use of obsolescent systems. Data lakes, in which all forms of data can be held and made accessible, can be a particularly effective tool.
Mike Haresnape, director at business intelligence consultancy Dufrain, explains that consolidating the material into this platform enables a firm to “model different data sets and join them together. The data can then meet any number of reporting, analytical or machine-learning use cases.”
He adds: “Organisations can cut costs and time spent on data extraction by ensuring that, whenever an application is enhanced, its data is moved to another platform. Taking advantage of relatively low storage costs and landing large data sets in the cloud or on the premises can also help with this.”
But it’s no secret that data consolidation programmes of this type are notoriously difficult to get right. Gartner estimates that around 85% of big data projects will fail to meet all their objectives, illustrating the scale of the challenge that businesses face when trying to get a handle on complex and disparate data from across the enterprise.
Using a mix of integration solutions is unlikely to succeed unless a comprehensive data governance system is embedded throughout the organisation. In practice, this will necessitate a range of standards, processes, policies and metrics that clearly define how data is used to best meet the goals of the business.
“One practical step towards making progress on data, while also keeping an eye on cost, is to align data towards business strategies,” Haresnape says. “A data governance strategy needs to add extra value, be tailored to the organisation and focus on data that’s relevant to its activities.”
A well-implemented data governance strategy can result in both data quality improvements and the ability to create a map indicating the location of business-critical material. With businesses continuing to generate large volumes of data on customers, suppliers, employees and more, data governance will become an increasingly important part of any method of dealing with silos.
There’s no question that the initial outlay required to break down data silos can seem excessive. But the potential benefits of establishing a ‘single source of truth’, enabling all decisions to be based on the same information, mean that it’s less of a cost and more of an investment in the future of the business.
