Last month, I explored the challenges created by data silos and how pragmatic approaches like a data lakehouse architecture and data steering teams can assist in breaking down the silos. In this month’s post, my attention on the topic turns to more strategic drives that can help achieve better enterprise-wide data sharing and overcome the limits imposed by data compartmentalisation.
A CIO article published earlier this year, titled ‘8 data strategy mistakes to avoid’, highlights how the data strategy must align with the organisation’s overall business goals. One particular quote in the piece attracted my attention – ‘Building a successful data strategy at scale goes beyond collecting and analysing data’. The reality is that there are several other factors to consider. Most notably, companies who do not prioritise their data initiatives and allocate the required resources to make it work might as well not embark on such a campaign at all.
Of course, you might ask what any of this has to do with data silos. Part of ensuring an organisational-aligned data strategy lies in establishing data stewardship, data sharing, and data accessibility across business lines. There must be an overall drive at the company to improve data quality and data governance. The data, information, and subsequent insights must be made accessible to everyone in the business. Therefore, breaking down data silos is something that must be integrated into the overarching data strategy.
The unstructured discussion
Most of the data that a company collects is unstructured. Just think of all the emails, workplace chats, notes, images, PDF documents, and other pieces of information that are created daily. It is therefore important that the data strategy must ensure that this unstructured data is not siloed. For example, important “how to” webinars, business policies, strategies and plans, and budgets and forecasts, amongst others must be accessible and searchable across the organisational environment.
While this is not technically unstructured data, I would like to include ‘spreadmarts’ as part of this. These consist of hundreds of spreadsheets that are extracted, copied, and sent around often containing sensitive and personalised business information. I want to make the bold statement that the number of spreadsheets used in an organisation is not only indicative of shortcomings in systems and reporting but is also a very good indicator of a highly siloed organisation. I have found that in siloed organisations, the employees on the ground use spreadsheets to transfer and communicate data across the business boundaries.
There are two problems with this. Firstly, the data in the spreadsheet is outdated the moment it gets extracted (or it gets updated inconsistently to the source system!). The second is that these spreadsheets often contain personally identifiable data. For instance, client lists, member lists, or even lists targeted for a marketing campaign. Having such data lying around unmanaged, especially across business boundaries, is a huge risk for compromise in the modern world.
Breaking down data silos is therefore a strategic initiative that must encompass the management and reduction of unstructured data and ‘spreadmarts.’
Decentralised data teams
According to the CIO article I mentioned earlier, isolated and dispersed data teams often create data silos. In turn, this has the potential to reduce the value of data and information across the business. It therefore makes sense that a centralised data, reporting, and analytics team can more easily work across organisational boundaries. They can also work at scale. Isolated teams are unable to do so, which adds to the risk of inconsistent reporting across business lines.
The worst example of this that I have encountered was a large organisation that had 21 data warehouses and data marts spread across its business units. More recently, I worked at a company that had three data teams and several other data analysts all reporting differently across the business. Fortunately, this business has started consolidating the data teams into a centralised reporting team. This will help create a unified data ecosystem that enables seamless data integration, sharing, and collaboration across the organisation.
There are many more aspects that can strategically contribute to breaking down data silos. Examples include organisation-wide data governance, organisational focus on data quality, and staff retention and appropriate talent acquisition approaches. What I have covered in this article clearly illustrates that breaking down data siloes is a mission-critical activity that should be mandated and supported at the top strategic level. This should then be part and parcel of the data strategy and its execution.
