For organisations that are early on in their automation journey choosing the right technology stack can be more straight forward, i.e., pick a data visualisation tool and start building dashboards (we’ll cover this tomorrow). However, to build automation at an enterprise level requires a bit more work, namely, devising a data tooling approach. In organisations which have an Enterprise Architecture (EA) function, this will likely be the most significant touch point with the business and its’ data & analytics team(s). To frame this topic, we will use the typical people, process, and technology approach.
PEOPLE. Before we can agree on the tools, the business must decide on whether it wants to democratise or centralise data. This is ultimately about deciding whether a centralised, fragmented or replicated structure is the best fit for the organisation’s needs. This choice defines whether the architectural landscape should be more heavily weighted in favour of end user tools or highly specialised technologies. It also means the difference between a single data platform that is controlled by a central team or a more open structure that provides greater access to data such as a data fabric for example. Defining this early, making the appropriate hires and communicating sufficiently will provide clarity to the wider business, increase cross-functional collaboration, and improve service delivery which will ultimately all lead to greater trust and better engagement.
PROCESS. From a process perspective there are 2 angles. The first is dedicating time to understanding the target (or existing) business process and how technology is intended to enable it will allow for better alignment / integration between the two. The other angle is ensuring robust processes are developed to ensure optimum usage and maintenance of data products. One of the main goals for both perspectives is to minimise any potential business or service disruption.
TECHNOLOGY. Taking in consideration all the above, when we are thinking about a holistic approach, apart from data visualisation, consideration must be made for technologies to enable several functions such as: data governance, data modelling, data profiling & querying, data lineage mapping, data preparation, data integration, configuration management, metadata management, reference & master data management, data science amongst others. Ultimately where an EA is available, they should be able to provide recommendations & guide the selection of an appropriate tool stack. However, for organisations that may be missing this function, it is highly recommended to seek professional advice. An investment early on will avoid an inflexible landscape that is expensive to redesign in the future!
