Embedded Files
Part of my work at NHS Supply Chain has been helping them decide how to build their data hub. They are convinced that the data they create and have access to, can be mined to bring value to suppliers, NHS Trusts, NHS Supply Chain and help improve patient outcomes.
There are real specializations in how you buy the medical supplies and those experts had historically been sat in separate organizations, with many of their own tools looking at their own universe of data.
There has been a central data warehouse that has served up data that many of these buying communities have an interest in, but they bring that data into their own orbit. There is mistrust regarding the provenance of data that bounces around between systems.
Recently, there has been a change in operating model to bring the buying experts back inhouse and to unify systems. The view of unifying systems could be:
to build a common data dictionary and move data into a canonical model at transactional level. The term canonical has its roots in the word canon, meaning the official law. There are plenty of examples of this model. I was fortunate enough to have been the chair of the Open Applications Group in my career. The whole mission of the organization was to define a canonical that would allow unmediated communications between applications.
I imagine many large organizations, especially those that are run as conglomerates, live in a world with multiple systems that provide overlapping functions. A canonical provides efficiencies in integration and consistency in analytics. It avoids the risks of taking applications away. However, it leaves in place the cost of redundant applications.
To nominate a set of applications in a target architecture and deprecate legacy or redundant applications.
I imagine organizations that desire stronger central control, and greater efficiencies prefer this model. It cuts out the need to unify the data in the mesh. You may be in a position to utilize the pipelines and transformations supplied by the vendor to create your data warehouse. However, an organization going down this route requires very strong change management. The cultural challenges are very real. Stakeholders may be satisfied at current local optima, and see little value to themselves in any changes. They will also need very good risk management. The company will be relying on expertise in one set of tools to solve problems that may be unfamiliar.
It should also be noted that there are risks in the canonical strategy that grow over time. A canonical data model may give you some efficiencies in integration, but it is not zero cost. Processes that rely on many overlapping systems being integrated tend to ossify. Systems become brittle and to make a small change requires recertification of many systems. Keeping expertise inhouse for how systems are integrated is itself a challenge. It means your resources are closer to the infrastructure and further away from the analysis needed to support business decisions. In the case of NHS Supply Chain, closer to data flows but further away from supporting clinicians and improving patient outcomes.
If you would like to start a conversation on how best to develop your data strategy please reach out to us as info@softwareStrategyConsulting.co.uk
Google Sites
Report abuse