Source system owners vital in data analytics projects

The lack of co-ordinated collaboration between source system owners and data teams may impact optimal functioning of the business.
Nathi Dube
By Nathi Dube, Director, PBT Innovation, PBT Group.
Johannesburg, 20 Feb 2024
Nathi Dube, director, PBT Innovation at PBT Group.
Nathi Dube, director, PBT Innovation at PBT Group.

Data analytics projects tend to focus more on business users who are the actual data owners. However, as much as business users understand their data and in which systems it resides, they do not usually know the details of how data is stored and how it can be extracted.

Source system owners help to fill this gap as they possess expert technical knowledge relating to the source system data models and how data is organised and stored in the database.

Source system owners therefore play a critical role in the data value chain, and in this piece, I explore this topic further.

Integral to the project

There are a few key data-related requirements aligned to any data analytics project and that source system owners support. These include:

Data availability and accessibility: During the development phase, source system owners often do not have a comprehensive view of the company’s data requirements, as these tend to be left too late in the project. The impact of this is that when certain data is required down the line when the system is live, it may not be available or could be too complicated to extract. When data requirements are part of the project, the source system can ensure data is available and easily accessible by design.

It is important to establish trust between source system owners and data teams in a mutually beneficial way.

Impact on front end design: Incorporating data requirements early in the project can influence how the source system front-end is designed, so that it is easy to capture all the required information.

Reduced query complexity: To assist data teams, especially when it comes to complex queries where combining/joining many entities/tables is required, the source system owner can ensure more information is populated into as few entities as possible to reduce the complexity of queries used to extract the data.

Business transformation projects consideration: In business transformation projects, upfront data requirements may help highlight the GAP in the new system which, otherwise, could have been missed and have this addressed early during the development phase, instead of the issue being identified much later in the project, which could result in serious business impact and costly system changes.

Data in the right format: It does happen that the data required is available but not in the format the business wants. When data requirements are clearly defined (and communicated) and factored in earlier on in the projects, the source system owner can ensure critical data requirements are not missed, and data is made available at the right time and as required.

Benefits of collaboration

It is important to establish trust between source system owners and data teams in a mutually beneficial way. It is easy to think that the relationship between data analytics systems and data sources is one way, but this is not always the case. Data analytics environments, like data lakes, can also serve as a source for upstream systems − so the relationship can be bi-directional.

Source system owners:

It should be noted that data platforms combine information from various sources and often have capacity to handle large volumes of data.

It is therefore possible to handle some resource-intensive data processing tasks outside of the operational systems, in a data analytics environment and then feed that information back to the operational systems once it has been processed.

Data analytics platforms also have capacity to store historical data over longer periods than operational systems. This comes in handy when there is a need to analyse historical data that is no longer available in the operational system.

Data analytics environments can also highlight data quality issues which the operational systems need to address. This creates a feedback loop in the data value chain and provides opportunities for continuous data quality improvements.

Data teams:

The benefits of this close working relationship for the data teams are immense. Over and above the improved accuracy of data, which is a big factor, there is improved overall efficiency in delivering data analytics projects, as the collaboration ensures the easy flow of information, especially during the design and analysis phase of the project.

The collaboration makes the execution phase smoother and more efficient, as questions around data mapping, data transformation rules and other source data-related matters are quickly addressed.

There is improved overall confidence in the process, as uncertainties pertaining to correctness of the data and the way it is being extracted are cleared. Source teams are also available to assist with any recons required, ensuring high levels of data quality are maintained throughout the development phase.

Impact of lack of involvement

A lack of co-ordinated collaboration between source system owners and data teams has undesired consequences that may impact optimal functioning of the business. This can also undermine organisations’ investments in data analytics platforms and lead to failed data projects.

When source system owners are not involved, the general quality of available data may be affected, causing a ripple effect across the data value chain, impacting other data-dependent initiatives, like the ability to rollout effective artificial intelligence and machine learning services.

Customer service may also be affected if correct and up to date information is not available to service the client.

Some source system owners make use of reference tables to store various configuration information, like account statuses, customer type, etc, and these may appear as just numbers and will need to be decoded and mapped to correct labels and descriptions.

Simple things like knowing whether monetary amounts are tax-inclusive or not, or if ‘1’ stands for ‘yes’ and ‘0’ stands for ‘no’ in the database can make a big difference in ensuring data is interpreted and used correctly.

When source system owners and data teams work in silos, the possibility of not meeting data requirements is higher, either because the required data simply does not exist, or the data teams do not possess the know-how to extract the correct data. This can have a huge impact on the business when it comes to critical data or reports.

In conclusion

A well-designed data analytics platform provides a springboard for innovation across the business. The success of a data analytics project is dependent on the full participation of source system owners. Whether the source system is supported internally or externally, data teams need to be supported by source system owners to ensure project success.

The configurations of IT teams must be such that it fosters close working relationship between data teams and source system owners. Management must support and encourage a collaborative approach among teams to ensure project successes.