In order to work data driven, it is important to consider the type of guidelines and processes to put in place. Although no two organizations are the same, we can distinguish three models when it comes to management and freedom of data. These are a centralized model, delegated model and finally a self-service model.
In this blog, we will discuss the three models and describe each one's advantages and disadvantages.
A centralized model
The data and analytics environment is managed by a central department in the organization. Often this is IT. This is a common model among organizations. The department is responsible for making the data sources available while ensuring that data are available to analysts and business users. The benefits are consistency is delivery of work, branding and clear ownership of the reports. Reasons for choosing a centralized model are:
- There is a lack of knowledge wider in the organization;
- Data is extremely sensitive and needs continuous monitoring, e.g., who has access and what data is being used.
- The organization has a traditional top-down IT or data strategy that also will not change in the short term.
Some of the challenges the model may bring:
- A centralized model can create huge backlogs of requests from the business. The owners of the data sources and thus become the bottleneck of the organization. This results in decisions being made later than desired or without the right information. This is exactly what you want to avoid when you start working data-driven as an organization.
- The business is often hardly involved in the process. So there is often poor communication.
- The gap in data and analytical skills will hardly be filled with a centralized model. The business is given little responsibility and thus has less need to develop this level of knowledge.
- There is often great technical and development knowledge but little knowledge (or time) to engage in creativity, readability and usability so that ultimately engagement with the dashboards can also be increased.
This model is strong on the technical front but often challenging when it comes to innovation and speed.
A delegated model
Unlike a centralized model, more responsibility is given to people outside the central department. This allows for faster innovation. Access to resources or dashboards is controlled and monitored by Site Administrators or Project Leaders. In this model, certifying data sources and content is important when a delegated model is in place. In some organizations, there is a centralized team as in a centralized model that is responsible for certification.
Reasons for choosing a delegated model are:
- More knowledge is spread throughout the organization, but in some cases the level of knowledge lags in certain departments.
- Especially in the case of sensitive data, there remains a core team responsible for certifying the data. This can sometimes cause delayed processes but does ensure reliable data that business users also want to use.
- This model is well suited when you want to make the move to self-service, this can be an intermediate step.
Some challenges that the model can bring:
- It is important to establish a process for how data and content are certified and validated by users.
- A customized training plan is necessary for analysts and end users to create or interpret appropriate content.
- Site Administrators and Project Leaders will need to be trained on what roles and responsibilities there are and what tasks go with them.
A self-service model
In a self-service model, content and data are created by Creators as we know them in Tableau or by Explorers in web-edit. The analytical knowledge is great in the organization and, in addition, many processes are established.
Reasons for choosing a self-service model are:
- End users would like to be able to analyze data themselves and make decisions when they need it.
- Demand for reporting, aka data-driven decision making, is increasing
- A self-service environment fits well when (most) data is not sensitive - an open organization that releases a lot of information.
Some challenges the model may bring:
- Monitoring of the Server environment and the ability to scale must be possible.
- You need more reports that allow administrators to understand who is accessing and actually using which resources.
- Without guidelines, there can be an overload of dashboards on the server. The site must be set up properly so that the right content can be found.
- Training is indispensable when a self-service environment is set up. This applies to Creators as well as Explorers and Viewers in the organization.
- Finally, a good community and support should not be missing. When you start working with data and the development or interpretation of data yourself, you will undoubtedly run into technical or how-to questions.
Which governance model suits your organization?
And how do you strike the right balance between establishing processes and guidelines and a self-service environment? Start by asking the following questions:
- Do we have a process in place for managed data sources and dashboards?
- What components are important for organization to capture?
- What can we improve about current processes?
Want to learn more about governance? Then read more information on our website or contact us directly.