There is a myth in Business Intelligence. The perception when we speak about Centralized BI is that IT will be a bottleneck in the process. Perhaps we can do a small recap of BI history to understand the reasoning behind this myth.
Back in the day, Business Intelligence used to be centralized by IT, but there were zero self-analytics capabilities. This meant that everyone relied on IT to analyze data, create reports, maintain the systems, etc. The process of data-analysis to insights was too slow. Business users would receive outdated reports and it was hard to make data-driven decisions at the right time. On the other hand, IT wasted a lot of time and resources creating time consuming reports.
People thought federated BI would be the solution. If users could work in their desktops without the need to go through IT to get a report, they would save time. In the short term, this seemed like a good option. But in the long run, we have seen that it is not. Data gets lost in silos when users work in separate desktops. Taking IT out of the equation resulted in no governance and therefore in data security breaches and other problems.
Having learned from past mistakes, the trend in the industry is to go back to centralization, but this time combine it with state of the art technologies and self-service capabilities that will save time and maintain security. The real solution to the bottlenecks is using a centralized, unified web based BI solution. This is the smoothest funnel of data that organizations can have.
Let’s Compare & Bust the Myth
In order to achieve a better understanding of why centralized BI is not causing bottlenecks, let’s compare the two main BI models: Centralized and federated.
In terms of connection
- While in centralized BI users work connected, in federated BI they work in silos – meaning the findings stay in the desktop of the analyst instead of sharing the knowledge in a centralized environment. This means that the analyst cannot be seen as a knowledgeable person that is critical in the analysis of facts and driving decisions. How can it be that silos are not causing bottlenecks in the organization? If people used to say that IT was a bottleneck, then in federated BI every single user working in a desktop becomes a bottleneck.
In terms of accuracy
- In centralized there is one version of the truth. In federated there are thousands of Excel spreadsheets that cannot be merged. There is no sharing of knowledge when users work disconnected. The normal scenario in federated BI is that an analyst sends an Excel file to another analyst; then they need to keep many versions of the truth. This is commonly known as Excel hell. It is basically a data disaster. In centralized BI there is only one version of the truth. There are no unnecessary copies of data. People can share knowledge and drive analysis in the same, accurate context. Are Excel disasters and many versions of the truth not bottlenecks? Is it possible that having to maintain hundreds of versions is better than a centralized version?
In terms of usage
- Centralized BI assures high usage because everything is in a unified solution. People can learn from one another. An analyst’s findings can inspire another analyst and they can open discussions in the context of the data. They can enrich their analyses with annotations and conversations. All this without jeopardizing security. Users can work with confidence, knowing that there will not be any data breaches or leaks of information. So imagine if each analyst needs to constantly request access to data from IT? In centralized BI the access permissions are set from day one, so no one needs to wait for someone else to get the information he has access to.
In terms of governance
- In a centralized platform, security and governance are IT’s responsibility. Analysts can share and distribute reports easily and quickly without worrying about security breaches. They can distribute a dashboard to thousands of sales people and every sales person will only see the data that they have access to. The job of the analyst shifts from answering the same questions all the time to distribute new findings. They can do this in a centralized system where people can learn from each other. And IT can administer the data transfer in the organization. So if the analyst is concerned of security breaches, how is he not a bottleneck?
In terms of deployment
- Centralized BI is easy to deploy and scale as more users want to use it. When IT want to provide access to new users, they do not need to install desktops. They only need to give them a name and password and set the data access permissions. IT can forget about installation and maintenance, as there is only one unified web solution to take care of. How can installing and maintaining desktops is less bottleneck than having one unified web solution to maintain?
After going through the key differences between centralized and federated BI and analyzing them in terms of “bottleneck potential”, we see that centralized BI as a bottleneck is just a perception, but not a reality. Federated BI is a bottleneck for the organization, while centralized BI will allow the organization to work connected towards meeting its goals.
But what does it mean to work connected?
Working together means setting clear decision rights and letting each team do what they do best.
- Business managers have authority over what services are delivered.
- e. Need beautiful dashboards for the executives
- e. Need KPI alerts for the sales people
- IT has authority over how the services are delivered.
- e. the system should be federated or centralized
- e. the system needs to be on premise or on Azure
If we learn to divide roles and let each team do what they are best at, we achieve our goals faster and better… this is centralization. Everyone working together, connected, successfully. And it is only one solution that is web based, that is the key.
On top of having all users working connected, for an organization to meet its goals, it must match the security and accuracy of centralization with the user experience that users expect to have. They should be able to benefit from dashboards, data mashups, connectivity, analytics, and KPI monitoring. And they should also benefit from critical and important enhancements, such as advanced analytics, collaboration and tailored solutions. It is important for the organization to use a BI solution with a centralized architecture that is secure and governed, and that can be deployed quickly.
Panorama Necto matched both worlds, providing state of the art BI within centralized BI.