Decision-making aids Self-service BI vs. custom frontend

If data is to be visualized, sooner or later the question arises as to which tool is best suited for this. One of the existing BI tools can be used for this, which already supports many functions out of the box, or a custom front end must be developed.

The advantage of custom front ends: they offer full flexibility. But how do you make the right decision and on the basis of which criteria?

Criteria for deciding on a suitable data visualization - BI tool vs. custom frontend

What is self-service BI? Self-service BI enables users to create reports and analyses themselves without the need for collaboration with the company's IT department. These tools and applications are intuitive to use.

The following criteria should help you to get a better feel for whether a custom frontend or a BI tool is better suited to your needs. Data visualization is suitable.

There are various knock-out criteria in the decision aid that exclude, for example, a self-service BI tool or a custom front end.

Criteria	Description	Rating	Knockout criterion
Budget	How much budget is available to provide a tool for data visualization? Is the customer prepared to use a license model?	BI Tools: They are cheaper to introduce than a custom frontend. However, the license costs and possible costs per user must be considered here. Requirements that cannot be realized with standard modules can be very expensive. (A precise requirements analysis is important here. Even small points can become a challenge here, e.g. zoom function) Custom Frontend: With custom frontends, more money usually has to be invested initially for development and security. However, the costs are minimized once the system is up and running. If the custom frontend is to be developed further, this must also be taken into account when considering the costs.	BI Tool: License costs Costs per user
Time	What is the timeline for introducing such a tool?	BI Tool: They can generally be used more quickly. The "purchasing process" for tools should be taken into account by the customer, as these can take a very long time. Custom Frontend: This must first be designed, developed and secured before it can be used.
Customer specifications	Does the customer already have specifications or guidelines for data visualization?	For some customers, certain BI tools are already predefined. Customer specifications can severely restrict selection options or make them overdue.	BI Tool: Customer specifications
Editor	Does the user have to be able to assemble dashboards independently?	BI Tool: An editor is included here. However, special users may be required for this, which can again be a cost factor. Custom Frontend: If the department / user is to assemble charts themselves later, this can be a major effort driver, as an editor has to be developed for this.
Embedded Design	Should the dashboards be integrated on other pages? Should the dashboards be adapted to the corporate design?	BI Tool: There are already embedded self-service BI tools such as those from Tableau. https://www.tableau.com/de-de/embedded-analytics. Important: the graphics created must not be too large, as scalability is limited. The self-service BI are usually limited customizable in design. Customer Frontend: On the other hand, there are the customer front ends, which can be customized as desired. They can be used very flexibly, there are hardly any limits, except possibly the budget.
End devices / Responsive	PC Smartphone Tablets ….	BI Tool: Tools may be omitted if the graphics work on certain end devices or have to be responsive. Responsiveness can also be a problem with self-service BI, but manufacturers are making efforts to close this gap. More information on Tableau, for example, can be found here: https://help.tableau.com/current/pro/desktop/en-us/dashboards_dsd_create.htm Custom Frontend: They offer full flexibility, but with increased effort. Before development, it should be clear which end devices are to be supported.	BI Tool: No support for specific end devices
Data sources / software	Which data sources need to be used / integrated? Azure / SAP / Excel / ...	BI Tool: Self-service BI tools usually offer numerous connectors to access a wide variety of data sources. It is therefore possible to use a wide variety of data sources without major integration efforts. Data sources can also be added relatively easily at a later date. Custom Frontend: Custom frontends usually require more initial effort to integrate data sources. However, a custom solution can have advantages for data sources that are not covered by a connector of one of the BI tools.	BI Tool: If there are no connectors for the data source
Quantity model	How many users will use the tool later? Which charts and how many charts should be displayed? Amount of data / size? Streaming? Real-time data necessary?	User quantity: A high number of users tends to speak in favor of a custom frontend if the BI tool has a per-user license model. Here you should consider which user types you need. For example, there is the typical "Viewer" role, which can only view reports / dashboards, and the Editor role, which can also create or edit reports / dashboards. However, there may be other roles such as the Admin role. Charts: It should be checked whether all chart types can be mapped using the BI tool. If not, a custom frontend is again recommended. Data volumes: With large amounts of data, many BI tools reach their limits (e.g. map representations or complex key figure calculations). With a custom frontend, you can react more flexibly to large volumes of data. Streaming: If real-time data is required, the extent to which it is needed should be checked. Depending on whether calculations still need to take place, self-service BI tools may be limited or not available at all. For example, Tableau is currently not able to process real-time data.
Data content	What data is displayed? Complex data? Do they need to be processed? Confidential? Authorizations?	BI Tool: There may be requirements that cannot be mapped with a BI tool or that lead to high costs. As a rule, the authorization structures of a BI tool should be sufficient. Custom Frontend: Authorizations are flexibly configurable.
User behavior	Do you already have experience with a tool? Is training necessary?	Training is required for both options (BI tool and custom frontend). BI Tool: The basic functions in a BI tool are the same. This means that you do not have to start from scratch if this BI tool has already been used in the company / project. In addition, with BI tools, you can fall back on training documents from the manufacturer and adapt them.
Display options	How should the data be visualized? Are there any special mandatory requirements here? Which charts are required?	BI Tool: Although standard software offers many options out of the box, it is limited in terms of display options if very specific ideas arise. Therefore, certain options may not be implemented. However, this also has a lot to do with the BI tool used. K.o. criterion. With Tableau you currently have a lot of options, whereas there are some stumbling blocks with QlickSense or PowerBI. In general, Custom Frontend offers greater flexibility here. Custom Frontend: You can respond precisely to the wishes of the user. There are no limits here, but it is a cost driver if many different charts are required.	BI Tool: Special requirements cannot be implemented
Care	Maintenance of the tool? Support?	BI Tool: BI tools offer professional support for the products that can be used. Custom Frontend: With custom solutions, a support team / structure would first have to be introduced and organized. However, if the solution fails, you can make the adjustment yourself -> independence from the manufacturer.	Custom Frontend: Support team must first be put together
Further development	New features necessary?	BI Tool: Further development / customizing can be expensive if the new requirements are not yet available in the standard software. Subsequent adaptations usually lead to significantly increased costs. It is therefore important to carry out a thorough analysis and also take known future requirements into account. Custom Frontend: The source code / know-how (developer) is available and can usually be flexibly expanded.
Data export		BI Tool: Various options are offered as standard Custom Frontend: With custom frontends, this in turn is a cost driver.	Custom Frontend: Exports are a cost driver

Which BI tool to use and when?

What if the recommendation comes down to a self-service BI tool - which BI tool should be used here? The devil is in the detail: Unfortunately, the perfect self-service BI tool does not exist, which is why you also need to check the requirements and framework conditions when making this decision. It is therefore important to consider the area of application and the existing software in order to minimize the integration effort and take advantage of the compatibility between the software systems. For example, Power BI is best integrated into Microsoft products such as Sharepoint or in the Azure environment. SAP Lumira, on the other hand, recommends a connection to the SAP backend. QlikView / Qlik Sense and Tableau are vendor-independent and therefore offer a wider range of applications.

Conclusion:

Depending on the application, you should carefully weigh up the above criteria and then decide. It is also worth taking a look at the product provider market at regular intervals. The reason for this is that they continue to develop the products and it is therefore a snapshot that may no longer be up-to-date in six months or a year.

From data integration and machine learning to data visualization