Choose the ‘perfect’ BI tool for your organization

1. Define your use cases and business case.
In order to select the perfect BI tool for your organization, it is important to start asking the right questions.
Why is there a need for a (new) BI tool in the first place? What is wrong with the current situation, or do we just feel we need to hop on this crazy train full of data-scientists, machine learning, big data experts, bi tools etc..? How many users are currently working with BI in the organization, which departments?
How many users would ultimately benefit from automated reports / dashboards in a controlled environment? What would be the added value ($)?
How are analysis currently conducted?
Who is currently creating dashboards in Excel? How would a new BI tool impact their daily routine? What is the budget and preferred license plan? Do you want a lightweight frontend tool or something that integrates more deeply with existing software?
Once you are getting answers to these questions, you will get an idea which area currently demands the most attention. You should focus on defining those use cases where most people benefit in terms of efficiency and money.

2. Define your requirements.
This is the step to put your requirements for a BI tool on paper. Take the use cases as your input. What requirements / functionality of a tool do I need to satisfy each end user of the use cases? A new BI tool for data analysts requires most likely more in depth (advanced) analytics capabilities, while sales managers may just be looking for a simple accessible dashboard with a few key KPI’s and some filters to interact. A comparison matrix will give you a good overview per topic and score of each tool. Define for each requirement whether they are (business) critical or not.

3. Start evaluating products and absorbing information.
A lot of work has already been done by others, so start absorbing information! Read the latest reports on all BI vendors from sources such as Gartner, Butler Analytics, PC Mag or G2 Crowd. Most BI vendors offer a product page, a free trial or a demo to get a first impression of their product. This should help you to narrow your BI vendor offers.

4. Work with your user groups and use case scenario’s for a POC.
After evaluating the results of the comparison matrix, you preferably end up with two favourite BI tools. In most cases the experience with the tool for the end user is more important than usage by the BI administrator. If you need 4 clicks to create a report vs 1 click in the other tool, but the end user feels more comfortable in absorbing / using the report in the other, that is what matters most (in most cases).
Since you are the person selecting the tool, it is hard to be truly objective to the evaluation. For a BI specialist some functionality is common sense, while an end user may have trouble interacting with some reports. Adoption of a new BI tool in a big organization is key to success, so now it is time to really engage with your end users and setup a proof of concept.
We are at the moment in the process of the POC. We invited two of our key stakeholders to help start creating dashboards and see what the limitations are per tool plus the key benefits. After this, one tool will be selected and a group of around 20 users will start using the reports on a daily base to make further iterations and gather feedback.

5. Draw your conclusions.
The POC together with the final commercial terms should give you enough information to make a choice considering the right arguments. The whole process to select the perfect BI tool can be very time consuming. In the meantime, new features from tools may be released and even influence your final conclusion. That’s why it is key to always look at your use case and the latest developments of each tool. Don’t get distracted by fancy features though! Will your company even make use of all those exciting new features? Maybe it is more important to have a BI vendor that really thinks with you and has a simple user experience for the end user. A ton of complex features and options won’t help at all with the adoption in an enterprise environment! Perhaps it is better to make your organization data driven to start with.