What are you Looking for?
Data time: When should you start thinking about a Business Intelligence tool?
This post aims to help business people or technical leaders with no business intelligence (BI) expertise understand when to start thinking about BI and how much it may cost. Here, we will compare a variety of BI tools: Google Data Studio, Power BI, Tableau, Pentaho, and others.
There are tons of products already out there, and something new pops up every day, so understanding which one is right for your business might be tricky.
To help you get a better picture, we first need to answer some basic questions and go through a little historical retrospective. If you want, you can skip this intro and fast-forward directly to the section of the article where the available solutions are listed. So, here we go.
What is Business Intelligence?
Business intelligence systems help decision-makers analyze data and stay on top of their business decisions.
But when should you start thinking about BI?Well, it’s when you need to get useful insights from your data. When you want to stay on top of what happens in your business and understand the trends. When you need quick turnaround on your decisions, which is essential in a fast-changing world. The better insights you have, the better decisions you can make. Getting insights takes having the right tools and solutions that will help you understand the data better.
And what about the technical side?
From a technical perspective, the many BI components always evolved in parallel with data storage and processing systems, among the most important of which is RDBM (relational database management system). As the data volumes grew, the need for more sophisticated (and, consequently, more expensive) data transformation solutions arose. That’s why historically BI has never been cheap.
New technologies such as cloud computing as well as modern processing and data storage systems have changed things in BI.
Modern business intelligence systems cover a broad range of on-premise and cloud-based architectures, whether with or without ETL. Take a look at the stage stack for comparison:
As you can see, modern approaches have virtually no borders or split components. You can connect your business data to a BI dashboard directly and get some insights. It simplifies things a lot, but there are still some important factors to consider.
Choosing a BI tool that’s right for you
To pick the right BI tool, you basically need to ask the right questions, such as the following:
- Data source
This is where your data comes from.
Does the BI platform you’re considering support your data sources?
Is it easy to extend if it doesn’t offer what you need out of the box? - Processing/ETL (extract, transform, load)
Even buried deep inside the cloud away from the user’s eyes, ETL and data processing are still important.
What if you need to pre-process data to get insights from your dashboards?
What type of data transformations are available?
How extensible is it? - Storage
How big might you need to go?
What is your final dataset size, and can the BI platform handle it? - Sharing strategy/price
You must understand how your dashboards will be delivered to end-users (and the permissions the users need to see the data). Since the price will depend on the strategy you choose and what is on offer in the market (cloud BI platforms are dominating the market at the moment), this is something you need to consider at this stage, too.
What are the options?
There are plenty of formal comparisons of BI tools. However, we at Broscorp decided to make our own shortlist based on the best practices established in our company.
Google Data Studio
We always start by checking if Google Data Studio might suit our problem. It’s free, it looks nice, and it has options for custom visualization development.
Where it excels
- Designed for small/medium businesses
- Easy design and development of new dashboards
- Integration with Google services — worth a try if you keep your data with Google
Where it falls short
- Row-level security (if you need different data subsets depending on the user)
- Self-service BI (development and sharing of new dashboards by other users)
- Integration of corporate authorization with your BI solution
- No ETL to prepare the data for complex calculations
Price
It’s free, so it’s always the first thing we suggest trying.
What else?
The solution that addresses most of the shortcomings of Google Data Studio is Power BI, so let’s look at it next.
Power BI
It is the most comprehensive and feature-rich BI solution available on the market. It offers one of the best analytic engines out there, as well as built-in ETL. Suitable for self-service BI, it supports advanced use scenarios, e.g., row-level security and scheduled data refresh.
Where it excels
- Designed for small to large businesses
- Complex role-based security
- Self-service BI
- Integration with MS services (e.g., Azure AD, OneDrive, Sharepoint)
- A good option for idea evaluation or PoC with nice-looking dashboards done with minimal effort
- Great mobile experience with dashboards accessible via a mobile app
- Built-in ETL engine for cases when you have diverse data sources and expect basic ETL, but have no infrastructure to run it
Where it falls short
- Because of tight integration with Azure Cloud, it can be pretty expensive to maintain your own authorization to share dashboards outside the Power BI web/mobile app (at the time of writing, a dedicated Power BI server could set you back ~$5,000 per month)
- Otherwise, Power BI gives you all the tools you need for successful business intelligence
Price
$10 per user monthly or ~$5,000 for a basic cloud instance of Power BI.
What else?
Looker offers pretty much the same feature set — check it out if you don’t need Azure — and Tableau is another strong competitor.
Tableau
One of the oldest market players out there, Tableau offers a good set of features — enough for most use cases.
Tableau has mostly the same strengths and drawbacks as Power BI. Choosing between the two, we take into account the following:
- The need to develop custom visualizations. Unlike in Tableau, it’s doable in Power BI, since it is built with the popular D3.js library.
- The ability to update underlying data in the database directly from the dashboard. It’s a tricky and unusual scenario for BI tools, but it’s available in Tableau and could become a life-saver for some teams
- AWS Cloud integration problems. Power BI is incompatible with some AWS cloud services natively, e.g., AWS RDS. As a result, we need to use a gateway to update the data in the dashboards that use them, which would add to the final maintenance price.
- Price. Despite the close initial price per user ($10 in Power BI vs $12 in Tableau), one should carefully calculate the final BI solution price before making a choice. The factors, such as the number of users and data refresh scenarios, should be taken into account because it could significantly change the final price.
Pentaho
One of the traditional BI tools, it has a rich toolset. It is open-source and extensible and has a free Community Edition. Compared to the other tools listed above, the development takes longer. Still, this one could be an excellent option for supporters of free software wanting to have full control over the infrastructure and code.
Where it excels
- Workable for medium to large companies and government organizations, but might be too expensive to bring any value to small businesses, as it might require additional development and administration
- A fully on-premise solution with full control over the data and the code
Where it falls short
- You will need developers to support solution installation and development
- You don’t have an infrastructure and sysops engineers
- You want a rich feature set out of the box
Price
Community Edition is free. However, since there is no cloud offering, the basic infrastructure costs are comparable to a BI server installation, e.g., renting an EC2 instance 24/7 with minimal requirements will cost you $750 per month, or $370 per month with an upfront 1-year payment.
What else?
Another open-source beast is Apache Superset. A more recent solution than Pentaho, it is potentially more promising in the long term. But we will look at it in greater detail in our next post. Stay tuned 🙂
Comparison matrix
To sum-up of the post, check out the feature matrix we use at Broscorp while picking a BI tool for a project. This list is by no means full, and we might extend it in the future[0], but the bottom line is that tool selection is a crucial choice we always make first.
[0] — There are some big names in this field like IBM Cognos, Microstrategy and QlikView, that are worth mentioning. We will try to add in-depth analysis of their features later
[1] — Pentaho has a separate component called PDI (Pentaho Data Integration) responsible for ETL.
[2] — Here I refer to PowerBI premium feature with dedicated instances deployed on customer’s Azure instances.
[3] — I didn’t include hosting costs here because it depends very much on existing infrastructure.
[4] — Some features supported out of the box, some of them require additional development effort