Your current Business Intelligence (BI) and Data Visualization requirements today will not be your requirements in the future. If you evaluate a BI tool, you need to make sure the tool you choose will grow and expand as your requirements grow and expand. Regardless of the Business Intelligence application, you decide upon, general flexibility will be essential in the long run as it is tough to predict your future needs.
Quick disclaimer. For those looking at BI through the lens of a 'business user,' it's good to understand that there is much behind-the-scenes technical work happening to allow you to view dashboards or even make your own dashboards by performing self-service tasks in your BI tool. The technical questions below need to be asked, and I highly recommend that IT be heavily involved in the early stages of any BI project. No amount of 'Ease of Use' will replace these technical considerations at the beginning of the journey or in the long run.
In my 15+ years in the industry, I've noticed that every BI vendor everywhere will open with a point about being flexible and customizable. The problem is that this term is nebulous, and it's difficult to understand what it means to be genuinely 'flexible.'
Flexibility is one area where you should be incredibly diligent in making sure the tool meets your needs. Let us look at a few questions you should be asking your BI provider to ensure that they are flexible in all areas of the application.
Today, I will give you a list of questions to ask your BI vendor to make sure they are serious about being flexible. If I pose an unfamiliar question, maybe it's worth looking into the requirement as it might be something that isn't on your radar yet.
I'll keep the format for these questions simple, the question "Q" and why it's essential to have an answer to these questions.
Q - Where and how is your BI tool going to be hosted?
Why? – Depending on your need, the location of your BI is of significant concern.
Do you have High-security needs? It would probably be best if you had an on-premises deployment.
Is your IT too cumbersome or busy? Cloud might be the way to go.
Also, if the vendor you are looking at only offers limited hosting methods, what will you do if your architectural needs change in the future?
Here is a video I did on the subject if you want to dig deeper.
Q - Do you connect to the data source that I need now?
Why? – Fairly obvious question, but many people test with simple excel data and then wonder why they struggle later at implementation when they try to use their actual data source. Excel snippets are the most straightforward possible data source and least likely to cause problems. Use actual data and perform some real-world testing.
Q - What if I need to connect to a data source that you (the BI vendor) have never even heard about before?
Why? – Things change. You may adopt new tools or decide that you need to use existing sources that were not initially considered during the evaluation. What will happen if you want to use something that the BI vendor doesn't currently offer? You need to make sure you're happy with their answer to this, as you don't want to be at the mercy of their release schedule or lack thereof.
DATA PREPARATION and MODELING
Q - How do you work with multi-dimensional or OLAP data? (if you use it)
Why? – Not all data is 'flat,' and some data comes in pre-structured formats for better analysis. If you have this kind of data, be absolutely sure that you can use it as is. Many tools will ask you to 'flatten' this data first, which defeats the purpose of structuring it like this in the first place.
Q - Do you support real-time data?
Why? – Believe it or not, some tools do not allow for real-time data refreshes or will charge you extra for it. Real-time is particularly important if you are looking to use BI for operations or performance management dashboards.
Q- Do you support role or user-based security?
Why? – It is a general requirement to filter a dashboard based on who is viewing it automatically. In this case, the viewer should only be allowed to view the data they are authorized to see. While many BI tools will have a way to do this, several require you to code this much-needed functionality rather than using a built-in mechanism.
Q – Are there options for data cleansing, blending, merging, and general data transformations?
Why? – Everyone's data is a mess and is seldom in alignment with your organization's reporting needs. This mismatch means that someone must take time to modify your data, which can mean either requests for IT or maybe the acquisition of additional tools to help. Some BI tools have a data preparation or ETL layer that will aid in this work. This functionality is a live-saver.
Q – How does data governance work, and how can it help your self-service users understand the data?
Why? – If you want to perform self-service BI or empower other users to do this, you need to investigate the data governance capabilities of a BI tool. It is a big subject, and here is my video on the matter.
Q – Does data correction exist to help temporarily relieve data issues?
Why? As mentioned earlier, most people's data is a mess, and it's very seldom that additional 'data work' isn't a requirement once you start developing reports. Data correction is a fantastic way to prevent downtime when you see a data issue. If a number is wrong, you can 'correct it' while simultaneously looking into the cause and potentially fixing the backend issue. Without this, you might have the take the report down to prevent people from making poor decisions based on these incorrect values. It's just a nice temporary stopgap.
Q – What are your write-back capabilities to support workflows and actions?
Why? If you're new to dashboards, you're probably not too worried about write-back, but this is one of those things that comes up as your requirements start to get more complicated. Data Input is 100% worth looking into if you want BI that will grow with you.
Q – Can I augment my data with R, Python, or Script to solve complex requirements?
Why? The adoption of Python and R is one of the best things to have ever happened in the Business Intelligence industry. These languages are massive and can allow you to do anything data-related, from augmentation and profound analysis of your data—a must-have.
Q – How do you support What-If Analysis?
Why? What-If Analysis is the type of capability that separates flexible BI tools from just offering the bare minimum. What-If Analysis allows you to predict results and be proactive with your decision-making.
Q – How can developers add new formulas to the BI tool? What is the expectation if a formula is unsupported out of the box?
Why? – No BI tool will ever provide everything you need out of the box, as everyone has different requirements. Formulas are a particular area of concern because you might need to calculate something with a specific need. It would help if you had the flexibility to allow for custom formulas.
Q - Do you support Ragged Hierarchies?
Why? – Sometimes, hierarchies can be 'unbalanced' or 'ragged.' Many BI tools will only support the simple case version or hierarchies. If they support more complicated versions, they will often force you to write a significant amount of code to get your data into a format that BI tools can ingest for use with these hierarchies. Proper hierarchy support is yet another non-obvious use case that will probably not come up in a cursory evaluation but can cause significant headaches down the road.
Q – Do you support advanced data visualizations like Cycle Plots, Radar Charts, Sanky Diagrams, Correlation Matrixes, or other advanced visualizations?
Why? – Bars, Lines, and Point Charts should undoubtedly make up most of your visualizations on your dashboards, but they are not the whole story. Depending on your analytics needs, there are many powerful options available, and they certainly shouldn't be ignored.
Q – What can I do if I need a chart type that you don't offer?
Why? – Like some of my other points above, no BI tool can provide everything you'd ever need out of the box. Some BI tools have more visualization options, others have less, but you should certainly know if it's possible to add new visualizations as your needs change, and of course, how!?
Q – Is there support for mobile responsive design?
Why? – Mobile viewing is here to stay, and it would be very inappropriate to choose a BI tool that doesn't correctly support mobile viewing going forward.
Q – Can you automatically shift a user between the mobile and desktop experience?
Why? – When it comes to mobile, there are typically two trains of thought.
1 – One dashboard should be created and should automatically morph itself between desktop vs. mobile use.
2- You should build two dashboards, one for mobile and one for desktop.
The reason behind this second mindset is that people use mobile dashboards differently than they would desktop. On a desktop, you have time, and you're more willing to drill in and look for insights.
You're probably about to walk into a meeting in the mobile use case, and you're more interested in summary information. If this is the case, you might as well build different dashboards to meet each need, and the BI tool should automatically shift you to the right one based on your device.
Q – How easy is it to add context like Period over Period comparisons?
Why? – In Business Intelligence, context is king, and there is no better way to show context than with period over period (PoP) comparisons.
Many BI tools don't automatically perform PoP and require you to modify your source data to meet every PoP variation that the user might need to see. Forcing a data change to support PoP is insanely impractical, and the BI tool should automatically handle this.
Q – Do you support report Bursting on top of the traditional scheduled reports?
Why? – In many organizations, the building of reports is only part of the picture, and many companies struggle with adoption. BI tools with scheduled reporting and scheduled personalized reports (usually called Report Bursting) are a great mechanism to get content in people's hands and create data-based decision-making behaviour.
Q – Can content like dashboards be embedded in another application like a portal?
Why? – One of the best ways to ensure that your content consumers have the most optimal user experience is to keep them out of the business intelligence tool. Embedding finished content and allowing users to view dashboards and reports as part of an existing website or portal gives users a streamlined experience where they don't need to navigate a BI tool to find content.
Q – How can you embed HTML content in your dashboard?
Why? – Sometimes, you have other outputs that would make sense to show on a dashboard. Perhaps an IT monitoring tool with a particular visualization would make sense to be viewed beside a traditional dashboard? Some of the best BI tools will allow you to bring in custom web-based content and blend it with the dashboard as if they came from one place.
Q – How do you deal with proper development methodologies like version control and change management?
Why? - Do you have one developer who will create all business intelligence content, or do you work in a team? In a team setting, you need to make sure that the Business Intelligence tool supports features so that your team is not stepping all over one another while they are building content. How much time are you willing to waste due to a poor workflow?
Q – Can you perform administration tasks using an API rather than the simple user interface?
Why? – When embedding your BI in another application or simply trying to bring all of the IT functions into one place, you may want to perform certain functions using code instead of a provided interface. Also, administration APIs are fantastic for automating large and tedious tasks.
Q – What kind of control do you have over the setup of the application? For example: Can you change the query timeout or white label the application?
Why? – It's great that the BI tool you choose can build dashboards, but can you customize the tool itself? What works for one organization may not work for you, and having precise control over the way the BI application looks and behaves can help it better match your business needs.
As you can see, there are many questions that you need to be asking to make sure that a vendor is truly flexible. Nothing hurts me more (as a consultant) than watching a company that clearly has advanced BI needs choose a simple BI product because their evaluation was misguided or way too focused on ease of use rather than flexibility. Hopefully, this list gives you a good starting point to doing a deeper BI evaluation so that you can choose the right tool for your needs and the future.
About the Author
Jeff Hainsworth is a Senior Solutions Architect at Dundas Data Visualization with over a decade and a half of experience in Business Intelligence. He has a passion for building, coding and everything visual – you know, shiny things! Check out "Off the Charts... with Jeff", his platform for great content on all things analytics, data visualizations, dashboards, and business intelligence. There’s something for everyone!Follow on Linkedin