Today's organizations are being flooded with new data from all directions and executives are expected to make smarter decisions with that information.
While small to midsize businesses (SMBs) may find using Microsoft Excel a convenient receptacle for data during their early days will soon realize that spreadsheets, however, are blunt and insufficient to deal with this kind high volumes of data.
There’s now a new crop of business intelligence (BI) tools to fill the gap.
These tools combine all of the sophisticated data hooks on the back end with a new style of front end that combines ease of use with things such as natural language querying to make using BI accessible to anyone.
These tools also provide new data visualization capabilities that let you turn your insights into clear and easily parsed graphics to help co-workers understand your discoveries.
Spreadsheets also fall down when the data isn't well-structured or can't be sorted out in neat rows and columns.
And, if you have millions of rows or very sparse matrices, then the data in a spreadsheet can be painful to enter and it can be hard to visualize your data.
Spreadsheets also have issues if you are trying to create a report that spans multiple data tables or that mixes in Structured Query Language (SQL)-based databases, or when multiple users try to maintain and collaborate on the same spreadsheet.
A spreadsheet containing up-to-the-minute data can also be a problem, particularly if you have exported graphics that need to be refreshed when the data changes.
Finally, spreadsheets aren't good for data exploration; trying to spot trends, outlying data points, or counterintuitive results is difficult when what you are looking for is often hidden in a long row of numbers.
While spreadsheets and self-service BI tools both make use of tables of numbers, they are really acting in different arenas with different purposes.
A spreadsheet is first and foremost a way to store and display calculations.
While some spreadsheets can create very sophisticated mathematical models, at their core it is all about the math more than the model itself.
This is all a long-winded way of saying that when businesses use a spreadsheet, they are actively sabotaging themselves and their ability to consistently get valuable insights from their data.
BI tools are specficially designed to help businesses better understand their data, and can prove to be a huge benefit to those upgrading from what a limited spreadsheet can do.
What Is Business Intelligence?
BI encompasses a fair amount of tools and processes that may not be standardized or that can be vague or nebulous.
Various types of software now offer some form of analytics which can feed into a businesses’ specific needs.
Whether it is user metrics, defining and anticipating trends, as well as predicting outcomes all fall under the umbrella definition of business intelligence.
In short, activities that help businesses turn raw information into actionable knowledge can be tagged as BI.
Now that businesses are generating more data than ever, it’s become more of a challenge to harness that data into actionable BI to increase profits and remain ahead of their competition.
Framed that way, BI as a concept has been around as long as business.
But that concept has evolved from early basics [like Accounts Payable (AP) and Accounts Receivable (AR) reports and customer contact and contract information] to much more sophisticated and nuanced information.
This information ranges across everything from customer behaviors to IT infrastructure monitoring to even long-term fixed asset performance.
Separately tracking such metrics is something most businesses can do regardless of the tools employed.
Combining them, especially disparate results from metrics normally not associated with one another, into understandable and actionable information, well, that's the art of BI.
The future of BI is already shaping up to simultaneously broaden the scope and variety of data used and to sharpen the micro-focus to ever finer, more granular levels.
BI software has been instrumental in this steady progression towards more in-depth knowledge about the business, competitors, customers, industry, market, and suppliers, to name just a few possible metric targets.
But as businesses grow and their information stores balloon, the capturing, storing, and organizing of information becomes too large and complex to be entirely handled by mere humans.
Early efforts to do these tasks via software, such as customer relationship management (CRM) and enterprise resource planning (ERP), led to the formation of "data silos" wherein data was trapped and useful only within the confines of certain operations or software buckets.
This was the case unless IT took on the task of integrating various silos, typically through painstaking and highly manual processes.
While BI software still covers a variety of software applications used to analyze raw data, today it usually refers to analytics for data mining, analytical processing, querying, reporting, and especially visualizing.
The main difference between today's BI software and Big Data analytics is mostly scale.
BI software handles data sizes typical for most organizations, from small to large.
Big Data analytics and apps handle data analysis for very large data sets, such as silos measured in petabytes (PBs).
Self-Service BI and Data Democratization
The BI tools that were popular half a decade or more ago required specialists, not just to use but also to interpret the resulting data and conclusions.
That led to an often inconvenient and fallible filter between the people who really needed to get and understand the business—the company decision makers—and those who were gathering, processing, and interpreting that data—usually data analysts and database administrators.
Because being a data specialist is a demanding job, many of these folks were less well-versed in the actual workings of the business whose data they were analyzing.
That led to a focus on data the company didn't need, a misinterpretation of results, and often a series of "standard" reporting that analysts would run on a scheduled basis instead of more ad hoc intelligence gathering and interpretation, which can be highly valuable in fast-moving situations.
This problem has led to a growing new trend among new BI tools coming onto the market today: that of self-service BI and data democratization.
The goal for much of today's BI software is to be available and usable by anyone in the organization.
Instead of requesting reports or queries through the IT or database departments, executives and decision makers can create their own queries, reports, and data visualizations through self-service models, and connect to disparate data both within and outside the organization through prebuilt connectors.
IT maintains overall control over who has access to which tools and data through these connectors and their management tool arsenal, but IT no longer acts as a bottleneck to every query and report request.
As a result, users can take advantage of this distributed BI model.
Key tools and critical data have moved from a centralized and difficult-to-access architecture to a decentralized model that merely requires access credentials and familiarity with new BI software.
This results in additional analysis becoming available to the organization, that experienced front-line business people can extrapolate and put to good use.
The emerging crop of BI tools all work hard at developing front-end tools that are more intuitive and easier to use than those of older generations—with varying degrees of success.
However, that means a key criteria in any BI tool purchasing decision will be to evaluate who in the organization should access such tools and whether the tool is appropriately designed for that audience.
Most BI vendors indicate they're looking for their tool suites to become as ubiquitous and easy to use for business users as typical business collaboration tools or productivity suites, such as Microsoft Office.
None have gotten quite that far yet in my estimation, but some are closer than others.
To that end, these BI tool suites tend to focus on three core types of analytics: descriptive (what did happen), prescriptive (what should happen now), and predictive (what will happen later).
What Is Data Visualization?
In the context of BI software, data visualization is a fast and effective method of transferring information from a machine to a human brain.
The idea is to place digital information into a visual context so that the analytic output can be quickly ingested by humans, often at a glance.
If this sounds like those pie and bar charts you've seen in Microsoft Excel, then you're right.
Those are early examples of data visualizations.
But today's visualization forms are rapidly evolving from those traditional pie charts to the stylized, the artistic, and even the interactive.
An interactive visualization comes with layered "drill downs," which means the viewer can interact with the visual to reach more granular information on one or more aspects incorporated in the bigger picture.
For example, new values can be added that will change the visualization on the fly, or the visualization is actually built on rapidly changing data that can turn a static visual into an animation or a dashboard.
The best visualizations do not seek artistic awards but instead are designed with function in mind, usually the quick and intuitive transfer of information.
In other words, the best visualizations are simple but powerful in clearly and directly delivering a message.
High-end visuals may look impressive at first glance but, if your audience needs help to understand what's being conveyed, then they've ultimately failed.
Most BI software, including those reviewed here, comes with visualization capabilities.
However, some products offer more options than others so, if advanced visuals are key to your BI process, then you'll want to closely examine these tools.
There are also third-party and even free data visualization tools that can be used on top of your BI software for even more options.
Products and Testing
In this review roundup, I tested each product from the perspective of a business analyst.
But I also kept in mind the viewpoint of users who might have no familiarity with data processing or analytics.
I loaded and used the same data sets and posed the same queries, evaluating results and the processes involved.
My aim was to evaluate cloud versions alone, as I often do analysis on the fly or at least on a variety of machines, as do legions of other analysts.
But, in some cases, it was necessary to evaluate a desktop version as well or instead of the cloud version.
One example of this is Tableau Desktop, a favorite tool of Microsoft Excel users who simply have an affinity for the desktop tool (and who just move to the cloud long enough to share and collaborate).
I ended up testing the Microsoft Power BI desktop version, too, on a Microsoft representative's recommendation because, as the rep said, "the more robust data prep tools are there." Besides, said the rep, "most users prefer the desktop tool over a web tool anyway." Again, I don't doubt Microsoft's claim but that does seem weird to me.
I've heard it said that desktop tools are preferred when the data is local as the process feels faster and easier.
But seriously, how much data is truly local anymore? I suspect this odd desktop tool preference is a bit more personal than fact-based, but to each his own.
Then there's Google Analytics, a pure cloud player.
The tool is designed to analyze website and mobile app data so it's a different critter in the BI app zoo.
That being the case, I had to deviate from using my test data set and queries, and instead test it in its natural habitat of website data.
Nonetheless, it's the processes that are evaluated in this review, not the data.
While I didn't test any of these tools from a data scientist's role, I did mention advanced capabilities when I found them, simply to let buyers know they exist.
For example, Microsoft Power BI is powerful while also familiar, certainly to any of the millions of Microsoft business users.
However, there are several other powerful and intuitive apps in this lineup from which to choose; they all have their own pros and cons.
We'll be adding even more in the coming months.
One thing to watch out for during your evaluations of these products is that many don't yet handle streaming data.
For many users, that won't be a problem in the immediate future.
However, for those involved with analyzing business processes as they happen, such as website performance metrics or customer behavior patterns, streaming data can be invaluable.
Also, the Internet of Things (IoT) will drive this issue in the near future and make streaming data and streaming analytics a must-have feature.
Many of these tools will have to...
