BI and analytics are both umbrella terms referring to a type of data insight software. Many providers use them interchangeably, but some use them in conjunction, claiming to offer both business intelligence and business analytics.

This of course makes us wonder: what’s the difference?

The short answer is: it depends on who you ask. As Timo Elliot, Innovation Evangelist at SAP so eloquently put it, “everybody has an opinion, but nobody knows, and you shouldn’t care.”

Elliot got it mostly right. You should care about the terms used to describe the applications you’re buying, particularly when it comes to embedded software destined to become part of your product.

The short answer is: it depends on who you ask. As Timo Elliot, Innovation Evangelist at SAP so eloquently put it, “everybody has an opinion, but nobody knows, and you shouldn’t care.”

It’s important to enter sales conversations armed with the knowledge that BI and analytics have no universal definitions. It’s up to you, the buyer, to probe for clarity around these and other ambiguous terms.

To aid you in this endeavor, let’s get into some common claims about the differences between business intelligence and business analytics.

BI for the Past, Analytics for the Future

There are four modes of data analysis: descriptive, diagnostic, predictive, and prescriptive.

Descriptive analysis, as the name implies, attempts simply describe what events took place, according to the data. Diagnostic analysis attempts to explain how or why those events happened. Both are concerned with the past.

Originally published here.

Predictive analysis uses past data to forecast what might happen in the future, and prescriptive analysis “takes that data and goes even deeper into the potential results of certain actions.”

Here’s how we might look at the same data set using these four analytic modes:

  • Descriptive: Sales increased by 2% this quarter.
  • Diagnostic: There was exceptional variation in macro-widget sales in particular, and from that we conclude that the 2% boost was largely due to this product.
  • Predictive: A linear regression of macro-widget sales suggests we’re on course to sell an additional 200 units next quarter.
  • Prescriptive: Adjusting the shipping schedule to coincide with peak business hours will likely increase profits by an additional 0.7% per quarter.

One school of thought distinguishes BI and business analytics along these past/future lines. Business intelligence solutions, they say, are for performing descriptive and diagnostic analyses whereas business analytics (BA) tools specialize in the predictive and prescriptive modes. But of course, an application marketed as either BI or BA could very well be capable of all four modes.

Take Excel, for example. Microsoft calls it “a spreadsheet” — neither BI nor BA. And yet, it can describe, diagnose, predict, and prescribe. As we learn from David Langer and Caitlin Johnson in this episode of Data Talks, you can use the spreadsheet to perform linear and logistic regressions. Both of these are predictive statistical tools.

“You can also do prescriptive in Excel using the Solver,” says Langer, “to, for example, optimize a supply chain.” Solver is an add-in you can use to perform if-then analyses; conditional logic is the key to prescriptive.

Statistical Nesting Dolls

So we know it’s not safe to assume that business intelligence and business analytics refer to different analytic modes. It’s also not safe to assume they’re two separate, unrelated entities.

Some consider one to be a subset of the other. Elliot says that at SAP, “business analytics” includes business intelligence, along with “data warehousing…enterprise information management, enterprise performance management, analytic applications, and governance, risk, and compliance.”

So we know it’s not safe to assume that business intelligence and business analytics refer to different analytic modes. It’s also not safe to assume they’re two separate, unrelated entities.

Dr. Rado Kotorov, by contrast, classifies analytics as a “function of a BI” — just one of the many things BI platforms do.

So which belongs to which? We’re left with a sort of Russian nesting doll conundrum in which each doll paradoxically fits inside the other.

Automagical Reporting

Last but not least, you may find “analytics” used to denote the automatic analysis of a data set.

This is less common in enterprise and OEM software than in SaaS, but “having analytics” means having built-in reports, dashboards, and data visualizations designed specifically for the data in question.

Think Google Analytics. The software is designed to handle website data, and all its reports automatically surface the insights that matter most to site admins, marketers, and web devs. (Note: thought they’re largely descriptive and diagnostic in mode, GA reports are, as the name mandates, considered “analytics.”)

Most web-based applications, particularly those with admin roles, have some sort of “analytics” feature. Sometimes these features are built in house specifically for that application, and sometimes they’re third-party solutions made to feel native to the host.

What they generally have in common are canned reports, which are premade expressly for the data set. All the user has to do is view them; they do all the work. This automatic data manipulation, which can be in any analytic mode, is sometimes what vendors and consumers mean by “analytics.”

Get Specific

Rather than rely on ambiguous umbrella terms like “business intelligence” and “analytics,” use the bolded words above to get clarity around the capabilities you care about. As I’ve said before, the only way to hack through a forest of jargon is with a machete of specificity, so don’t be shy about demanding detail. As the consumer, it is your right.