Early bird tickets for Inspire 2023 are now available! Discounted pricing closes on January 31st. Save your spot!

Analytics

News, events, thought leadership and more.
MichaelPeter
Alteryx
Alteryx

From report writers to OLAP to dashboards to search, the world of business intelligence (BI) and data analytics has seen steady evolution over the past 30 years. The consistent objective has been to make it easier for individuals to get answers to their questions to make more informed decisions.

 

Seems like a straightforward goal, and you would think that after 3 decades of advancement, businesses would be in great shape. So it is troubling to see studies such as this from the International Institute of Analytics showing most companies are still struggling to become more analytics driven.

 

image001.png

Curious about your own analytics maturity?  Take the self-assessment here, It's free!

 

While there are likely different factors at play in each individual organization, I’d like to posit there is a common root cause found in two words of the stated objective: Questions and Answers.

 

Now you may be thinking, “wait a minute, that’s the whole point!”  But is it? Let’s take those words out of the statement: The consistent objective has been to make it easier for individuals to make more informed decisions. Hmm…

 

All of the technologies mentioned above excelled at improving Answers, but they are all still constrained by the Question part. Before a report or dashboard can be created, someone has to decide what questions are being addressed. That will dictate what data is used and how it will be displayed. Search-based BI has further improved flexibility, but you must still start with a question.

 

The question is: what is the right question to ask?

 

Questions are curious things.  They, of course, emanate from what you hope to learn from the answer, but at the same time, they are often influenced by what is already known, or at least expected, based on your experience.  This can result in confirmation bias affecting the decision process. As soon as an expected answer proves true, you are inclined to go with it, make the decision and move on. But what if there was a better decision based on answers to questions you didn’t think to ask?

 

I could go on with this bit of diatribe, but that introduction is already getting long; my purpose was to get you thinking about the whole Question and Answer focus that has been fundamental to every BI solution since the term BI was created. And to set the stage on why Alteryx Auto Insights is a revolutionary solution, not just an evolutionary technology.

 

Since I’ve been focusing on words, let’s start with the name: Auto Insights rather than Auto Answers. The words insight and answer are sometimes used interchangeably, particularly in marketing activities, but a quick look at the basic definitions illustrates the difference:

 

Answer: a thing said, written, or done to deal with or as a reaction to a question, statement, or situation (Oxford)

Insight: a clear, deep, and sudden understanding of a complicated problem or situation (Cambridge)

 

Another way to consider this is answers to questions are essentially 1:1, whereas insights are usually multifaceted, effectively combining the answers to many questions, both explicit and implicit. Let’s consider a simple example to demonstrate how this difference manifests in a traditional data visualization/dashboard tool compared to Auto Insights. 

 

To start, we do require an idea of what needs to be understood, for example: “How did my APAC region do last month?”  (note: for this example last month is August 2021)

 

That question is rather vague, so the typical first step would be a dialog to get to a more specific question:

 

“What do you need to know to determine that?”

“Revenue is important. I’d like to see it broken down by the different countries in the region”

“Great! I can do that!”

 

And at some point, you would probably see something like this:

 

image002.png

 

I implied this dialog happened with a BI developer, which is typical in most organizations.  Certainly, if you can learn to navigate the UI below, you could create the chart yourself.  But you would still talk yourself through the dialog.

 

image003.png

 

That chart accurately answers precisely what was asked. But does it tell you what you want to know? Probably not.  And as is usually the case answering one question leads to several more, such as what happened in Indonesia? But here again, you can’t answer that vague question without first getting to more specific questions.  And so on and so on... This represents a common cycle to create dashboards that, in the end, still address only a specific set of questions.

 

It illustrates the point I alluded to at the start: the need to define specific questions is the Achilles’ heel of BI, and I believe the reason it has never lived up to the promise.

 

Now let’s see how Auto Insights would handle this.

 

First, understand that you don’t need to build anything, simply point Auto Insights to your data, and its AI models generate everything. As this example data contains more than just APAC, you would need to tell it what to look at (APAC Countries) by simply selecting those at the top of the page. That requires just a couple of clicks, and Auto Insights gives you this:

 

image004.png

 

And this…

 

image005.png

 

And this…

 

image006.png

 

And for good measure, gives you these too…

 

image007.png

 

So, do you now have “a clear, deep and sometimes sudden understanding of the … situation” in APAC?  Maybe that wasn’t sudden enough: so, what did happen in Indonesia?

 

One click and Auto Insights shows that Amos closed a monster deal with Kozey-Skiles.  Go Team 1!

 

image008.png

 

Also, notice not just what it told you but how; it created a comprehensive story of revenue in APAC last month.  While there are charts, the emphasis is on detailed descriptions in natural language that provide rich context around the data being displayed. This is important because while charts can be visually appealing and convey large amounts of data efficiently, they must be interpreted, which is a skill on its own and has its own risk of error. Hard to misinterpret something when it is clearly articulated, like in the examples above! 

 

And this is only one set of insights available in this data; all provided without needing to know what questions to ask nor build anything to reveal the insights.  I was not being hyperbolic when I used the term revolutionary.

 

Insights, not answers, are what is needed for your organization to truly excel. Now you can get them automatically!

 

Comments