Stacia Varga (@_StaciaV_) is the founder of Data Inspirations, a consulting and education services company that helps organizations manage and use their information assets effectively. We recently asked Stacia for her insight on the evolution of business intelligence, here’s what she shared:
Can you tell us about the mission behind Data Inspiration? Why did you start your company?
Our mission at Data Inspirations is to help people find and use their business data, transforming it into information that helps them make decisions. I started Data Inspirations not only to provide consulting services to a diverse group of clients, but more importantly to teach them how to create business intelligence and analytics solutions for themselves.
How has business intelligence evolved since you started your career? How has the way companies manage business practices changed?
Back in 1999, The Data Warehouse Institute described business intelligence like this: “With a single entry point, decision-makers quickly find, organize, automate, deliver and aggregate all forms of information supporting the decision process … without IT assistance.”
In short, the goal in 1999 was to enable self-service BI, the holy grail of most companies attempting to implement BI technologies. However, at the time, there were very few true successful self-service solutions because the tools available for self-service were few and there were a limited number of people within an organization who understood the data and the tools well enough to successfully use the information.
Every year or so after that, a new trend or marketing buzz word would appear in the BI market. I remember very clearly that in 2010 the emphasis was on self-service BI. At that time, I reflected that it had been a decade after which we were supposed to be providing that capability and only in 2010 were we starting to get closer to that goal. Every vendor was proclaiming the self-service capabilities of their tools, and I thought … finally!
But even then, there were varying degrees of success. At the end of the day, the tools are only useful if data is relatively clean and structured in a way that adequately supports analysis. Some of these tools tried to put this restructuring of data in the hands of the users, which enabled a lot of freedom with data, but still required a better-than-average understanding of the tool and data. Consequently, the self-service BI revolution was not broadly realized at that time, but it was a step in the right direction.
Now in 2016, I think we are the closest to the self-service goal than ever before. Much has changed to make this possible. The combined capabilities of hardware and software to process bigger volumes of data and to support more sophisticated types of analyses has been a significant contributing factor. BI software is easier to use and we can build solutions faster. Furthermore, more people in companies are using BI than ever before. More companies have adopted a data-driven culture and have a greater expectation that a broader set of employees have access to and use information on a daily basis.
An additional observation is that the problems that companies are trying to solve with better access to information have also evolved. They have moved beyond simply trying to count things or events and measure changes in these counts over time, such as you might find in traditional sales analysis. Now they are applying analytical techniques to assess whether observed changes are statistically significant and to use historical information to predict future trends.
Can you share an anecdote about how using business intelligence transformed a business? What’s one of your favorite client success stories?
My favorite story is how one of my clients is transforming the livestock industry by using data to understand whether and how specific traits of parent animals, feeding and treatment strategies and herd characteristics among other criteria affect the health and growth rates of their offspring. As a result of this access to detailed data, which I affectionately call the Facebook of Cows, this client is able to help ranchers improve the overall health of their herds and thereby increase ranch profitability.
What should businesses be doing today to stay ahead of how data will be used in the future?
As the saying goes, you cannot manage what you cannot measure. To add to that maxim, you cannot measure what you do not have or what you cannot find.
Storage costs have decreased dramatically, so businesses can save more than ever before. However, this is a double-edged sword because the ability to save more does not mean that what you can save is worth it. Nonetheless, staying abreast of how non-traditional data sources, such as text, voice or images to name a few, can be used effectively in business intelligence and analytics projects is important to determine what can and should be saved.
Furthermore, the ability to save more data can make it more challenging for business users to know what’s available and how to use it. Implementing technologies that support the ability to manage metadata at the organizational level is going to be crucial. Data quality continues to have an impact on the effectiveness of BI and analytics solutions. Sometimes data quality improvement can be automated and this should be explored. In other cases, assignment of responsibility for monitoring and addressing data quality issues should be a priority. Again, there are various ways that technology can be used to support this monitoring effort that organizations should evaluate.
What are the most common mistakes or oversights you see businesses making in how they use data?
When I teach data modeling for BI and analytics, I emphasize approaching the task from “right to left” because the biggest problem that I see is developing from “left to right.”
Imagine a traditional data warehouse diagram in which you see an array of data sources on the left, a data warehouse and its reporting and analytics layer the middle, and users on the right. A “left to right” development approach consists of trying to design a dimensional model based on what you see in the data sources whereas a “right to left” approach analyzes the ways in which users think about data and ask questions of data and then designs an appropriate model.
The main problem that I see in the use of the “left to right” approach is that solutions can become convoluted and overly complex or completely miss the mark and fail to answer the real questions that business users have.
What Big Data trends and/or headlines are you following today? Why?
I remember back in the late 1990s when the Internet became a big part of our conversation in the industry, but there was still a lot of speculation about what could or should be done to move businesses onto the Internet. Here we are now in 2016 with the hindsight of what worked and what didn’t.
Either way, our world in business and as consumers has changed in remarkable ways. I feel that we are approaching a similar pivot point in which now we can see enough of what is coming to speculate on what will change, but not far enough along to see what will be normal in a decade or two.
In particular, the trends that I find most intriguing and full of potential are cognitive computing and machine learning. As we get bombarded with more and more data, the ability to sift through it all to find meaning that’s actionable becomes more and more challenging. That’s where cognitive computing and machine learning can help and can likely find important information that would be difficult for a human being.
The next logical step as a result of this capability is embedded analytics in which applications can leverage machine learning to predict potential outcomes based on current information and to recommend next steps. For this next step to be realized, dynamic data integration, real-time data access, and data quality on-the-fly, are areas that are important contributing technologies and therefore am watching trends in these areas as well.
What piece of advice do you find yourself repeating to clients over and over?
I’m very passionate about the “right to left” development concept. I’ve seen time and time again how “left to right” fails to meet users’ needs and contributes to project delays whereas “right to left” leads to an excited and engaged business user community.
Another key aspect of this approach is to prototype, prototype, prototype. Meetings with users to discuss requirements at a theoretical level using terms with which they are not familiar are rarely productive. However, when you can show people what a solution looks like if they make design choice A versus design choice B, you can validate that you are on the right track. The business users will move mountains to ensure project success because they want what you’re building for them as soon as possible. It’s very satisfying to hear someone say, “This is what I’ve always wanted!”
Looking for more options for managing your business intelligence? Browse Syncsort’s Big Data solutions.