“Simplicity is the ultimate sophistication.”
- Leonardo da Vinci
Let’s examine one of the underlying premises in modern business technology, namely that the more data you collect, the better your decisions will be. This logic dictates that if you only collect enough data across the entire enterprise, you will be able to optimize each of the decisions that impact every line of your P&L.
If this were true, by now, we would be getting pretty close to achieving the vision of universally optimized decision making. In the last year alone, we have generated as much new data as has ever existed before in history. And in the coming year, we will double the total amount of stored data yet again.
Furthermore, we’ve hired teams of data scientists and invested heavily into advanced business intelligence, statistical and AI tools to help us turn all that data into real business insight. Yet, despite this continued doubling in the amount of data, odds are your organization isn’t 2X smarter than it was 365 days ago.
Why is it that despite having so much more data, so many business decisions still happen as they always have? In silos. Driven more by habit and by hierarchy than by data-guided analysis and collective input. What’s missing from this equation?
I think we can agree that the idea of “more data = better decisions” is a bit of a myth. But that doesn’t mean we must give up our dream of optimizing decision making across the enterprise. It just means we need to understand the limits of big data and find solutions to break through them.
While powerful, big data and the corresponding systems of expertise and tools combine to form barriers that limit the impact they can have on your business. For example…
- #1 Big Data needs… a LOT of data. The statistical tools used with big data require astoundingly large amounts of data to provide good predictions. There need to be several hundred occurrences of a given event in order to create a conclusion. And those occurrences may be buried within millions of events. And yes, there are areas of the business where collecting massive amounts of data is easy… and these areas have often been the first targets of big data. But what happens when you are trying to do something relatively new? What if the data you have is relatively spotty or is measured by the number of lines in a spreadsheet versus the number of terabytes of storage? Despite massive data growth, the lion’s share of your business decisions operate in settings of small- and medium-sized data. Settings where existing tools simply don’t offer enough value.
- #2 The Black Box Nature of Big Data Tools Limits Business User Input. Ideally, analytical decision making should be like a conversation. The data should give you an answer, and you should be able to provide additional intelligence (first-hand business experience, deep understanding of the situation) to hone those answers closer and closer to an actionable decision. Yet, the vast majority of AI tools operate in a black box. They provide an answer, but they don’t show you the WHY. This makes it very difficult for business users to challenge those results and iterate.
- #3 Complex Big Data Tools Present Barriers To Business Users. That ideal conversation is further hindered by the complexity of so much data and the tools and expertise needed to make sense of it all. For instance, neural networks are many things, but they are not simple. Big data and advanced statistics require tools far too complex for the vast majority of business users. So, we often rely on data scientists and analysts to handle these “data conversations” on our behalf. This indirect and asynchronous problem solving works well enough for certain challenges. But this arrangement simply isn’t designed for the speed and data scarcity surrounding so much of your business.
There is no doubt that Big Data has a significant role to play for many organizations. Yet, the clear limitations involved have led to a search for new approaches to pick up where Big Data’s effectiveness begins to fall off.
- #1: Algebraic Geometry & Data Scarcity. You can relax. As much fun as some math nerds (myself included) love to talk about abstract concepts such as Algebraic Geometry and hypercubes, we won’t turn this article into a math lecture. Just know that practical applications in the field of Algebraic Geometry enable us to recognize signal from much smaller amounts of data than the majority of statistical and Big Data approaches. In effect, tools based on Algebraic Geometry, such as MondoBrain, significantly reduce the need for large and highly complex data sets, enabling powerful insight across a much wider range of decision making.
- #2: Algebraic Geometry and Explainability. Another benefit of tools based on Algebraic Geometry is that the results are all clearly explainable. Rather than the “black box” of most artificial intelligence, you can easily see the robustness of the answer AND you can see which factors contributed and to what degree. This clear visibility enables business users to understand the WHY behind answers and to add and adjust constraints in line with business reality. This back-and-forth actually enables that essential “dialog” of iterative decision making.
- #3: Simple, Easy To Use Interface. Even with the inherent advantages of Algebraic Geometry, if a tool is packaged in a complex way, it still places those advantages outside the reach of most business users. This is why MondoBrain is designed using color codes and common interface concepts associated with basic data visualization. No need for learning programming languages or long training courses. If you can use a dashboard, you can use MondoBrain to optimize decisions. We’ve also added collaboration tools to open review and input by peers around the organization to enable broader perspectives, consensus, and compliance. This is why we talk about MondoBrain harnessing the power of artificial, human and collective intelligence. It’s truly the combination of all three of these that results in optimized decision making.
As previously stated, we are big believers in Big Data as a critical development that will continue to grow in importance. Yet it’s vital to understand that the hugely complex world of Big Data won’t give us all the answers.
This is exactly why we created MondoBrain. There are so many facets of business characterized by data scarcity, speed, and the need for human / collective intelligence. In these areas, better decisions won’t come from added complexity, but rather from deriving value from the data you already have using simple and accessible tools.
As Peter Seeger famously said, “Any darn fool can make something complex; it takes a genius to make something simple.” In this case, the reverse is true too. It takes something simple to enable your organization to operate like a genius.