Guest Post: Data-Driven to Win

Is your use of data equipping your organization to be a comfy sedan or a Formula 1™ race car?

Data. Whatever you may or may not know about it, it’s a safe bet that you’ve seen, heard, or read pundits say that (a) it’s everywhere, and (b) that any organization that doesn’t embrace it, especially its latest derivative Artificial Intelligence (AI), will inevitably go the way of the dinosaurs. True or not, other than raising the alarm, there is little follow up in the way of practical guidance regarding either how these predictions apply within the context of your own organization, or what you can do about it. We can do better.

We can compare your organization to driving and maintaining a car. As a responsible driver, you want to be safe on the roads, not exceed the speed limit, try to save on gas, and know when your car needs maintenance.

But what if the car you’re maintaining isn’t your own, but a Formula 1 race car? This is a new situation: as well as keeping your drivers safe, you’re also there to beat the competition.

Most data and analytics solutions are designed to provide a “family sedan” level of information to those operating the organization—ensure everything is operating well and try to foresee any potential problems. Many organizations, however, have additional priorities that are more like those faced by a Formula 1™ team. However, these needs are not well served by today’s typical Analytics practices and Business Intelligence systems.

Data and Dashboards: Looking backwards in time to support current operations

Let’s accept as given that your organization probably has access to more data than it knows what to do with. Modern IT and Communications (ITC) systems produce logs, summary and tracking data as part of their normal operations, and there are many more data sets that can be acquired either in the public domain or from commercial vendors. Almost certainly, today, availability of raw data is either not a problem in most organizations, or not one that cannot be solved (of course, access to data, along with assuring high levels of accuracy, completeness, and trust are also important, but these are matters for another discussion).

With all this data, most of the focus of the last several decades has been on ways it can be managed and visualized. Many powerful solutions exist today for collecting, organizing, querying, and presenting data. Most common is the visualization of aggregated data in a “Dashboard” using attractive graphical charts and similar elements that are both aesthetic and informative. The more advanced solutions allow users to “drill” into the data, extracting further details as they do, to “slice and dice”—that is, interactively re-organize and filter the data to highlight particular aspects—or proactively calculate “warning signals”, when trends in the data indicate that a key performance indicator is either out of range, or heading that way. Collectively, these capabilities are referred to as “Business Intelligence” solutions and can provide tremendous value to the operational effectiveness of the organizations that use them.

Management decisions: Looking forward in time and seizing opportunities

But this is where the distinction between the needs of the casual driver (and the status-quo business operator), and the needs of the racing team (and the forward-looking business leader) begin to diverge. Unlike the casual driver who gets everything they need to know from the dashboard, the racing team needs additional information that the dashboard can’t give them. For example, how will our car perform at the track we are racing on next week?  How can we configure the car to maximize its strengths, and minimize its weaknesses on that track?  What should we do if it rains?  When should we take our pit stops to change tires?  How do we change our plans depending on the decisions our competitors make?  Questions such as these are not operational in nature, but rather are about the future. The current Business Intelligence paradigm provides little to support decision making in this context.

This helps explains the chasm that all too frequently separates the needs of decision makers and the products created by the data experts who support them. Management often complains that they are not being provided with useful information, while analysts feel their work is being ignored. In the end, despite the mountains of data, thoughtful analysis, and colorful presentations, important decisions with major consequences are often made by “gut feel” and the promise of better outcomes from evidence-based decisions goes unfulfilled.

This is due neither to deficiencies in the product the analysts deliver, nor is it the fault of the decision makers who use that product. The problem is that the paradigm within which the work is being done is optimized to support operational goals, not future planning. Racing teams understand this and use very different tools when planning the strategy for a race as compared to making sure the car operates properly during the race. What can we learn from them, and how does their experience translate to the broader business context?

Ferrari pit wall with engineers tracking automobile telemetry, 2006

Tools for using Data to plan the future: models and simulation.

On race day, the “pit wall” team huddles over dashboards that could be produced by any of today’s top Business Intelligence suites. But before race day, back at headquarters, a very different set of tools was used to help make the strategic decisions that will be implemented while the cars are on the track. Two of the most important of these tools are:

  1. Models, and
  2. Simulations.

Between them, models and simulations cross the chasm that has separated the work of analysts from the information needs of decision makers

Put most simply, decision makers generally need to know “If a given course of action is taken, what will the outcomes be?”. This lies beyond the reach of Business Intelligence, but is at the core of the new discipline of Decision Intelligence.

See the source image
F1 simulator. Source: horizon inspirasi

Elite racing teams apply vast amounts of effort to develop and refine models of almost every aspect of what they do. The car is modeled, the driver’s inputs and responses are modeled, the track, the strategy of the competition, and much more. These models are then run to simulate how various decisions the team might make are likely to play out, and the outcomes of these simulations analyzed to determine what decisions should be made under the various circumstances that may occur on race day. When it comes time to make those decisions in real time, encountering similar situations during simulations arms the management team with foreknowledge of what outcomes are to be expected for each course of action.

In this aspect, the needs of the racing strategists are no different from the needs of a business leader who has to make critical decisions in complex, rapidly changing environments that contain external agents ready to pounce on any mistakes (such as those of competitors). Unlike top-shelf racing teams, however, most business leaders do not have access to models and simulators, nor to the enormous budgets and talent required to create bespoke versions of these. Fortunately, the increasing prevalence of Decision Intelligence (DI) makes modeling and simulation-based decision support tools much more available and readily affordable. In DI:

  • The role of Artificial intelligence is to automate the creation of models from data, where high-level, specialized statistical analysis talent was required in the past.
  • Modeling techniques such as Causal Decision Diagrams (CDDs), along with emerging tools that support them, systematically integrate quantitative models and human subject matter knowledge.
  • Similarly, business-level simulation software that runs these models through the various scenarios that represent different decisions are also becoming available and, by exploiting elastic computing in the cloud, are much more scalable and cost-effective than in previous eras.

Traditional Business Intelligence dashboards can be created to analyze, visually display, and allow users to interact with the outputs of simulations. Therefore, no new skills are required to consume the output of Decision Intelligence solutions. The main difference between these and traditional dashboards is that, while Business Intelligence concentrates on providing a view of the present and the past, Decision Intelligence gives you — like that Formula 1 team — data from the future.

Mark Zangari
Co-Founder and CEO at | Website

In addition to his duties as CEO, Zangari leads Quantellia LLC's Scalable Solutions division, where he is responsible for financial technology, telecom, and Covid-19 solutions. Zangari is also the architect of the company's World Modeler solution suite.  Before joining Quantellia, Zangari spent 15 years as CTO for a spatial GIS company, where he specialized in providing solutions to utilities and telecoms.

A physicist by training, Zangari's papers on cosmology are still referenced today.

You may also like...