Driving with Data: Do You Truly Want to Be a Data-Driven Business?

Getting to the stage where your business processes and decision-making are “data-driven” sounds appealing. After connecting all your data sources to the necessary artificial intelligence and machine learning mechanisms, you can just flip the switch: Decisions are made for you, and changes to business processes to meet customer needs can run automatically.

But is that a stage you actually want your business to advance to? As an analogy, are you ready to sit in the back seat of your car and trust it will self-drive to your destination? The technology is available today that enables cars to transport people safely and take the quickest route possible on their own. But are you ready to let go of control?

Perhaps Data-Informed Is More to Your Liking

Instead of aspiring to become data-driven, many businesses prefer to reach the stage where they are data-informed. In other words, leverage AI and machine learning to predict what will happen and prescribe the actions to take in light of those predictions—but still allow humans to make the final decision on what course to follow. Humans would also continue to drive the ship.

Going back to the car analogy, being data-informed is like utilizing Waze or Google Maps to tell you which route to travel, or having your car alert you when the vehicle ahead of you suddenly brakes. You still retain the right to take an alternate route, and only you can push the gas, step on the brakes, and turn the wheel.

Whichever data stage your business aspires to, it’s important to take a phased approach as to how you process your data. It’s impractical and often a recipe for failure if you just jump to trying to be data-informed or data-driven—without first accomplishing the key steps that come first.

The Five-Step Data Analysis Maturity Lifecycle

To build a data-informed or data-driven business, here’s a rundown of the step-by-step analytics process. Your business may need to evolve following these five methodical phases:

Descriptive—react tactically to what has happened in the past
Diagnostic—determine why things happened in the past
Predictive—identify what will happen in the future
Prescriptive—assess strategically what changes need to occur (a data-informed state)
Autonomous—the system takes action without human intervention (a data-driven state)

The order of these data analytics phases is critical. It’s what we call the “Crawl…Walk…Run…Sprint” data maturity lifecycle—where you progress through each phase before you attempt the next. The successes of each step also create a foundation for building out the subsequent phases so that you can learn how your business needs to process data and invest incrementally to spread out your system costs over time.

Where Do Your Systems Rank Today?

The first step in completing the data maturity lifecycle is to evaluate where you’re currently at within the five-step process. Your level of data analytics maturity will likely be different across business units and the database systems within each business unit.

To get a sense of where your systems stand, consider the attributes those systems currently exhibit:

Descriptive analytics systems tell you what has happened. This is often handled via transactional lists that focus primarily on cost monitoring but can also include performance evaluation and operational analysis. Advanced descriptive systems generate data in near real time; those that lag may produce reports once per week or once per month, and some reports require manual intervention.
Diagnostic analytics help you determine what is happening and why the business is performing at its current level, although the reporting may be uneven across the different operational areas of the business. Most reports are generated automatically and soon after the time period being analyzed.
Predictive analytics provide insight into what is likely to happen in the coming month, quarter and year. Management also has access to departmental scorecards with the ability to drill down into each data point to understand what happened and what it is likely to happen in the future. Charts show trend data and these reports are produced and delivered automatically with no manual intervention and typically cover the operations of an entire business unit or enterprise.
Prescriptive analytics shape your perceptions of what has taken place in the past and provide data to impact your decisions as to what actions take in the future. You are also integrating your internal data sources with external industry and customer data so that you have context around your KPIs. All analysis takes place in real time.
Autonomous state is reached when senior management feels it can trust the system to make decisions on strategic planning for the future and kick-off automated processes for making changes to how the business is run on a day-to-day basis. For those businesses that reach the data-driven autonomous state, humans will of course always monitor the decisions and have the power to override any changes.

A Matter of Trust

The level of data analytics maturity that is best for a particular database system, business unit, or an enterprise will vary according to what works well for each entity and the available budget to invest in the required technologies. The long-term goal of most businesses will be to reach the prescriptive data-informed phase.

Whether to go all the way to the data-driven autonomous state will likely come down to how comfortable the senior management team is with the available technology and how predictable processes and outcomes are for the specific function. Humans are often swayed by emotions, and gut feelings are wrong just as often as they are right.

Computers, however, remove all the emotion and base their prescriptive actions only on the logic built-in by the business as well as the historical and future trends. It all comes down to who, or what, you trust most to run your business profitably. Your management team will need to carefully consider its course of action.