Do We Need Data Scientists, or Do We Need Decision Scientists?

If you're thinking something along these lines: "We have all this data, and we need data scientists to tell us what we can do with it," you're at the wrong end of the trail - and you're walking backwards. Blindfolded.

Do we need data scientists, or do we need "Decision Scientists"?

If you're thinking that you need data scientists to tell you what you can do with your data, you don't need data scientists. Step back from the hype; step away from the storage salesperson trying to convince you to "store everything and we'll figure out what to use it for later." Step back from the "Big Data Initiative" you are in the midst of planning or executing.

Here's why:
If you don't know what to do with your data, spending time and money analyzing it is not going to get you where you want to go - because you don't know where you want to go.
Enter the Decision Scientist, and a decision-making methodology that makes everything else fall into place in a nice, orderly, relevant way:

Step One: Start with the objective in focus. What is the end game? What is the outcome, or what are the outcomes you wish to achieve? What decisions can you make that will enable those outcomes, and how will those decisions impact the future?

Step Two: Now determine what information is necessary to analyze the decisions and their impact on the outcomes. Is there information to be gleaned from Big Data? Great. Likely you also have information embedded in various systems: in a data warehouse, in spreadsheets, in predictive analytics, and even information that isn't documented anywhere, living in the heads of the people around the table. This information can all be leveraged, and the work already being done in data management is utilized - even if the data isn't perfect or even if it is missing.

Step Three: What external factors will also impact our outcomes? What assumptions do we need to make about those externals? Are we assuming, for example, that hourly rates for resources will not deviate more than 2%? Are we making decisions based on an assumption that our competitors' prices will remain constant? All of these factors must be weighed in, as there could be a profound impact on your business based on factors that you can't control.

Step Four: It's time to create a decision model. Now, this is really hard to do with a spreadsheet, but it's possible. World Modeler™ by Quantellia is one of the few (if not the only) software platforms specifically designed for this purpose that is visual, easy to use, and in "normal" language. In creating the decision model, the dependencies between decisions, externals, and outcomes are calculated, and the impact of decisions made today on outcomes are clearly shown - in a 3D visualization.

Step Five: Experiment with the decision model to see how changes to decisions will ripple through time to impact outcomes. Are the outcomes realistic? Can you act on the decisions you've determined are relevant? Watch the model change, and you'll start realizing how decisions today will impact the future.

Step Six: Act on the decisions, monitor the externals, and manage the decisions based on actual information as it accumulates - knowing that you are doing so with a clear direction toward the outcomes that are critical for your organization to achieve.

Maybe you do need data scientists. Starting with the outcomes defined, and developing a decision process will allow you to position those expensive resources in the places where they will have the most impact on the organization's goals.