What are some good "beginner level" data modeling/analytics approaches to kick start a data science/analytics team?

How does your organization optimize the value in your business data?

• One of the biggest challenges that I've seen as organizations strive to become data driven is the ability to identify who within the organization uses what data and the value it provides. There's a seemingly unending supply of data and sources, but determining what is important and to who can be difficult, especially when an organization is larger and on the low end of the data maturity spectrum. Two techniques that I've employed to assess the value of business data and build out a data strategy has been 1) conducting comprehensive surveys of all current data products (reports, dashboards, analytic output, etc.) and attempt to "map out" data usage patterns throughout the organization, or 2) start clean with a top down question what business problems should be solve and identify key measurements to drive the use of data to address these problems. The issues with the first approach is short value half-life of some data (and it's hard to know which), while also limiting innovation of data use to things that have already been done before. It is also true that this type of exhaustive product level analysis can misinterpret the intention of data products and lead to ill-informed new data requirements. The challenge with the second approach is the difficulty organizations have with identifying their key data measures and tying these directly to business decisions that impact organizations success. There is also the notion that this level of direct interaction with organizational leaders from the top down has a significant resource cost that can be difficult to justify or build interest in. For the average mid-sized organisation both approaches require high cost in terms of resource utilization, and unless there is an organizational mandate and desire to mature the use of data to drive the business, key business resources are hesitant to devote the time necessary provide a clear picture of their data use and value. What techniques have you employed in your organization to identify the business value of data and map that out to organization use to deliver a strong data strategy? I'd love to hear about the experience of other data focused leaders.

• Answer:

  To answer this question most efficiently I think we need to shift it, slightly.   “How do we optimize our business processes and help our employees work smarter, using our data?”   This shift does something very important to the way we approach the problem. It directs our attention towards the people and the processes of the business instead of starting head first in the data. This is important, because the idea of optimizing anything is fundamentally an issue of prioritization. Optimizing the value in business data simply means applying it to the highest priority problems, first.   Identification of the highest business priorities is the first trick. Here are three steps to go about it.     1.       Study the Business. Where is most of the revenue coming from? Where are the costs highest? Where is marketing reach greatest? Where is conversion rate lowest? This step is probably the least effective at truly identifying the highest priority needs, but doing some if this first will give you enough information and a foundation for conversation for the following two steps. You may also want to search for and read some case studies on how other companies are effectively using data to give you an even greater breadth of ideas.   2.       Talk to People. Get out of your cube and go say hello to some folks in various departments of the company. Marketing. Sales. HR. Executives. Whatever. It doesn’t have to be a long, formal conversation. Ask them about their day to day and what holds them back. Ask them what information they wish they had access to. You are likely to come away from this exercise with a gold mine of ideas for where you can apply your business data. And you might make a few new friends.   3.       Socialize the data. If your company has an internal social network, use it to socialize the data.  We use Kona. The truth is that most of the people in the company have no idea the amount of data available to them and how they could use it. Start posting ideas you have, interesting data sets/visualizations that are available, and some potential use cases. Ask for needs and feedback. Because each business is so unique, this socialization is critical to find the nuanced uses of data within yours.     The methods above are more grass-roots and organic, rather than comprehensive exercises. The reason the organic approach works so well is because businesses are large and very complex entities. You have hundreds of people, each doing different things and working in different ways, to help the company move forward. The idea that one person or team could possibly understand the information needs of everyone in the company in such a perfect way as to be able to optimally distribute data value efficiently just isn’t going to work.   But the next closest thing is to have a socialized priority identification method that an analytics team can respond to it in an agile manner. As you start to see needs come up during this process, prioritize and fulfill them. As more needs keep flowing in, continue the process.   It’s essentially agile development, but for analytics.