It’s 2017, and you’re tired of hearing about how other organizations are getting value from their data; your own spreadsheets are hideously complicated, and take hours to prepare! What are you missing? Is there a tool on the market that can fix your data woes? Unless your organization is a lemonade stand, you’re probably sitting on a mountain of data (editor’s note: No offense intended to the modern amateur lemonade services industry). This data likely already drives some decision making, or at least some “hindsight” analysis: how were sales last quarter? What is my most popular product? But is that data “activated”?
Oracle Marketing Cloud defines Data Activation as “…the concept of unlocking value in data through development of insights and turning those insights into action.” Typically, data activation is discussed in the context of marketing data (particularly around customer insights). Simply put, data activation is taking the primordial soup that is your hot mess of data and getting value from it. Over the next couple of blogs, I’m going to talk about a hypothetical organization and how it might evolve into a truly data driven operation.
To do that, pretend with me that we’re part of the lemonade stand titan Limon & Lemon Citrus Beverage Services Incorporated. We operate 117 lemonade stands across 14 states, with daily lemonade sales in the tens of thousands of cup-fulls. We manufacture our own lemonade, offer a customer rewards program, we have an ongoing advertising campaign (you know the one: “When life hands you lemons, head to Limon & Lemon”). What might our data infrastructure look like? We certainly have multiple disparate systems:
- Order/Point-of-Sale system
- CRM system tied to our customer rewards program
- Product database
- a formulation database that tracks our secret sauce recipes
- an HR system
- …and many others.
Currently, we have some executive dashboards to view sales by lemonade stand, most popular products, etc. This data is useful, and certainly drives decisions… but it’s also quite basic. This information is truly no better than we could’ve calculated by hand years ago. Further, it’s no longer scalable. Our operations are large enough now that integrating that data by hand each month takes an inordinate amount of time. We are plagued by data quality and integration issues, and they aren’t getting resolved.
Finally, our new leadership has a mandate: transform us into a data driven organization. We believe that if we are to thrive in the highly competitive lemonade stand industry, we’ve got to innovate. That innovation should be driven by intelligent, data-backed decisions. We want to subscribe to external datasets, and utilize machine learning to look for trends and insights of which we were previously unaware.
With this mandate, our data architect gets to work on envisioning the architecture. Our operational systems with structured data will be integrated into a cloud-based Enterprise Data Warehouse (EDW). ETL processes should move the data from the source systems to the EDW and apply some data quality rules. The EDW will then feed subject-oriented data marts for traditional reporting. Later, we’ll invest in a Data Lake with unstructured data and external data sources we subscribe to. We will combine that with data from our EDW to create an analytics capability for the most hardcore data geeks on our team.
This high-level architectural diagram is just the beginning of a major effort, but it’s a place to start. In future blogs we’ll take a deeper look into how this will play out as well as the change management and data governance capabilities we’ll need to build in our organization to get the most out of our data.