I just read an article this morning talking about Big Data and a “Data First” approach to BI. I generally try not to deal in absolutes, but this one gets me riled up.
I’ve seen many organizations take a “Data First” approach to BI and in most cases, it’s a huge waste of resources. This type of approach is often driven internally by IT. It’s more comfortable to start extracting and consolidating data than it is to actually go out and talk to business users. If we are honest about how most of us in IT operate, we'd rather just build something than go out and talk to someone in sales or marketing.
Business Intelligence initiatives should be “Business First”. Before embarking on any BI projects, we need to understand two dynamics:
- How will this data drive the overall goals of the business?
- What’s the effort to get that data?
If those two dynamics can't be expressed, we need to take a harder look at why we'd move forward. It's wasted effort to pull a large amount of data into a Data Lake, Operational Data Store or Data Warehouse if it doesn’t move the organization toward its goals. It is also a potential waste of resources to go after data the business wants, but is costly to retrieve. Just like we'd do with any other organizational asset, we need to understand the costs and benefits of analyzing data.
Years of experience has lead us to develop a methodology for working with the business to understand what their key drivers are and how much effort it will take to measure them. We take an iterative approach to implementing key initiatives based on a balance of business priority and overall cost. This approach has been effective for many organizations and helps them avoid the high cost pitfalls of a “Data First” strategy.