I received pushback last week from my article about the need to embrace analytics. (See “From Reporting to Analytics: A Pathway to Greater BI Value”). One BI leader who runs an effective BI shop and whom I know well wrote:

Although our data warehouse and reporting are certainly more mature, we still have a long way to go…. [Meanwhile] our new big data/advanced analytics practice continues to show promise, but quite frankly, still hasn’t delivered any tangible results. But our [company’s] resources go to [the analytics practice], while our existing BI architecture doesn’t get the attention it deserves.

Those of you who follow my work know that I believe there are two worlds of business intelligence (BI). One is the top-down world of reports and dashboards populated with certified data from a data warehouse, and the other is the bottom-up world of ad hoc exploration and discovery. You also know from my recent posts that since 2010 the power and resources in organizations have shifted from top-down BI to the bottom-up world. (See “The BI Power Struggle: A Strategy for Success.)  

What I forgot to mention in my paean to analytics last week is that analytics is not the whole answer. In fact, analytics is not even half the answer. Although every organization should covet analytics because it can deliver critical insights, it is not the alpha and omega of a data-driven organization. In fact, analytics is really just a valuable adjunct to BI, not a replacement for it. Executives need to recognize that they need to fund both BI and analytics in equal portions, not favor one at the expense of the other.

Analytics Without BI?

Organizations can survive without analytics, but they can’t make it through a single day without BI. BI does the blocking and tackling in a data-driven organization. It cleans and integrates data; it delivers reports and dashboards that support the business on a day-to-day basis; it tracks the performance of core processes and shows executives and managers their progress against plans. When the data warehouse goes down or dashboards slow to a crawl, business people scream.

BI Complaints. Of course, business people also scream about their organization’s lousy data warehouse and BI solutions. We’ve all heard the complaints: the data warehouse is too slow, too hard to navigate, and doesn’t contain the right data or timely data. We’ve also heard complaints about the BI team: they don’t know our business; they take too long to build reports or data marts or add new data sets to the warehouse, and so on.

Some of these complaints are legitimate: there are many data warehouses that are poorly designed or in desperate need of a technological facelift. And many BI teams are too far removed from the business to do a good job building relevant solutions quickly. Unfortunately, these complaints intensify if executives fail to properly fund the BI program.

The Role of BI in Analytics

Unfortunately, many of these complaints are misguided. They usually come from business analysts and data scientists—the folks in the bottom-up world who need quick, ad hoc access to any and all data and whose requirements can’t easily be ascertained in advance. While these folks can benefit from a data warehouse and BI-style reports and dashboards, what they really need is a dedicated bottom-up analytical architecture comprised of various sandboxes and visual analysis and data mining tools.

Certainly, the BI team can and should play a role in supporting these bottom-up players.  The BI team can help create and maintain the analytic sandboxes; they can purchase and support analyst tools; they can provide training and support about how to navigate the data warehouse and uses the analytic tools to support root cause analysis of top-down issues; and they can create a BI/analytic center of excellence and analytical ecosystem that tie the top-down and bottom-up worlds together. This is all the more reason to fully fund the BI program. A healthy BI program can spawn and support a healthy analytics competency.

Some may ask whether an organization needs to implement or perfect a BI program before commencing with analytics. Although an active BI program with a robust, dimensionally modeled data warehouse can provide a strong foundation upon which to deliver analytics, it’s not required. A department head can hire a few analysts or data scientists and let them loose on the data—where ever it may lie—and derive some value. In fact, this is how the majority of organizations operate, unfortunately.

Creating Synergy. The optimal environment is where both top-down and bottom-up worlds work synergistically. Their respective heads communicate frequently, report to the same executive, and work together to create a unified BI/analytics organization and architectural ecosystem that supports both top-down and bottom-up requirements in an optimal way.

To create this data nirvana, business executives have to be enlightened about the dynamics driving a data-driven organization. They need to fund BI and analytics in equal portions, not favor one at the expense of the other. This will help avoid the predicament faced by the BI leader mentioned at the outset. In his words:

So while our marketing Leaders continue to experiment with predictive indicators from social media, our CFO simply wants to know “What is our customer renewal rate for each product”. Right now, we can’t reliably provide that to him [because all the investment goes to advanced analytics.]