No one likes a stalemate. You know, that point in a game where no one can win, all the moves are exhausted, and with a little better game plan, we might have produced a winning outcome. Lots of effort, nothing much to show for it.
Business Intelligence initiatives wind up, more often than not, in a stalemate, or worse. Maybe that’s why BI initiatives have remained in the top 5 CIO to do’s year after year for decades. We hear from dozens of companies each year who are embarking on their latest “new” BI initiative, but with the same old game plan, they’re not likely to see winning results.
Much has been written about this recurring outcome. Most of the narrative has focused on technical issues like data quality, master data management, data warehousing best practices et al, as primary failings for BI initiatives. Not that these are unimportant, but in today’s state of data and systems, these are the more obvious issues, and by and large, most firms have solved these, but still have not solved Business Intelligence. The good news is, these technical issues are the harder ones to solve; the bad news is, we keep missing the easier ones. So what are these, and how can we put forth a winning game plan? Let’s take a look.
IT often points to lack of C level sponsorship for their latest BI stalemate. Resources like Aberdeen Group or Forrester make a compelling C Level case showing higher performance of “leaders” vs “laggards” based on their BI strategy. These studies show earnings (EBITDA) are 10-15% higher, or cash flow 7-10% better for companies who are BI leaders. Executives would kill for just a few points, so seeing what they could gain by being a BI leader in their market segment should earn their sponsorship. Executives are also intently more aware of the exponential growth of Compliance, the penalties for non-compliance, and their connection to good BI. Going forward, there should be no excuse for not having executive buy-in, because no company can afford not to turn its data into profit.
Yes, this may be a more politically correct way to address the nerd factor. IT approaches problems in rows and columns, very left brain. End-users want to see a solution in a “day in the life” context, very right brain. Trouble is, there is no middle brain, so we’re at another stalemate.
As a consulting firm, we are often asked to umpire this stalemate because IT has asked the business, “What do you need?” to which the business answers, “What can you show me?” The left brain response to this is, spoken or not, “they (the business) don’t know what they want.” In reality, the business does know, but the left brain can’t answer the “show me something” question very well if “something” hasn’t been defined.
Today, there are a number of vendors who will lend you their software to allow you to be a little more right brain toward your customer, the business – Tableau, QLIK or Microsoft’s PowerBI, to name a few. In a matter of hours, you can pull workbooks or data sets into demo-ware and very quickly show some pretty fancy dashboards, KPIs, ad hoc queries or reports. And in doing so, you just went right brain. The benefit of showing something helps us learn what the customer doesn’t want, which begins to define what they do want. Eventually the stalemate is broken.
Here’s a very common scenario. Your C ranks do understand the value of their data. And they have a pretty ripe appetite for it as well, with a hefty diet of operational dashboards, daily flash reports and lengthy backlog of ad hoc requests. To deliver this, you have a team of over-worked developers with tribal knowledge of where the data is and how to get it. So you calculate the 100s or 1000s of work hours each week for these processes, and tell the CFO you can save a boatload of money. You present this and wonder why the response is “No!”
This may be false ROI. Yes, on paper you’re going to save or recapture these hours to productivity. But to the CFO, this is really false ROI until those savings show on the bottom line (more often than not, that is how s/he is getting paid).
If your ROI doesn’t center around one or more of the “3 C’s” – Cash, Competitive Advantage and Compliance – then it’s probably false ROI. Simply put, Cash means you can show a “bottom line” contribution from better BI. Competitive Advantage means you can show a “top line” contribution. And Compliance means you can demonstrate an “under line” contribution – that is, reducing the lines under an Executive’s eyes because s/he is going to be able to sleep at night. Don’t laugh, this may be the most important of the C’s, and often most missed. Show real ROI in one or more of these areas.
Many have assumed an enterprise data warehouse (EDW) is required to deliver BI. The “build it and they’ll come” mantra has cost a lot of companies millions, with little to no return.
Here’s a better approach. Every EDW starts with an operational data store (ODS). This is where disparate data sources from heterogeneous sources is staged, homogenized and layered with business logic like master data, security trimming and reporting hierarchies. A true EDW might answer questions like “how do customers of a certain demographic behave on rainy days in March?” The ODS is where we ask more immediate questions like “how are we performing month to date” and “what can we do to change the outcome.” An EDW represents years and millions of dollars of investment. The ODS is an iterative alternative measured in months and tens of thousands of dollars. In time, an ODS can evolve into and support an EDW, and it’s a much more cost and ROI effective place to start.
We are often asked to participate in RFPs for a BI Solution. Maybe you know the drill: Let’s ask the top 10 vendors to respond to an RFP that vaguely represents a requirements definition, score them, ask the top 2 or 3 to present, and finally select one. Most of these are orchestrated by IT with insufficient input from the business community, and result in shelf-ware, aka, we bought it but can’t seem to get the users to adopt it. The opposite, but equally flawed approach, is to let the business buy its own flavor of BI tool, and expect IT to support it. This is like building a new house, letting the kids design their own rooms and telling mom and dad they are responsible for managing the household.
This is really the easiest of the stalemates to avoid. Think tools last. Before you demo shiny objects in front of end-users, be sure you have a good blueprint with business requirements and success factors signed-off by the business opinion leaders and executive sponsors. Tools should be the last thing you think about for successful BI.
To sum up our strategy for breaking the BI Stalemate: