One of the key challenges in business intelligence (BI) is balancing corporate and departmental interests. This is a thread that weaves through most discussions of BI governance, data governance, and IT governance.
The tension between corporate standards and departmental autonomy goes well beyond the domain of BI and even information technology. In fact, it's endemic to the human condition, a driver of human history. In the United States, we fought a civil war over the scope of states' powers within a federal government. Closer to home, most of you have probably felt the heavy hand of a new corporate parent after an acquisition and have had to adjust your expectations and work habits accordingly.
Wayne, the Renegade
Ironically, I write articles like these to help corporate BI teams rein in renegade BI users, yet by nature, I'm one of those renegades. I started at TDWI when it was a private, almost family-run operation. But by the time I left, I had lived through two corporate parents, one benign and ineffective, the other bold and bossy. Given where I started with TDWI, I must admit that I bristled at taking instructions from corporate headquarters--almost on principle--especially when those standardized, sanitized instructions made little sense for our domain or cost us time, money, or customer goodwill. My mantra was "think local and resist global." I wanted complete control over my own domain. In short, I was a BI manager's worst nightmare.
Pondering my role as a potential BI "ne'er do well" got me thinking: how would I, as a corporate BI chieftain, deal with me, the proverbial BI bad boy? How would I get Wayne, the rebel, to willingly adopt corporate standards and tools and stop rushing into projects with short-term, duplicate solutions? How would I get Wayne to wait a little longer for applications that are more aligned, scalable, and durable than what he can build himself to serve his own parochial interests? How would I get him to trust the corporate BI team and rely on us instead of fight us?
Viewed through this lens, the tactics to rein in a BI renegade became clear to me. And I validated them through gut feel: if I felt they might work on me, then perhaps they might work for you in the BI trenches.
Some are bottom-up tactics that you, as a BI manager, can employ directly, such as appealing to my ego and self interest or just plain impressing the hell out of me with what you can deliver. Or you can rely on top-down tactics in which you need to recruit the CEO to buy goodwill through monetary incentives or force compliance with threat of dismissal. I've discussed many of these tactics in prior blogs, but here is a quick synthesis:
Bottom Up Strategies - What the BI Manager Can Do-
Impress me. Show me (Wayne the Renegade) that you can deliver what I want on time and within budget. Trust is hard to build, easy to destroy. Of course, since I never really know what I want until I see it, and what I need changes monthly based on ever-changing business conditions, good luck! Ideally, you know the business as well or better than I do because of your long tenure at the company and can anticipate what I want before I ask. Or you have skilled programmers who can build solutions incredibly fast without a lot of committee meetings or the need to rigidly adhere to architectural standards. In some cases, you recognize that it's more important to go fast than get it right and are willing to reset your architecture once a year to get things back in sync. Bottom line, if I think you can move faster than I can with my own resources, then you are my man (or woman)!
Give me a title. Give me a nice, ego-fattening title to entice me to run the BI working committee. Tell me you want me to help shape the BI program to address my needs. Ask me to recruit my buddies in other departments to serve on the committee with me. Repent of your team's perennial shortcomings ("too slow," "too costly," and "too structured.") Acknowledge that you want to do better, but emphasize that you can't do it without my help. I will likely be flattered by your blandishments. And once I chair the committee, I'll be unwittingly sucked into a position of protecting global interests ("my buddies") not just my own.
Give me a job. If I'm too busy to participate on a working committee or too skeptical of corporate BI, then up the ante. Recruit me to work on your team or perhaps even run it. Yes, make me your boss and give me control of your budget and people. Successful BI teams recruit people from the business, who know the business, and can talk the language of business. And if they are renegade spreadmart users, they know enough technology to pick up the rest. These people build bridges of trust between the business and IT, paying considerable dividends.
Sell me. If you can't recruit me to join the steering committee or your team, sell me on the benefits of rebuilding my application. Find out how I'd like to improve my spreadmart--e.g., add more data sources, make it real time, add a better visual interface--and the create a proposal that does that. And if I like it the way it is, show me a comparable application at a competitor that shows me the untapped potential of my application. Recruit vendors to deliver a slick demo and maybe donate some free software. Calculate how much time and money you'll save me once the application is built and trumpet how good it will make me look in the eyes of my boss.
Co-opt me. If I really don't trust you or your team, then give me access to your platform. Let me use your ETL tools and create my own data marts on your data warehouse platform. Let me add my own data to the data marts and use my own BI tools to access the data. Provide basic guidelines for using the ETL tools and platform, such as naming conventions, scheduling, and error management, and insist that I use corporate definitions for all shared data elements.
Bribe me. If all else fails and I'm a key player with sizable political clout whose support could benefit the BI program, then bribe me. Give me a freebie under the table. Take me out to lunch and learn about my mission-critical spreadmart which executives rely upon to make key decisions. Then make a secret pact in which you build out my application free of charge--or at least an initial module as a proof point--and I agree to fund the development and maintenance if what you develop meets my requirements and timeframe. Tell me you will assign your best developer to my project--someone I like and trust and who has the collective skill sets to build the entire application by himself. If you can deliver such a quick win, you'll make a powerful ally for life.
Top Down Strategies - Enacted by the CEO-
Force me. The quickest way to get buy-in to an enterprise BI program is to have the CEO require everyone in the organization to adhere to corporate standards in the use of data and reports. An edict is needed to set the stage, but won't be sufficient to stamp out renegade activities. To accomplish that, the CEO needs to "reassign" people who spend the majority of their time creating spreadmarts. However, before he restructures the workforce, he should make sure there is a viable enterprise BI alternative to pick up the slack. Unfortunately, over time, spreadmarts will creep back in to the organization unless the corporate BI team can get ahead of requirements and data sources.
Incent me. People go where the money is. So the CEO should base the bonus of a few senior managers in part on the growth of the company. The CEO should require them to serve on a cross-functional steering committee that oversees the data warehouse and BI functions. By paying them to think global, they will more effectively align their groups with global mandates to support an enterprise data warehouse and BI service. They will more effectively restrict or curtail unauthorized usage in their own departments, understanding the value that will accrue to the organization (and themselves) by adhering to a single version of enterprise data.
To rein in renegade BI users like me, you need a combination of tactics above. I can vouch that they will have some effect, if not work outright. Bottom-up strategies are things you can work on now, while top-down tactics require the cooperation of an enlightened senior executive. Nothing is a surefire bet, but if you use common sense and business savvy, you'll find common ground with even the most recalcitrant user, like me.
Posted March 25, 2011 1:49 PM
Permalink | No Comments |