Here’s the best question I’ve ever been asked by a client:
“What’s not in this report that you think we should know?”
Now, I wish I could tell you that the answer was “nothing”, because of course if there was something important to say, I would have put it in the report. After all, that’s what they paid for. Instead, I reflected on the question for a few seconds. It was a small group and we had become friendly over the course of the engagement, so I decided to tell it to them straight:
“Ok. First, as long as Carl [not his real name] is in a position to influence data strategy, your enterprise approach will not succeed. I think you already know this – his personality makes it impossible to work collaboratively across groups. Second, you should put Makayla [not her real name] in charge. The work that she did to facilitate and implement a data strategy in the transportation unit may be the best I’ve ever seen. I know she’s very interested in taking what she’s learned to the enterprise level. It’s difficult for me to understand why she wasn’t selected for that role to begin with.”
I guess I over-corrected a little. Pointing out individuals by name in that context really wasn’t a smart move. The good news is that, without getting into details, it ended up working out well for everyone involved. Career crisis averted.
Over the years, I’ve tried hard to be as forthright with my advice as possible, while being much more diplomatic and sensitive in the way I communicate. One way I’ve carefully side-stepped naming names is instead to identify characteristics of people who are well suited to drive enterprise data strategy. While these characteristics are useful beyond data strategy, they’re especially important for developing a vision and plan for enterprise data – and implementing that plan – because of the complexity, the cross-functional nature of the work, and the impact data strategy has on virtually every major business initiative within a large organization.
Characteristic #1 – The ability and willingness to establish and communicate the right objectives, with seriousness
The details will vary by organization, but there are essentially three objectives for an enterprise data strategy:
- Provide the data needed to support funded business initiatives across the organization
- Ensure the condition and management of data effectively supports business operations
- Establish and continuously improve trustworthy, shared data resources
But identifying and articulating the objectives are not enough. They must be communicated with seriousness. You can communicate your goals with a formal a mission statement, principles, and objectives – or not. You can hang posters all around or not. You can develop beautiful, multi-colored, graphical presentations or go with black and white with a long list of excessively verbose bullet points. Any of these choices can work to communicate objectives. The thing that really matters is your seriousness.
Characteristic #2 – The ability and willingness to act with sincere intention and continuous learning
Once the objectives are established, you’ll encounter plenty of advice on how to meet them. But even more important, as the entire team internalizes the objectives, they’ll emulate your sincere intention and find ways to accomplish the goals. Data modelers will learn how to build extensibility into their data structures for the long term as they focus on the data needed for business use cases in the short term. Business analysts will parameterize reports for easy reuse and seek out ways to share and rationalize analytics. Data stewards will learn how to prioritize data quality issues to support application projects and ongoing business processes. As projects move forward, team members will use their creativity and innovation while being open and receptive to the right ideas because the objectives are set firmly in mind while acting and course-correcting every day.
Characteristic #3 – The ability and willingness to work proactively with strangers and adversaries
A successful enterprise data strategy requires cross-functional cooperation. If the data you’re delivering supports major business initiatives – sponsored elsewhere in the company – you’ll need to work closely with those areas on a regular basis. Let’s be honest, that can be intimidating. We all like to work with people we’re comfortable with. It makes our day much more pleasant. It’s no fun to risk presenting your ideas only to have someone you hardly know throw verbal obstacles in your way at every turn. But the leaders with the ability to proactively press forward, working with people who make them uncomfortable, will develop the skills needed to navigate these interpersonal challenges. And of course, the easiest way to work with adversaries is to not have any. Good leaders know how to make that happen – most of the time. If, instead, you avoid collaboration and deliver data as a general “foundation” for the enterprise, working within the relative isolation and comfort of your own team, you’ll probably end up with lengthy and costly projects that just don’t deliver the results you hoped for because the work won’t have the right urgency associated with close alignment to important business initiatives across the organization.
Characteristic #4 – The ability and willingness to worry about the right things and let go of the rest
Directing worry appropriately is just another way of saying that you should institutionalize effective risk management. When planning and building enterprise data resources, a well-known area of risk, for example, is data quality. I’ve seen many projects and programs derailed due to excessive worry about the wrong data quality issues. If you’ve established the right objectives (characteristic #1) then the actions you take (characteristic #2) should be focused on the specific data quality issues that align to the business initiatives and operational processes you’re trying to support – no more, no less. This dramatically reduces the scope of work while ensuring that every action is in support of the in-scope, prioritized business goals, not an unrealistic expectation of getting all the data perfect. There are plenty of other other anxieties that do you no good – excessive focus on maturity models, arguing about technology decisions too far in advance of the need, or redesigning and re-architecting highly complex solutions for no reason other than to “modernize”, which, if done perfectly, would give exactly the same results already in place, likely at higher cost and risk. But again, if the real objectives are clear, serious, and internalized, then the unnecessary stress associated with these tangents will easily fall away.
I’ve examined many large-scale data and analytics programs up close, and there are always improvements that can be made – in the organizational structures and processes, the technology choices, architecture and design decisions, the overall strategy, and so on. But, really, none of it matters if leaders of the program don’t have (or develop) the right characteristics. With these success factors in place, you have a very good chance of success, no matter how complex and daunting the mission.