Top-Down Leadership for Data: Seven Ways to Get Started

Leaders can initiate successful data strategies by focusing on data quality, building organizational capabilities, and putting data to work in new ways.

Reading Time: 7 min 

Topics

The New Leadership Mindset for Data & Analytics

This MIT SMR Executive Guide offers new insights and strategies for how leaders can help accelerate their companies’ data efforts, from identifying the type of talent they need to shaping a company vision that supports a data-driven culture.

Brought to you by

AWS
See All Articles in This Series

For many senior executives, data presents a conundrum. Naturally, they want their data programs to succeed. Indeed, they’d like to help in some way, and even provide leadership, but beyond providing funding, they’re not sure how. Business leaders’ lack of technical knowledge exacerbates their uncertainty and makes them feel like outsiders to data science teams.

But data programs cry out for business leadership, and there are many ways senior executives can play a bigger role, even if they don’t fully understand data. This article explores seven ways leaders can accelerate their companies’ data efforts, derive near-term benefits, and gain a better understanding of the roles data can play in advancing their business objectives. Think of these as opportunities. Not all of them will suit your needs, interests, or style, so just pick one or two and get started.

Focus on Quality

We begin with three opportunities related to quality. Counterintuitively perhaps, improving data quality saves money. Further, all data strategies (and more and more business strategies) depend on high-quality data.

1. Break the quality logjam. Many leaders readily admit that they don’t trust their company’s data, at least not enough to use it when it really matters. It’s no wonder, given how often they encounter bad data: Two numbers don’t jibe, some report looks too good to be true, and the finance team complains every quarter about the overtime needed to produce a routine report. Still, most companies manage to convince themselves that their data is “pretty good.” And an uneasy stasis results — you don’t have the trusted data you need but are frozen from doing anything significant about it.

To break the logjam, undertake a review to determine how good — or bad — your data is. Tell people there will be no rewards or punishments for good or bad results — it is just important to establish a baseline. Instruct your teams to make a simple data quality measurement, using what I call the Friday Afternoon Measurement. It entails laying out a sample of the most important data your teams use every day, marking the obvious errors, and then counting them up. This measurement returns a score between 0 and 100 — the higher the better.

Although calculating this measurement takes just a few hours, it may take longer for the team to work through the implications. On the one hand, you may find out that your data can be trusted (scoring above 96), and you can then move on to other improvements. However, few teams score this well, and any score below 85 signals real problems. Solving these problems will require more work, but breaking the logjam is half the battle.

Send a signal that your company must change its approach to data quality, from passively dealing with errors to making them go away — for good.

2. Get to the bottom of something. When you raise questions about numbers that just don’t look right, your staff goes off in a frenzy to find the mistakes and correct them. Tragedy averted, or at least delayed.

This little drama plays out dozens, hundreds, or thousands of times a day (depending on the size of your company) — sales cleans up data it receives from marketing, operations cleans up data from sales, and finance cleans up data from everyone. Note the vicious cycle and the incredible drain on company resources. You can lead by example to break this cycle. Next time an issue comes up, lead an improvement project to get to the bottom of it — finding, then eliminating, the root causes.

In so doing, you will send a signal that your company must change its approach to data quality, from passively dealing with errors to making them go away — for good. Those who report to you (and maybe some peers) will follow suit, making this perhaps the single most transformative action you can take.

3. Get everyone on the same page. People in all but (maybe) the smallest companies complain, “Our systems don’t talk.” You’ve probably experienced this yourself, when people cite it as the reason why different answers might exist throughout the company to basic questions like “How many customers do we have?” Unfortunately, the problem of systems not talking camouflages a more fundamental issue — namely, people not speaking the same language. This has real consequences: It is more difficult to work across departments, complexity grows, and your technology department wastes time better spent elsewhere.

To find out if you have this problem, pick three or four important terms — for a financial services company, security, buyer, and client are good examples. Ask everyone (yourself included) to write down their definitions of these terms and bring them to your next staff meeting. Then read them aloud and see how closely the definitions align. Remember, there is no right or wrong. If everyone agrees, great! If not, you may have a serious problem.

This is the headiest of the seven opportunities described here. But once you know where to look, you will see how common language can resolve all sorts of business issues.

Put Your Data to Work

4. Bring data science to your strategic decisions. A common misconception about data science — with terms such as big data, artificial intelligence, and machine learning — is that it’s reserved for areas where data is plentiful. In reality, when done well, data science helps ensure that a business problem is clearly articulated, reveals hidden biases in your thinking, provides a fresh perspective on whether the data you have is good enough to address that problem, seeks new sources of data to close gaps, analyzes all the data in powerful ways, and reduces your uncertainty. So, pick an issue, invite a first-rate data scientist to join the fun, treat the person as a full member of your team, and listen to what he or she has to say.

5. Determine how your data sets you apart. Your next opportunity involves competitive differentiation. As you know, companies don’t compete based on the ways in which they are similar to others but on the ways in which they differ from the crowd. It is all too easy to view your data as a byproduct of your real work and not see it as a source of value in its own right. It is worth the time to examine where strategic opportunities may be overlooked.

Start by charging a multidisciplinary team with answering the question, “Do we have data that no one else does?” Chances are, you do — after all, your data is uniquely your own. If so, this data qualifies as proprietary and may be a source of competitive advantage. Clearly enough, the next step involves figuring out how to exploit it. Of the opportunities described here, proprietary data probably signals your best chance at creating near-term revenue.

Build Needed Organizational Capabilities

6. Separate the management of data and technology. Slowly, and usually with great difficulty, data is invading every nook and cranny of every industry, company, and department, creating both opportunity and risk. Yet today’s organizations are unfit for data — they lack talent, silos get in the way of data sharing, data and business strategies are poorly connected, and it is unclear who is responsible for what in the data space. Over the long haul, these issues will require a lot of your time.

Far and away the best thing senior leaders can do for their data programs is to clarify managerial accountabilities. Fortunately, step one is relatively straightforward: Move lead responsibility for data out of your IT department. Data and technology are different kinds of assets, and comingling their management has slowed progress on both. Work department by department, from HR, to finance, to operations, to administration, to find a better home for your data program.

7. Demonstrate your commitment to data on your board. It’s important to mirror the data improvements you make internally on your board of directors. If you haven’t done so already, make it a priority to appoint a director to your board who understands data. You need someone who will push you to go further and faster than you might otherwise be inclined to, to serve as a sounding board as you evaluate options, and to help you anticipate and deal with resistance. The ideal candidate has both an expansive vision and experience with hard-fought battles in advancing a data agenda. The sooner you find this person the better, so start looking right away. Interview as many people as you can, and don’t decide until you find someone you really trust.

Data may well represent your best chance to grow your business and distance yourself from your competitors, but getting even a fraction of the value data has to offer is tough. It’s no surprise that unlocking this potential requires strong stewardship from senior leadership. Executives can use the seven ideas in this article as entry points for getting started.

Topics

The New Leadership Mindset for Data & Analytics

This MIT SMR Executive Guide offers new insights and strategies for how leaders can help accelerate their companies’ data efforts, from identifying the type of talent they need to shaping a company vision that supports a data-driven culture.

Brought to you by

AWS
See All Articles in This Series

More Like This

Add a comment

You must to post a comment.

First time here? Sign up for a free account: Comment on articles and get access to many more articles.

Comment (1)
Andrew Kingery
I could not agree more about data quality improvements being a cost saver -- and getting the initiative visibility at the board. Great points thanks for sharing Thomas.