10 steps to success as a data-driven organization
Looking through job titles on LinkedIn, I couldn’t help but notice an unfortunate reality when it comes to data roles in the business world: it’s a mess. Companies are putting a lot of effort into collecting and analyzing data, yet there’s so much ambiguity regarding responsibilities between different departments, that it’s going to have an influence on business insights.
There are a number of reasons for this, from internal politics to a lack of knowledge how an organization should structure itself in terms of data functions.
Over the years, I’ve worked with dozens of companies. I’d like to suggest a few strategies to optimize structure and tackle issues which arise from duplication of roles and functions, and insufficient efforts by various departments, to help your enterprise get back on track and be more data-driven.
Here are 10 strategies to make any company a more successful data-driven decision-based organization.
1. It starts at the top. The CEO must appreciate that data is a basic element in the success of a company, and that the ability to quickly and accurately analyze data for business insights is of critical importance to maintain a competitive edge. Otherwise, there will be problems across every department. It’s as simple as that.
2. The CDO (Chief Data Officer) is the most critical and powerful role in the data organization. The CDO must be a strong C level position who should take part in all of the data-related responsibilities. That doesn’t mean that a specific department can’t analyze on its own; but it does mean that data management and analysis is strategic, and requires an effective leader who reports directly to the CEO.
3. If the customer is fundamental, then the data department must be recognized as a service-oriented department, beyond simply a technical role. The most successful companies I’ve worked with were the ones that saw data analysts as providing the services and mechanisms that enabled BI (Business Intelligence) tools, including data monitoring and analysis tools. “Although digital business thrives on data and its analysis, we still see that data and analytics only play a supportive role when it comes to business initiatives,” says Mike Rollings, research vice president at Gartner. This has to change.”
4. The data department also cannot function as a “dictator.” It’s a common mistake to believe that everything regarding data should stay inside one department. But, it doesn’t work. Not all the data tools in an organization need be managed by the data department, either. That may sound counterintuitive, but if a tool like Google Analytics serves only the marketing department it makes sense that the employees there will purchase, maintain, and use the tool for their own needs, and won’t involve the data department.
5. Data Analyst? Business Intelligence? There’s a lot of confusion in the market regarding job titles and functions. This creates an untenable challenge for HR companies and departments trying to find the right people for essential positions. Are you looking for a BI or a Data Analyst? The first step is to understand the various responsibilities. I see BI as just one part of the equation. BI tends to generate dashboards; the data analyst, on the other hand, has two main roles: a) to train other teams, and create dashboards and alerts for them, and b) to provide in-depth analysis and function as an additional layer in the business insight chain.
6. Data Scientist vs ML (Machine Learning) Engineers. We’re hearing the term data scientist a lot more lately, but what really is a data scientist? Let’s be clear: it’s a lot more than just running Tensorflow. A data scientist should have a MSc or PhD with deep mathematics knowledge, and the ability to take advanced algorithms in machine learning and modify them or write new ones. A ML engineer, on the other hand, is necessary for companies with simpler needs in machine learning. Indeed, all you need are some Python libraries and Tensorflow or Theano, and a good knowledge of Python and that’s it. You can ask my 18-year-old son. (To his credit, he’s very smart).
7. Data Engineers. The Data Engineer has a critical role in the basic elements of the data organization. Without his or her ability to drive data projects, the data team will not be able to provide valuable insights, and this frustration will move across the organization. It’s vital that the CDO, VP, and Data Engineer have continuous communications on data engineering projects, roadmaps, and delivery schedules. In fact, the Data Department is a “customer” of the Data Engineer’s team. It’s important to emphasize that the Data Engineer, at this time, has almost nothing to do with the data content itself; The DE is often called “the plumber” to keep the data in the right scale environment, and up and running 24/7. Unlike other roles in data, the Data Engineers are more engineer oriented, and as a result, in most organizations are part of the engineering department, but this can be short-sighted.
8. DataOps. In another article I wrote, “From DevOps to Dataops,” I emphasized the importance of the role in monitoring both the data and the business. Data integrity, things like business glitches, may seem small but must be monitored 24/7, on a minute-to-minute, hourly, and daily basis.
9. In the words of Star Trek’s miracle worker, Chief Engineer Scotty, “How many times do I have to tell you -- the right tool for the right job!” The market is presently flooded with analytics and data tools; the ones that will be most successful across your organization are the ones with built-in collaborative functionality, able to manage multiple users and teams, without cumbersome permissions functionality for the organization.
10. Be part of the design and deployment process. Every change in the schema, design, or distribution of a new product version, can dramatically affect consistency, benchmarks and internal quality of data. For that reason, it’s critical that the data department is always a part of those design discussion, just as Devops/IT is part of hardware capacity design for new features. Only the data department can provide the proper sense of importance during design meetings on the measurements, data implications, and quality of insights.
Remember, your data only has value if it’s actionable. If you follow the steps above and prioritize data throughout your organization, you will have a more perceptive sense of all your business activities, and be better able to meet your business goals. Becoming more data-driven isn’t easy, and there’s far more at stake than just the data itself. It’s about taking your company into the future by positioning yourself as ready and able in a data-driven world.