The data revolution is in full swing: data science practitioners are prospering and creating huge value for their companies. Despite this success, data science leaders across the industry are facing stress and difficult conditions. Data leaders must avoid these pitfalls to succeed and generate value for their organizations.
Limited Experience
Pitfall #1: Because the field is new and growing rapidly, many of today’s data leaders are operating with limited experience. With little knowledge of general management, new data leaders may make costly career mistakes.
Solution #1: Invest in professional development. Many tech companies don’t have training programs so the impetus for skill development falls to the individual. Courses in management, finance, and public speaking are advisable. Be honest about your limitations, as well as your intention to improve.
Lack of Understanding By Executives
Pitfall #2: Most business leaders and executives are aware of the hype surrounding data science. Yet, due to the young age of the field, knowledge of the actual capabilities and shortcomings can be limited. Coupled with inexperienced data leaders, this adds up to the potential for miscommunication and conflict.
Solution #2: Spend time crafting a narrative about your vision and goals for the use of data in the organization and stick to it. Anticipate questions and be ready to justify choices in data use, model selection and implementation.
Unclear and Changing Requirements
Pitfall #3: Data project management is prone to ambiguity in requirements. There are many unknowns about the nature of the data at the outset of a project. Adding to this, business stakeholders may lack knowledge of data science so expectations will evolve as they learn. This will cause requirements to shift throughout the project, creating more work for data leaders and stress for their team members.
Solution #3: Roll with the punches. Data leaders must learn to be creative and open to changes in order to have successful projects and harmonious relationships. As much as possible, data leaders should also shield their teams from headaches like unnecessary meetings and discussions.
Access to Data and Engineering Resources
Pitfall #4: Data scientists need access to high quality, timely data for analysis. Engineering resources are also required to deploy models developed by data scientists, and to continue gathering data on its performance. It’s unlikely that data scientists themselves can be expert in all pieces of this process, from data identification, and gathering through to development, QA, deployment and continuous monitoring. Without access to these resources, data leaders can not succeed.
Solution #4: It is necessary for data leaders to form the relationships necessary to access such resources and expertise. Formal processes may not exist. Data leaders must get buy-in and make their case to busy Business Intelligence (BI) and engineering teams in order to get their work into production.
Recruiting & Talent
Pitfall #5: Recruiting talented data scientists is difficult. The massive growth of positions has proceeded faster than newcomers can accumulate experience. For the time being, a relatively small number of experienced data scientists are heavily recruited. Recruitment is a sink-or-swim challenge for today’s data leaders.
Solution #5: Develop projects that are both technically interesting and deliver important results for the company. Data scientists choose and keep jobs that have interesting and important projects. Also, starting an internship program will allow your company to take advantage of the tremendous interest in the field from newcomers.
Beyond Data Science 