You are here:  Home  >  Data Education  >  BI / Data Science News, Articles, & Education  >  BI / Data Science Blogs  >  Current Article

Don’t Understaff and Overstretch Your Analytics Development Team

By   /  May 19, 2014  /  1 Comment

by James Kobielus

Enterprises increasingly worry about whether they can fulfill all their new data science requirements with qualified candidates. But an even larger issue has been simmering for years in many organizations’ data analytics practices.

The issue is that it’s typical for many companies to short-staff traditional positions, such as business analyst, report developer, database administrator, and ETL specialist. That’s why the people in these positions tend to stay overworked, stressed, and stretched to the limits. And it explains why they avidly seek solutions that can automate their workload, why they gladly offload some of it to consultants and outsourcers, and why they’re often happy when users pick up some of the development burden through adoption of self-service applications.

For many of these more traditional positions, there may be plenty of qualified candidates on the market. There may be a budget for hiring some of them. There may even be some degree of management support for expanding the data analytics team and hiring data scientists and other new specialists. But, even when all these stars align, there may still be no clear consensus within the organization on how many people of what specific skills are needed to shoulder current and projected workloads. There’s a clear downside to hiring people to support projected workloads that never materialize, or to handle a spike in requirements that turned out to be transient. No one likes to fire good people, especially when the people in question were perfectly qualified and capable for their positions.

However, the risks of data analytics team short-staffing are even worse. As this recent article notes, placing the entire burden on one full-time staffer or, much worse, a part-timer, is a recipe for disaster. Under those circumstances, the proverbial hit-by-bus scenario always looms, as does the risk that your entire data analytics team will simply cross the street to the competition some day.

As the article states, there is no hard-and-fast formula for determining how many people of which skills you should have on your data analytics staff. However, it nicely spells out the principal components of a full-fledged internal practice (program owner, project managers, technical staff, analyst staff, etc.). Clearly, each of these positions should have a sufficient number of personnel to delegate, share, and otherwise ensure that the work gets accomplished day in and day out in spite of potential showstoppers (e.g., deaths, departures). And the number, range, and skillsets of these individuals must be commensurate with the number and requirements of the functional groups and users requiring analytics support. Likewise, the staffing matrix must ensure adequate coverage for each of the data analytic platforms, tools, and applications in use across the organization. And the roles must cover all the requisite skill categories (e.g., business analyst, report developer, DBA, ETL specialist, data scientist) for the data analytic initiatives in production or in the works.

Even within those broad parameters, the optimal team size will depend on a host of variables that must be assessed on a case-by-case basis. The article presents a good illustration of how team-size recommendations might differ based on the specifics of two hypothetical companies.

Any such guesstimate must be couched within a comprehensive risk analysis. Having too many people on staff is a clear risk to the bottom line. Having the wrong mix of skillsets jeopardizes your ability to achieve the objectives of your data analytics initiatives. And having too few people–in aggregate, or in particular roles–creates the hit-by-bus risk, but may also hasten those people’s departures as their workloads become unmanageable and they go elsewhere to restore their work-life balance.

About the author

James Kobielus, Wikibon, Lead Analyst Jim is Wikibon's Lead Analyst for Data Science, Deep Learning, and Application Development. Previously, Jim was IBM's data science evangelist. He managed IBM's thought leadership, social and influencer marketing programs targeted at developers of big data analytics, machine learning, and cognitive computing applications. Prior to his 5-year stint at IBM, Jim was an analyst at Forrester Research, Current Analysis, and the Burton Group. He is also a prolific blogger, a popular speaker, and a familiar face from his many appearances as an expert on theCUBE and at industry events.

  • With the advent of more capable Data Warehouse Automation tools, modern-thinking organisations no longer need to invest in large development teams or expensive ETL tools. As long as developers hang-on to outdated and outmoded ETL tools and their forced-by-default Waterfall development processes, then your average BI project is doomed to being another statistically-inevitable failure.

You might also like...

Artificial Neural Networks: An Overview

Read More →