Generalists and Specialists in Data

Our article about roles within data helps break down the roles and responsibilities of general data archetypes. An additional layer is how well someone fits within one — or multiple — of these archetypes and how much they can stretch across multiple archetypes to fit the needs of your organization.

Generalists versus specialists

The three archetypes we identified rarely form three mutually exclusive groups and generally look more like this:

Though some data practitioners gravitate toward one of the main archetypes, some may have a mix of skills across some or all of the archetypes. We refer to these folks as generalists, or those who have a breadth of knowledge across the data spectrum and can hop in and out of multiple archetypes as needed. On the opposite end of this spectrum are specialists, or those whose skills fall more cleanly into one of the archetypes. These folks tend to have an unparalleled depth of knowledge in a particular area.

So how do you identify whether generalists or specialists would be better for your organization?

What you have will influence what you need

Using the results from the questionnaire you completed can help highlight what is needed in terms of data — and the severity of those data needs — for your organization. Further, your needs can vary based on a variety of factors about your company, including its value proposition, size, and existing team.

Value proposition

Your organization likely has nuance and requires some specialization with regard to skill sets:

  • If your business is specialized or in a specific field or industry, you may need a specific skill set (e.g., an Actuary or Claims Analyst)
  • If your business is product heavy, you’ll likely need someone experienced in experimentation (e.g., Statistician)
  • If your product or service is data-heavy, you may over-index on machine learning folks compared with other organizations (e.g., Machine Learning Engineer)

You’ll likely start by hiring folks with the specialty skill sets necessary for your business. What can the specialist hires do? More importantly, what will they not be able to or not want to do? When you revisit your data plan and remove the work your more specialized data folks can tackle, which necessary skill sets remain?

Company size

Though your company’s size does not always dictate how many data folks you might need, it can be a good way to think about the kind of folks to hire.

Company Sizemay want to hirewhobut may lack
Smaller start-upgeneralists(1) are good at thinking about the business at large, (2) have a breadth of skills across all three archetypes, and (3) can quickly get data projects off the grounddepth within each part of the data spectrum
Scaling or large start-upspecialistscan (1) develop subject matter expertise and (2) go deep in specific areasbreadth across the data spectrum

This is not to say that smaller start-ups should exclusively hire generalists and scaling or large start-ups should exclusively hire specialists. Rather, smaller start-ups will likely lean toward hiring more generalists versus specialists, and vice versa for scaling or large start-ups.

Some common generalist roles include those in which data folks:

  • Load and transform data while bringing transparency to the organization through reporting
  • Implement event tracking while designing and analyzing experiments
  • Build machine learning models while enabling the machine learning infrastructure

At a smaller start-up, the work changes quickly and in unexpected ways. Generalists can take on new, vague, and unknown projects in stride, flexing their breadth of knowledge. As companies become more mature, the work becomes more consistent, well-scoped, and robust. This is typically when organizations begin to switch from generalist-leaning hires to specialist-leaning.

For scaling or large start-ups, data specialists may be able to complete some projects across the data spectrum, but would be far more valuable — and likely happier — working within a specific part of the data spectrum. For the most part, these folks would flex skills within their archetype while collaborating with others on the data spectrum as needed.

Existing team

Your existing team holds valuable institutional knowledge and represents a considerable investment. You’ll want to make sure your subsequent hires increase and complement the existing team’s capabilities.

If you have an existing data team, you’re not starting from scratch and your needs will be different from an organization with no data resources: 

  • If you have a very specialized team and you’re looking for more flexibility across the data spectrum, you may want to consider finding folks with a more generalist skill set
  • If you have a team of generalists, you may want to consider your biggest areas of opportunity within the business for specialists and target those skill sets

Your existing non-data team can also influence your data hires. For example, some non-data leaders may be stronger in data than the average person in their role or have folks on their team who can do a lot of the data work, obviating the need for additional data resources

There are likely other factors that you may need to take into consideration, but there are some general principles that can help you better navigate your unique situation.

General principles

Hiring your first few data folks can be exciting, but it’s also critical that you find the right kind of skill sets to get you started. These principles can help ensure you’re starting off on the right foot.


Regardless of your organization’s size and structure, you’ll need to have a solid foundation upon which your data consumers — including Analysts, Data Scientists, and other stakeholders — can successfully access accurate and reliable data. It’s important to note that there are many more consumers of data than there are folks who are skilled at building a data infrastructure.

The good news is that many software start-ups have popped up in the past few years that help organizations build a solid data foundation without an internal Data Engineering team. From ingesting third-party data to managing ETL pipelines, these tools make it easier for non-infrastructure folks — especially those familiar with data — to implement and manage their data infrastructure. 

However, there are a few things to watch out for if you choose to go the tools-first route:

  • It’s likely a short-term solution
  • There are some advanced topics for which existing tools may not yet solve (e.g., advanced security and data handling practices)
  • Someone with infrastructure experience would be helpful in instrumenting the tools effectively

As you scale, you’ll want to make sure that your data foundation scales at the same pace as your business. In some cases, the sooner you lean into infrastructure that is owned and managed by your team, the better. You may need to consider finding folks with infrastructure skill sets early as this will enable you to not only build a more nuanced and customized data foundation, but also implement more robust tools to manage your increasingly large and complex data.


Whether you’re trying to understand how your company is doing in general, determine what to build or iterate on next, or measure the impact of your efforts, your first data hire is likely going to be a generalist who can be your Swiss Army knife of data.

When you’re looking at your data needs, you may determine that you need a few skill sets from each of the three archetypes. This does not necessarily mean that you should hire one of each, as the amount of work may not require a full-time resource of each archetype. Instead, you might want to try and find someone who has some experience in two or three of the archetypes, but does not specialize in any one area. 

For example, if you’re looking for someone to build a data foundation and get some core dashboards up and running, it’s probably not in your best interest to hire someone who wants to build complex models. Similarly, if you want someone to build core dashboards and build a straightforward forecasting model, your ideal candidate would be someone who has experience in the Analyst or Scientist archetypes and is willing to work across the two.

Though generalists can specialize, their superpower is being good at a lot of things and not going deep into any one thing. At some point, you’ll likely need to hire folks who specialize. The good news is that specialists can complement generalists well, and having an appropriate mix of generalists and specialists can be powerful for your organization.

Execute, evaluate, and be prepared to pivot

Smaller and scaling organizations can move incredibly fast; similarly, your data needs are also likely changing quickly. Constantly evaluating your position based on new information will help you focus on the best path forward, even if that means changing course.

If you’ve been thoughtful about what you need to make progress on your data goals and are regularly evaluating your progress, you’ll be able to see how your needs are changing over time as your staffing and business evolve. For example, progress on the infrastructure allows you to focus more on analytics efforts, and the implementation of a new experimentation platform enables you to focus more on the experimentation process.

Learning from your data and answering critical business questions may also change your data priorities — a big win for your data-driven efforts! Initially, you may have set your sights on a recommendation engine, but after learning more about your business, you might realize there are higher impact projects you can tackle without needing to hire additional folks. Adapting to your constantly changing data landscape and priorities will pay dividends and ensure neither your data priorities nor data team is outdated.

How to put this into action

If you are interested in learning more about what this means for your organization, fill out this questionnaire to get started!