In October 2012, Harvard Business Review published an article by Thomas H. Davenport and DJ Patil titled "Data Scientist: The Sexiest Job of the 21st Century." The title was, by design, provocative — and it worked. The phrase has since become close to inescapable in discussions of STEM careers, quoted in job postings, conference keynotes, and university program announcements. Five months on, it is worth asking what the article actually argued, and whether the career category it named is turning out to be as durable as the headline suggested.
Patil's authority on the subject was not abstract: he had, together with Jeff Hammerbacher, been credited with coining the term "data scientist" itself while building analytics teams at LinkedIn and Facebook in the years before the article's publication. Davenport, a longtime analytics and business-technology researcher, brought the framing of data science as a business discipline rather than purely a technical one. Together, the two authors argued that a new hybrid role had emerged — part statistician, part software engineer, part business analyst — that existing job categories and existing university programs had not caught up to.
What the Article Actually Claimed
The core argument was narrower and more specific than the "sexiest job" headline suggests. Davenport and Patil described a role that combined three distinct skill sets that had historically lived in separate departments: the statistical and mathematical training of a quantitative analyst, the software engineering skill needed to work directly with large, messy, unstructured datasets at scale, and the communication and business judgment needed to translate technical findings into decisions that non-technical executives could act on. Their claim was that people who combined all three were extremely scarce relative to the sudden demand from companies sitting on unprecedented volumes of data — web logs, transaction records, sensor data — that they did not yet have the internal capability to interpret.
The scarcity claim is the part of the article that has held up best under early scrutiny. University statistics, computer science, and business programs in 2012 were, for the most part, still training students in one of the three component skills rather than all three together, and hiring managers in the months since publication have repeatedly described the same gap Davenport and Patil identified: candidates with strong statistical training but limited engineering skill to work with large datasets directly, or strong engineers with limited statistical grounding to know what questions were worth asking of the data in the first place.
The Immediate Hiring Effect
In the months since the article's publication, "data scientist" job postings have visibly proliferated — not only at the technology companies (Google, LinkedIn, Facebook) whose early data teams the article drew on for examples, but across retail, finance, telecommunications, and media companies building out analytics functions of their own. Recruiters and hiring managers report that the article's framing has been directly useful in internal conversations: naming a role and describing its distinct value has made it easier for companies without an existing analytics function to get executive buy-in to create one and to justify the salary premium the role commands relative to more traditional statistician or business analyst titles.
This naming effect should not be understated as a career-pathways matter. Before the article, a hybrid statistician-engineer working on large datasets had no consistent title to search for, no consistent job description to aim a resume at, and no consistent salary benchmark to negotiate against. The proliferation of a shared vocabulary — however manufactured by a magazine headline — has real practical value for job seekers trying to identify and pursue this kind of work, and for university programs trying to design curricula that produce graduates employers are actively looking for.
What Remains Genuinely Uncertain
It would be premature, five months after a single magazine article, to declare a permanent new professional category settled into the STEM landscape. Several open questions bear watching. First, whether "data scientist" solidifies as a distinct, durable job title with its own career ladder, or whether it proves to be a transitional label that gets absorbed back into more established categories (statistician, software engineer, business analyst) once the current scarcity eases and hiring practices mature. Second, whether university programs will develop dedicated data science degrees and tracks, or whether the skill combination will continue to be assembled informally by individuals combining separate statistics and computer science coursework. Third, and most relevant to WIGSAT's audience, whether the demographic composition of this new hiring category will differ meaningfully from the demographic composition of its component fields — statistics and computer science both have long-documented, and quite different, patterns of women's representation, and it is not yet clear which pattern the blended "data scientist" role will follow.
On this last point, the honest answer in early 2013 is that no comprehensive demographic data on the emerging data scientist workforce yet exists publicly. The category is too new, and the companies building analytics teams have not, for the most part, published disaggregated hiring data. It is a question worth tracking as the field matures over the coming years, rather than one that can be answered from the available evidence today.
Why This Matters for STEM Career Planning Now
For a student or career-changer weighing STEM career paths in early 2013, the practical implication of the Davenport-Patil article is not that "data scientist" is a guaranteed safe bet, but that the underlying skill combination it describes — statistical fluency, software engineering competence with large datasets, and the ability to communicate technical findings to non-technical decision-makers — is a combination that multiple industries are independently discovering they need, regardless of what the eventual job title settles into. Building competence across all three areas, rather than specializing narrowly in only one, appears to be the more durable career strategy suggested by the pattern the article describes, even if the specific title "data scientist" turns out to be more transitional than permanent.
How Universities Are Beginning to Respond
One visible early effect of the attention the article generated is movement inside university statistics, computer science, and business school departments, several of which have begun discussing — and in a small number of cases already designing — dedicated data science coursework or certificate programs distinct from their existing statistics and computer science majors. This is a slower-moving process than corporate hiring, since new degree programs typically require formal curriculum committee approval and can take a year or more to launch even once a department decides to pursue one, but the direction of movement is notable: several research universities have publicly confirmed they are evaluating whether a combined statistics-and-computing track, an expanded set of cross-listed electives, or an entirely new program is the right response to employer demand for the skill combination Davenport and Patil described.
This lag between employer demand and formal academic response is itself informative for someone planning a STEM career path in 2013: for the next several years at minimum, a student who wants to build the specific statistics-plus-engineering-plus-communication combination the article describes will likely need to assemble it deliberately from existing course offerings — a statistics sequence, a computer science sequence emphasizing data structures and databases, and deliberately seeking out communication-heavy coursework or project experience — rather than relying on a single dedicated "data science" major, which in most cases does not yet exist as a formal credential.
The Risk of Overcorrection
A phenomenon worth naming directly, five months into the surge of attention the article has generated, is the risk of overcorrection: as "data scientist" becomes a fashionable job title, some employers appear to be applying the label to roles that do not actually require the full combination of skills the term was coined to describe — relabeling existing business-analyst or reporting-analyst positions as "data scientist" roles without meaningfully changing the job's actual technical requirements or compensation. Job seekers evaluating postings using this title should look past the label to the specific technical requirements listed — whether the role genuinely requires hands-on work with large, unstructured datasets and independent statistical modeling, or whether it is closer to a conventional business analyst role that has simply adopted the more fashionable title. The gap between the title's prestige and a specific job's actual requirements is likely to be a real and recurring source of confusion in the hiring market as the term's popularity outpaces any standardized definition of what the role actually requires.
Frequently Asked Questions
What did the "Data Scientist: The Sexiest Job of the 21st Century" article actually argue?
Published in Harvard Business Review in October 2012 by Thomas H. Davenport and DJ Patil, the article argued that a new hybrid professional role had emerged combining statistical training, large-scale data engineering skill, and business communication ability — and that people combining all three were extremely scarce relative to sudden corporate demand for the role.
Who coined the term "data scientist"?
DJ Patil, co-author of the article, is widely credited with coining the term alongside Jeff Hammerbacher while the two were building analytics teams at LinkedIn and Facebook in the years leading up to the article's publication.
Is "data scientist" a stable, permanent job category?
It is too early to say with confidence. As of early 2013, hiring for the role is proliferating rapidly, but whether it becomes a durable, distinct career ladder or gets eventually absorbed back into existing statistician and engineering job categories once the current talent scarcity eases remains an open question.
What skills should someone build to pursue a data scientist career path?
The role as described combines three components: statistical and mathematical analysis skill, software engineering ability to work directly with large, messy datasets, and the communication skill to translate technical findings for non-technical decision-makers. Building competence across all three, rather than specializing in only one, matches the skill combination employers are reporting difficulty finding.