top of page
Search
  • Writer's pictureDeandra Cutajar

Hiring a Data Scientist


When I started this blog, I intended to share as many data science elements with the broader public to bridge the gap between data science and businesses or natural persons. Sharing data science knowledge around techniques and algorithms is one way to do so. The other is to speak directly to the heart of the business, costs.

**All the estimates below will be linked to their sources to help businesses plan their costs and understand their expenses.


When hiring a data scientist, the company must consider different factors to estimate the cost of hiring and its sustainability. The four pillars upon which this can be achieved are:

  • Business Needs

  • Seniority of the New Hire

  • Budget Allocated

  • Resources to be Allocated


A. Business Needs


The team or company requesting a new data science hire must provide metrics to support the budget allocated for the new employee. This is important when considering costs and return of value for the new data scientist. One obvious metric is the RoadMap and Project backlogs. The other is Stakeholder Management.


A data scientist usually works on two or three projects simultaneously, depending on experience and seniority. Thus when justifying a new hire, one metric required to determine the need for this position is to quantify the capacity of a data scientist and compare it with the project roadmap or backlog. I suggest that the roadmap would be already approved by the business to ensure that all the work is a requirement. Assuming that the projects planned in the roadmap are usually fixed for three months with some flexibility after that, the projection ought to justify why the current team cannot successfully work on the roadmap when considering the team's priority tasks and development.


Another important metric, or additional requirement in data science job descriptions, is the ability to communicate with business stakeholders. Data scientists are expected to discuss data science techniques with non-tech audiences. This is especially the case when companies transition to a data-driven culture. Such a position may not be subject to projects and roadmap but rather a need to educate. In another article, I will discuss the importance of this element against costs!


Determining the reason behind hiring a new data scientist will ensure that the budget allocated for the role is within financial planning and ideally sustainable for years.



B. Seniority of the New Hire


New hires' experience level will reflect their expected salary and how quickly they can adapt to the business during probation. The company must understand what would make a best first or new hire. A junior data scientist expects lower salaries but usually requires a mentor to understand the business and its processes. While the coding level may be somewhat similar to a senior data scientist, their limited experience means some things are taken for granted. Moreover, communication with business stakeholders may not run smoothly. There are exceptions to the norm, and these should be exposed or highlighted during the interview process. A junior-mid data scientist can quickly focus on building models and hands-on work if a team and support are available.


A senior data scientist has much experience in stakeholder management and leveraging the tools provided to deliver a project. Sometimes, data scientist refuses to engage with stakeholders, but their programming skills are exceptional. Senior data scientists have enough experience to interview new hires and grow the team, even if they didn't manage a team per se in the past. More often than not, senior data scientists work closely with their managers, provide technical feedback, and support other data scientists. They are involved in discussions around data science processes, even building data and AI strategies. Of course, senior data scientists ask for higher salaries and are usually not keen to fill a role requiring basic skills.


The company should consult with external parties when hiring the first data scientist. If the new hire is an additional talent to the team, then the manager or first hire should be able to guide these matters. External consultants also can support this process.



C. Budget Allocated


The salary of a data scientist is one of the first discussions around budget costs. However, to get candidate leads, the company must promote the vacancy. The prices of such promotions depend on the experience level the new hire is expected to have, the influence in the field, and the business requirement.


Getting leads into the interviewing process can be done in two ways:

  • Advertising the vacancy on different platforms.

  • Hiring external recruiters to head hunt professionals in the market or top candidates.

Advertising a vacancy on platforms can be free or based on a subscription. LinkedIn allows for a free job posting, or companies can decide on a budget for promoting the vacancy. For more expert positions, the budget would be higher. It also depends on how long the ad should run for, and so on.


According to Indeed, external recruiters charge around £3,000 per new hire or 15-30% of the salary. Considering an average salary of a data scientist to be £ 57,000, as seen on Talent.com, that amounts to £ 8,550 to £ 17,100. When deciding between these two, there are some factors to consider.


When anyone can view the job vacancy, there is a higher chance that many applicants do not meet your desired expertise. That means one or more employees must be allocated to vetting these applicants. Usually, these would be HR personnel and the hiring manager.

The average salary of a data scientist in the UK is around £ 30 per hour, as indicated on Talent.com. This amount varies by seniority, but assuming this ballpark value, if one employee spends a week vetting applicants, and considering the working week is 40 hrs, the company spends £ 1,200 per week on vetting alone. The more employees that are involved in this, the higher the cost. One week is conservative, and usually, the vacancy is open for at least a month, if you are lucky, during which interviews will start. Assuming the process ends in one month, the company would spend around £ 4,800 for a new hire. This number excludes licenses and costs of using any tool to conduct such interviews.


Looking at these numbers, we now understand that if the process takes two months or includes two employees, the cost of hiring a data scientist exceeds that of recruiting recruiters. Most data science vacancies come with three or four interviews, some even 5 to 7 rounds. Therefore, at best, the estimated cost of hiring without a recruiter is minimum.


On the other hand, hiring an external recruiter can be out of budget. However, it may lead to quicker results. Recruiters call the applicants or contact potential candidates in their network. They ask the questions necessary to the hiring manager and begin to filter the candidates out, thus minimising the effort of internal employees. In this case, the company will continue its business as usual because its employees continue to work on projects rather than putting that on hold. The longer an employee stays away from an ongoing project, the longer it will take to return value.


D. Resources to be allocated


When calculating budget costs for hiring, companies often overlook the resources required for the new data scientist hire to carry out the job from day one. These include laptops and other devices, which would be necessary but also licenses to software, access, permissions and CPU quotas for the scientist.


A junior-mid data scientist may spend the first week or month shadowing other members or getting an introduction to the company. A mid-senior data scientist is usually invited to meetings to catch up. During the first week, a senior data scientist may need the resources to develop code, primarily since the increase in headcount is usually motivated by the lack of capacity for the projects' backlog and require immediate deliverables. That means the senior data scientist will often be allocated to a project from the first week and expected to deliver during the probation.


Knowing what the company needs and supporting a new data scientist hire with business demand is the easiest and most sustainable way to hire new employees. Data science is often regarded as a nice-to-have, but companies have quickly realised that the cost of not having a data team is to miss out on opportunities and get behind the competition.



In another article, I will discuss costs around a data science team, including maintenance, learning & development, budget allocation and overhead costs.


For Data Science Team Building consultation, kindly reach out on LinkedIn or email me at deandra.cutajar13@gmail.com.




116 views

Recent Posts

See All

Comments


bottom of page