In any field, data science is an emerging area where roles can be hard to define or quantify by hiring managers. For those looking to employ data scientists within their business, this role can lead to a wealth of unanswered questions. This chapter from our Data Solutions for Drug Development Report, sets out key tips for hiring and resourcing data scientists, and looks at how best to resource and retain data scientists within the pharma industry.

HIRING AND BEYOND: HOW TO RESOURCE DATA SCIENTISTS IN YOUR BUSINESS

EXECUTIVE SUMMARY

• specific expertise within pharma, an understanding of the role’s requirements and good communication skills to different levels of audience are vital in a new data scientist hire
• data scientists coming from the consulting sector work well in pharma, given their broad industry knowledge and communication training; those with a background in retail or technology have fewer translatable skills
• employers should ensure staff are resourced sufficiently, including attending conferences and events to communicate with the wider community to also keep up within the changes in data science
• managing career expectations and understanding specific employee motivation for data science are key to staff retention

INTRODUCTION

In any field, data science is an emerging area where roles can be hard to define or quantify by hiring managers. For those looking to employ data scientists within their business, this role can lead to a wealth of unanswered questions: What does a good hire look like? How should data scientists be resourced? And how do you retain skilled professionals in such a difficult and complex field?

The good news is that hiring, while often a nebulous practice, can be done right. Understanding the best methods of hiring, the scientists’ relationship with the rest of the business, and most importantly the role itself, can go a long way to hiring the best people for the role and growing your data ROI considerably.

KEY TIPS FOR HIRING AND RESOURCING DATA SCIENTISTS

Dr Nuray Yurt, Data Science Lead for US Oncology, Novartis, said that hiring the right data scientists for pharma was a challenge. She argued that while general technical proficiency and out-of-the-box thinking were important, specific industry experience was also vital. The key to hiring well was to find a balance between the two.

Regardless of specific role, all data science candidates should be proficient working with:

• analysing data and presenting results to others without a technical or analytical background
• the relevant software or mathematics to work alongside others in their analytics team with different expertise

With data science, the implications of hiring the wrong people for your business can be far greater than doing so in other sectors. Given pharmaceuticals’ particular way of collecting and processing data, there are industry-specific data processes which new hires must know in order to create accurate and useful models in their work. This means it’s critical for new hires to not only know the necessary technical aspects of the job, including how to model generally or work with machine learning as a function – greater specialisation than normal is also needed to ensure candidate success.

Communication is also something very often overlooked by hiring managers. To ensure data scientists provide real value for a company, employees should have the ability to talk to other departments, such as medical, marketing and sales fluently and understandably. This is both a highly-valued and valuable skill, and a difficult one to find.

Dr Yurt recommended those scientists coming from the consulting sphere: those with consulting experience typically work across industries in their career, and can apply previous models to new circumstances. Those without such experience often don’t know the differentiation between industries, leading in some cases to the wrong choices being made when picking models to use.

Dr Yurt claimed that another thing to look out for were those scientists who have previously only worked in retail or technology: those backgrounds differ in terms of data structure and quality to pharma, and as such often lack the experience and understanding that a better-focussed role might have.

RESOURCING DATA SCIENTISTS

Regarding resourcing data scientists, Dr Yurt suggested more trainings and learning opportunities should be taken into consideration, particularly regarding investment in training. In particular, training around how other areas of the business work, including manufacturing and clinical trials, would allow scientists to develop further and have greater understanding of the relationship between their own work and the wider business.

Ensuring data scientists attend relevant conferences and events was also cited as a key area of consideration. This would again ensure greater connectedness between scientists and the wider community, propagating the growth of the business and enhancing the employees’ knowledge base at the same time.

RETAINING DATA SCIENTISTS

Hiring is, of course, not the only issue around employment that businesses need to consider. According to an article published recently by the Financial Times, the average data scientist “spends 1-2 hours a week looking for a new job”. Machine learning specialists are also cited in the article as being the most likely type of developer to be looking for a new job, at 14.3%.

The biggest reason for these retention troubles is often cited as the fact that role expectation rarely matches reality. This occurs for a number of reasons:

• data scientists are often hired without a suitable infrastructure in place for working with data, something they are often tasked with fixing as a first role
• often the impact of the role is limited to improving the company’s data, with limited impact outside that business
• a lack of understanding from other members of the organisation around the role, and distorted expectations as a result

These difficulties, damaging to a company while they remain, are luckily easy to banish entirely. An employer or company can fix their retention issues by:

• ensuring the goals of the data scientist are aligned with that of the company—possibly the most important issue to focus on, doing this will mean employees will not enter the business with unrealistic expectations. Letting them know the early focus of their work, the team they will operate in and the expectations of their specialties and area of operation will ensure they enter the business knowing their role and willing to do it well
• ensure senior or experienced data practitioners are hired before junior ones, so employees are not overwhelmed by requirements or tasked with problems well beyond their prior experience
• ensuring employees in all related areas who will work with the data scientists know fully what the role entails, and what each individual cannot do

CONCLUSION

While hiring a data scientist is a more unique challenge than finding employees for other roles, it shouldn’t have to be an onerous duty. Advertising the role where those with experience will see it – at conferences, in specialist groups and on LinkedIn – will increase the number of relevant candidates considerably, and hopefully shorten hiring time as a consequence. Understanding the vital requirements for candidates, and ensuring they are aware of the precise role and expectations within it, will not only ensure your employees are the best possible but that they stay within your company for a great deal more time.

CASE STUDY: METHODS OF RECRUITMENT

THE PROBLEM

Dr Yurt described a situation in which she was hiring a data scientist for a particular board role, and was finding it a challenging prospect.

Hiring through the regular recruitment channel, which functioned largely by posting adverts on Novartis’ website and utilising headhunters to find the right candidate, was turning up individuals who were not right for the role, or lacked particular qualities.

THE SOLUTION

Dr Yurt, realising the situation was not working for hiring data scientists, changed her recruitment method to something considerably more specific, and for her hands-on.

This involved:

• posting the role on LinkedIn for individuals doing selective searches
• posting the role personally as the hiring manager, and someone who had great experience working with and understanding the role of data scientists
• connecting with data science conferences and working with them to identify suitable individuals

THE GAIN

As has been previously stated, the role of data scientist is one which requires more particular experience and background than do other roles, or even as does the same role in different sectors.

To hire the right individual, it is important not just to rely on the traditional methods of recruitment but look further afield, where there is greater potential for meeting professionals with the correct skills to work well in your business.

Working with relevant conferences or professional bodies or groups is a particularly innovative way to encounter talented employees without sorting through a large number of other individuals without the relevant skills.

Data Solutions for Drug Development Report

This chapter is from our Data Solutions for Drug Development Report, which focuses on the data developments most critical to the drug development industry. The report looks to provide the solutions that pharma companies need to stay ahead of the curve and understand the importance of data in their workplace. Using both in-depth contributor analysis and high-level case studies from businesses leading in this field, its intent is to demonstrate the importance of data today and provide solutions to a number of problems many pharma companies, both large and small, are facing. You can download your free copy of the full report here – http://info.frontlinegenomics.com/data-solutions-for-drug-development

The chapter was put together with help and input from Nuray Yurt, Executive Director, Enterprise Analytics & Data Science, Novartis. Nuray will be presenting a case study on Shaping Your Business Strategy with Advanced Analytics at our upcoming D4 (Data-Driven Drug Development) USA Conference, taking place in Cambridge, MA this October.

You can find out more about Nuray’s session and the full agenda of evidence-based case studies from the world’s leading minds in pharma by visiting the event website here – https://www.d4-pharma.com/

Unable to join us in Cambridge? We will also be hosting D4 Europe in Basel this October, you can view the agenda and speaker line-up for the European event here – https://www.d4-europe.com/