If your organisation is seeking to appoint a data scientist, or it is your current job title, our recent online panel discussion about all aspects of the role offered lots of food for thought.
The webinar, hosted by our Committee Member Nicola Askham, brought together four data science experts to share their views:
- Marisa Murton, Head of Data Science at Incited
- Dr Sarah Schlobohm, Head of AI at Kubrick
- Linford Bacon, Co-founder and CEO of Ngenious AI
- Dr Shorful Islam, Co-founder and CEO of Be Data Solutions
The lively discussion was a whistle-stop tour of many aspects of data science: from trying to nail down a definition, to helping firms to identify the optimal time to create and fill the role.
While many organisations are in agreement about the growing importance of data scientists, the role doesn’t seem to have a single definition. The Government’s own career guidance website reveals how tricky it is to apply a standard job description: “Data science is a broad and fast-moving field spanning maths, statistics, software engineering and communications.”
This is a problem for organisations. With more than one million data scientists already employed at UK businesses by mid-2021 it’s vital for firms to understand what these data experts can do for their business. Organisations often rush to recruit them without first thinking how they’ll really help the business - and what needs to be in place before they arrive so that they can hit the ground running.
The panel began by discussing some of the misconceptions about the role and how this issue affects current performance. Participants pointed to a “broad church” of inter-related disciplines mentioned in job descriptions, before attempting their own definitions.
Marisa believed the data science role exists to use computer algorithms to extract and apply insights from data, informing real-world decisions. Shorful explained the scientist component describes using data to generate and test hypotheses about unresolved business challenges.
The panel ultimately agreed that a single definition isn’t available - and might not be helpful for business anyway. Their view reflects the fact that organisations now have access to such vast volumes of data - beyond “big data” alone - that its uses for data science are also broad.
There is an imperative to have a clear understanding of the need to hire a data scientist, and your expectations of the value and outcomes that you want them to deliver, before you set about hiring one. In other words, what can data science do for your business, and how do you create the conditions for them to be successful?
Knowing when to hire a data scientist
With this in mind, the panel explored how organisations can make a business case for appointing a data scientist once they’ve understood how the role can make a difference to operations and outcomes.
The benefits of data science identified by our expert participants include:
- the ability to understand changes to consumer behaviour
- meeting more savvy customers’ demands for a smooth, joined-up experience when interacting with an organisation
- pinpointing opportunities to create new products and services; and
- ensuring the business can keep pace with rivals that are also employing data scientists (as long as you have agreed your objectives in advance of hiring one, of course!)
Clearly, it’s not a decision to take lightly - and there’s a big risk of jumping on the bandwagon by hiring too soon. Many factors across an entire business must be considered to lay the groundwork that will make the data scientist’s efforts as successful as possible.
Linford stated an existing business intelligence function is vital. An organisation’s raw data should already be available and catalogued, with data governance in place. The data scientist will then be able to understand the data and start to use their “advanced data toolkit” as soon as they’re in post.
Then there’s the perennial issue of relationships between data teams and the IT department. A data scientist who starts in the role without agreement over the necessary data permissions and appropriate controls could have a fight on their hands to achieve their own goals and those of the business. Involving IT during the role scoping process can prevent problems further down the line. For instance, the data scientist might build a product only to find it isn’t compatible with IT’s parameters.
Communication is crucial. There can be a fear of introducing data-led decision-making, agreed the panel, with employees worried that automated processes created by data science will eventually replace them. Sarah said that’s why the whole organisation should be reassured that, in fact, they’ll be able to do higher-value work as a result of data-driven operational improvements.
Finally, in counterpoint to the issues set out above, the panel believed that much smaller businesses that are keen to appoint a data scientist should be bold and hire early. The right expertise can help the firm set up successfully, building the correct business model based on the volume and quality of data available.
Getting the best from your data scientist
Summarising the discussion, the panel returned to their personal experiences of data science as a profession. They cited the components required to ensure the role is as beneficial for the person living and breathing it on a daily basis, as it is for their organisation.
- Room for manoeuvre - Ensure the data scientist has scope to explore, hypothesise, and test and learn.
- Opportunities to collaborate - While the nature of data analysis is often highly individual, the data in question lives in many different departments, therefore highly varied outcomes are desired. The opportunity to build strong relationships with key stakeholders is imperative.
- Measure more than ever - Measurement is a major part of the data science business case and expectation management. The firm needs to ask the right questions of data science. That could even mean creating a “conduit” role between the two, to challenge the business request but also help the data scientist stay on track and articulate the findings back to the organisation.
Conclusion: The era of data science
It’s undoubtedly an exciting time to seek a career in data science. The panellists have followed diverse routes into the role, and believe it draws on many skills, subject areas and even personal characteristics.
How has your own blend of these attributes determined which path you’ve taken into a data science role? And how would you describe your own organisation’s approach to data science? We’re keen to hear about your experiences.