In 2012 the Harvard Business Review identified ‘data scientist’ as ‘the sexiest job of the 21stcentury’. The marketing world collectively responded with excitement – name dropping various data science-y buzz words in blog posts and meetings. Since then, we’ve seen data scientists appear on the client, agency, and publisher side. However, data science itself is still perceived to be mysterious, technical and elite.
Despite the term data scientists becoming widely used, there’s no industry-wide consensus on what the title means. Depending on who you ask (and where they work), one person’s ‘data scientist’ could be another’s business analyst or hybrid market researcher.
This is confusing clients and agency people alike. To truly understand the value that data science offers, and to leverage it to the best of its ability, there must be increased awareness of what data science can enable. Perhaps the starting point is a broader understanding of what – and who – data scientists are.
New Tools & Titles
Some confusion has been caused by the new tools coming into the market that make previously difficult-to-obtain data more accessible. Analysts and marketers with little technical background can access previously inaccessible data via third party tools such as social listening, emotional language analysis, and natural language processing platforms.
Further, tools such as Power BI make it straightforward to automate the reporting of large, disparate data sets on a user-friendly front end. No coding, no crunching, just smooth sailing. All good things, but this also contributes to the murkiness of the data science role.
Let’s look at a few examples of the different types of data scientists…
At a conference recently, I spoke to the data lead at a large global FMCG brand. Throughout our discussions around this topic, it became clear that this brand’s ‘data scientists’ are what other agencies would call analysts or statisticians – highly skilled analysts who can run advanced predictive and propensity modelling based on large customer datasets.
On the other hand, I’ve worked at media agencies where ‘data science’ loosely refers to teams that specialise in market mix modelling or econometrics. More recently, I was at a research conference where there was a debate around whether anyone who works in data is, in fact, a data scientist (in a way, data in marketing and communications is looking at the science of data, no?).
As you can see, it gets quite confusing and the lines are blurring. ‘Data scientist’ has become a catch-all term for people who extract data in a variety of ways – and a cynic might argue the term is being used to make a job sound sexier (I mean, I’ll take it…).
So, while the term data scientist can mean different things to different people, it’s essential to ensure we aren’t missing the opportunities to maximise the potential of ‘hard-core’ data scientists.
Now you’re probably wondering about these data scientists – the wizards who can find any kind of data from any source. Where do they fit? Well, it’s all about the language.
All Hail the ‘Hard Core’ Data Scientists
Some people speak the language of love, others speak Java or Parseltongue (excuse me, Python). Largely speaking, data scientists are considered by many to be ‘hard core’ coders (I’m clearly writing this from the point of view of a data “artist”). These are the people I consider the ‘hard core’ data peeps. Those who can extract data without relying on plug-and-play software – they can instead seek data through APIs and other means of extraction.
They can build their own interfaces and visualisation tools, apply machine learning and AI, and give businesses access to data and insights that aren’t readily available.
For example, I recently learnt that a data science colleagues specialises in computer vision. She can (and has) built bespoke tools that recognise images in pictures and videos and correlate performance of content based on these images. That’s truly the magic of data science – and this is just one example.
The problem (and opportunity) is there’s so much data scientists could do, but clients and agency people don’t know about it. Even in my role as lead of a data team, I am consistently in awe of what’s possible. Furthermore, it’s become clear that the briefs we give data scientists are often limited by our understanding of what’s possible. Therefore, we risk giving them basic tasks; an effective way to demotivate a skilled person.
Tell Clients What Data Science Can Offer
So perhaps the take-out is twofold:
1) Address awareness of what data scientists are truly capable of. Ask them how they can push the envelope further and move beyond basic text analysis, visualisations and communicate what’s possible with the agency and upskill
2) Improve at communicating the possibilities to our clients – and not get too hung up on job titles, roles and responsibilities
Invite your clients in but don’t just tell them what you’re capable of, show them. Data scientists need to be brought to the table and work alongside analysts, strategists and CX if we’re to be able to fully understand what data can do and then show this to clients.
Ultimately, I don’t think it matters too much who does the work, or what their job title is. The important thing is what we can achieve together. We don’t want to become self-limiting. The key thing is that we need to change the conversation and explain what businesses could achieve if they give us – and yes, our data scientists – the opportunity to shine.
