Data Science

Meet the experts who take data science beyond the buzzwords   

Karsten Shaw

And find out why data science really is important for your business

Data science, machine learning and AI. The world of analytics is full of technical buzzwords, but we mustn’t assume that everyone knows what these terms really mean, especially now they’ve found their way into the language of brand and customer insight. We asked one of our resident data science experts to shed some light on why businesses really do need a deeper understanding of data and how our data science and analytics team is set up to support it.

Karsten Shaw is Analytics Partner at Yonder Consulting. Over the past six years, Karsten and his colleagues have built a team that’s truly embedded at the core of Yonder, which allows them to work seamlessly with other teams to get to grips with what clients really need. Gone are the days where data scientists and analysts were siloed. Instead, Karsten and the analytics team are set up to support and collaborate with partners across the business.

Insight is the fundamental starting point for all strategic action, both internal and external, and a deep understanding of data is where insight truly begins. A deep understanding of data means data science, machine learning and AI (to name a few of those buzzwords) actually start to make sense. Our team of analysts take the most complex insight, garnered from customers, and translate it into simple data that can inform business and brand strategy.

Throughout this series, we will be translating complex analytics and explaining how it impacts business. Our first conversation is with Karsten Shaw, so we’re going to start at the beginning by asking, how would you describe data science?  

Karsten: There are huge amounts of data available, and it ranges from survey results to the kind of complex unstructured data that record natural speech or images. We work with many different clients across many different sectors, from charities to tech to finance in any given day, and there’s a glut of data to work with. The way that organisations are now run means they’re collecting more and more data all the time, through websites, loyalty cards, you name it. Many will even have data on you that you don’t know about. Data science is what allows us to process and analyse these many types and sizes of data, so they become useful. But that’s not all, we also need to understand this data within the context of a client’s business question.

This process has three pillars: statistical knowledge, programming skills and business understanding. Put another way, you should understand the business or organisational challenge before you even start the analysis, and you need the technical skills to look at all types of data that may be useful. Finally, you need to have a good and broad understanding of statistical models to choose the right one for the type of data you use, and the business objective you have. In combining all three skill sets, we see the whole picture and gain the flexibility to tailor our approach to different client needs.

As overwhelming as this sounds, clients are in safe hands with Karsten and his team when it comes to deciphering all this data; Yonder has built a team where everyone brings something slightly different to the table – be that different backgrounds, interests or strengths, this diversity creates a winning combination that’s essentially the secret sauce of a high performing data science team.  

Karsten – who has many years of working for big market research companies under his belt – is also a member of leadership groups for the Market Research Society and Royal Statistical Society. He shares all this experience with new generations of analysts as a lecturer at the London School of Economics.  

Karsten: You really can’t do this job without having an innate sense of curiosity. We have such a diverse team with skills that are so different but complement each other well. These range from segmentations expertise to text AI and machine learning, which means the team can choose the right method for the client’s needs and tailor the approach as needed.  It’s because we have such different backgrounds and skills that we’ve worked hard to create a culture which feels as safe as it is challenging and engaging for the whole team. Anyone can ask any question and know that it’s not silly – asking questions is how we innovate. We really encourage the team to try new methodologies so we can keep getting better at delivering what customers need. 

We do this by not limiting ourselves to only using certain techniques or products. We don’t look at a client’s problem through the lens of a specific data tool, we think about that problem in depth before we decide which tool to use. And we certainly don’t restrict ourselves to pre-existing knowledge or tools. Constant learning amidst continuous evolution is at the core of what we do. Data tools are developing all the time, so you must keep learning to keep up. Plus, we’re always learning from each other, from our clients, and from colleagues across Yonder.   

Data insight is integral to Yonder, but data science and analytics seems particularly specific, so where does your team fit into what Yonder do for their clients?

Karsten: Everywhere! If there’s a client project, we become fully integrated as part of the project team. This isn’t really something you’d expect an analytics team to do, but we prefer to be involved in all stages. We like talking to clients and hopefully they like talking to us! It’s really important that when clients come to us, we don’t just understand what they need, but make sure they understand what they’re asking for too, because we work in a collaborative way. Clients may ask for one thing, but we might reshape their ask to make it more impactful and applicable, which is precisely why they come to us. They know we’ll get to the bottom of the project, especially if it requires thinking outside of the box. It also happens that less complicated thinking is sometimes needed. Someone can ask us for a snazzy bit of analysis or some complex data, and we’ll look at the problem and simplify the whole thing instead.

There’s clearly no one-size-fits-all approach to analytics, especially in what sounds like an ever-changing landscape. So, with this in mind, is there any one thing that customers need to be aware of right now when it comes to the world of data?

Karsten: If you asked ten people in our line of work, you’d probably get twenty answers. Things are continuously evolving, but one of the major developments for us has been the ability to integrate survey data with other sources of data. We have made advances in understanding our client’s internal data structures then layering this with our own tools and techniques.

Market research or survey analytics tell us ‘why’ people behave in certain ways, being able to link that to ‘who’ and ‘what’ they do is key.

We did this for a telecoms firm recently. We combined their customer survey results with their customer database, which meant we could tap into customer behaviours like spend and usage, and then link all of this to other crucial information, such as how important privacy is to them and how tech savvy they are. Combining so many layers of data multiplies the impact of what we do, and the importance it holds for the client. It’s a rapid evolution in terms of what we can deliver for our clients. 

Next time, we’ll put the spotlight on a special award-winning project with one of the largest healthcare charities in the UK…

In the meantime, take a look at some of our case studies here, or contact our teams for more information.