Lorna – Can you start just by telling me a little bit about your background and how you came to be where you are now in your career?
Bethan – I started out doing a maths degree at Lancaster University. And then I did a master’s in operational research, which was data science really, just before it was cool enough to be known as data science! As part of my dissertation I read a book by Malcolm Gladwell which talked a lot about marketing and statistics and working out the influences within networks and everything like that. Reading that got me looking into the use of statistics in marketing. Then I went into market research straight from Lancaster. I’ve always been working in the analytics side of market research so then worked for a few research agencies. Now I’m working at Boxclever as part of their analytics division.
Lorna – So what’s your role at Boxclever?
Bethan – I’m Analytics Director working with the analytics team here. We’re a very integrated team, we all get involved in each of the various parts of a project. I get involved with all stages of the process from speaking to clients, submitting proposals and checking briefs right through to doing the work such as designing and setting up the project, analysing the data, and finally debriefing that back to the client and helping them immerse that insight back into the business. A large part of my role is the more complex, statistical analytical side of it – getting my hands dirty with the data myself. I’m always sat there with some sort of program with a ridiculous amount of data in front of me. Normally cursing at the screen!
Lorna – What appeals to you about the process of analytics?
Bethan – I’ve always enjoyed playing with numbers and being a bit of a geek in that respect! Doing that in the research industry means that I’ll get involved in client conversations and help answer stakeholder questions, working also on the softer side of things which I enjoy. I also like working out the answer to a problem after you’ve worked out what the problem itself is, and then being able to pass that insight back to people and help them work out what to do with it. I like working out the uncertainty, getting my hands dirty with the data, finding an answer and watching that answer transform some aspect of the real world.
Lorna – Does Boxclever focus on any particular kind of client?
Bethan – We don’t restrict ourselves to one industry or size of client or anything like that so we’ve got quite a varied set of clients and that is probably one of my favourite things about the job. It means that you get a lot of exposure to a lot of different areas and it keeps things interesting. That said, I do a lot of work in retail in particular. We also do a lot of work in utilities, and I’ve done a lot of work in leisure and tourism. So, it’s a huge breadth of clients and industries. It also means that you get to work for huge global organisations, and experience the opportunities and challenges that brings, and then also get the experience of working for much smaller organisations perhaps on their very first research project, and see the benefits and the challenges that come with that as well. I quite like the variety, I guess!
Lorna – Do you perhaps see the benefits of the research more quickly with a smaller client?
Bethan – I don’t know if that’s necessarily true. I think it depends on the types of people working in the organisation. Obviously with larger clients, you tend to find a bit more bureaucracy and paperwork as well as a lot more stakeholders to manage. But sometimes with the smaller clients there needs to be more steps in the process to help you understand what’s happening and what to do with it, and you might be working with somebody who doesn’t even really know what research is. But I suppose the results and the changes are sometimes easier to implement, because you don’t have that red tape and those processes to follow.
Lorna – Do you see common kinds of analytics problems being addressed across different industries and different sizes?
Bethan – Yes. A very common question that we are asked very broadly is “We need to understand our customers – we need to know who we’re talking to who we’re selling to. What can analytics tell us?” Another one that’s becoming a lot more frequent now is around value for money and what that means now to the consumer, given the cost of living increases and everything else is happening. Companies want to know what they should do about that, for example whether they should change their pricing strategies or otherwise navigate that challenge. That’s now a common question that we’re being asked.
We also get a lot of questions around branding and how organisations can keep their branding relevant, especially given things like a greater focus on sustainability and social responsibility and stuff like that, because organisations want to make sure that they’re doing something that’s important to people but without it seeming far-fetched for their brands to be playing in that space. They want to understand whether consumers will actually see and appreciate what they’re doing or whether they’ll expect organisations to be doing something different.
Lorna – And what’s the quality of the data you’re dealing with generally? Do people have the data they need to be able to answer those types of questions?
Bethan – We do sometimes run into data quality issues, especially if we’re working with internal data. We normally collect primary data that’s specifically designed to answer particular questions when the internal data does not do the job. Obviously then I’d like to say that the data is good quality because generally I’ve helped to design
ed the questionnaire! When you start working with internal data you do tend to run into problems. Sometimes internal data is a bit patchier or you can run into issues when you’re looking at a legacy data system and trying to understand what is in there. In those situations, it can be very useful to have some primary data to help alleviate some of the problems. With primary data you know what you’re dealing with whereas internal data sometimes is a bit more of an unknown quantity.
Lorna – Can you talk a little bit about the tools that you use?
Bethan – Yes – we use a real range of different tools. Obviously, we use SPSS but then we also use Q for a lot of cross tabbing and stuff like that. We use R for data exploration and modeling. I think that Excel is really overlooked as an analysis tool. You know, people don’t really think of it as that but it’s actually pretty powerful, particularly if you just want something simple and quick. It’s also easy for clients and stakeholders to understand because they’re familiar with it. That can be quite powerful as well sometimes.
Lorna – What sort of things do you use SPSS for?
Bethan – We use SPSS a lot for things like segmentation, particularly in terms of creating the clusters and checking them and running discriminant analysis and things like that. SPSS is also useful for data manipulation before you even begin the analysis or for cleaning up the data and doing some ‘data stacking’ if you want to make sure your data is in the right shape before you start.
Lorna – Are there particular projects that you’ve worked on that were particularly rewarding or interesting?
Bethan – I really like segmentation projects. I like that there isn’t a definitive answer. You run your analysis, you get several different solutions, and you have to qualitatively work out which one you like, which one has the most commercial value, which one is going to be most useful. You might run some qualitative research to bring some life to the segments. These are always quite chunky projects and often generate a lot of excitement from internal stakeholders, which I really like, and they tend to be the kinds of projects that can make quite a big difference to a client internally, which is very cool.
Lorna – Do you have any advice that you would give for people who are thinking about a career in analytics?
Bethan – Yes – the best thing that somebody can do is really just to get their hands dirty. I realise now that when I was at university a lot of the datasets that I was using were almost perfect. Then when you come into the real world you realise that actually human beings aren’t as black and white as that and you’ll never really be working with data that clean or that complete. The only way to really learn how to deal with those sorts of challenges is to just jump in and do it. These days you can actually go online and download somebody’s dataset and play with it. You can even download their code to rerun the analysis yourself step by step if you want to. Then you can rerun it on a different data set and make sure that you know how it all works. It’s not until that you start doing that, writing the code yourself and coming across those common errors and getting frustrated and solving your own problems that you really start to learn. So just jump straight in, get your hands on a data set and start doing something with it!