What’s your background? How did you come to be working in analytics?
I came from an accounting background and then moved into retail where I worked across a variety of teams in finance, marketing analytics and strategy for various different retailers. I then left that to join a consultancy working with loyalty card data. I really got into analytics in that role, and from there I moved across into the gambling industry which of course is very rich in terms of customer data and analytics. I’m now Head of Yield Management and Analytics, leading a team of eight people doing the analytics for the retail and the multichannel arm of Ladbrokes Coral.
What does that role involve?
It’s really about using available transactional and customer data to inform business performance and strategic initiative as well as day to day yield management, to ensure that we’re putting the data at the heart of decision making on a day to day basis. The team is involved in all commercial and trading aspects of the business.
An example of that is that we’ve got nearly 4,000 shops in the UK. It’s one of the biggest retail estates in the country, so there’s a lot you can do in terms of test and learn and trying different things and tweaking a few things. The time-to-market when making decisions is incredibly short which means you can measure things very quickly and make an informed decision based on statistically representative test and control groups and then roll out from there.
Can you give me an example on how that might work in practice?
Yes, it could be things like scheduling races and other events that we manage within our shops, so it’s all about understanding which customers might be in our shops at particular times, and then which of those events would give us the biggest bang for our buck. Another example might be deciding what hours our shops should open for.
What kind of challenges do organisations like Ladbrokes face today?
On the analytics side we use the data a lot to try to model the impact of some very high profile regulatory changes are coming our way. We’ve been leading the dialogue with the government and Industry bodies on the impact of some of those changes, and that’s been a very data-led exercise using combination of transactional data and customer research data and putting it all together.
It’s very competitive market. Customers are very promiscuous. It’s a commodotised product, if you want to bet on Liverpool versus Man United you can do that anywhere, so we’re using our data to understand how we can be sure that we’re front of customers’ minds when it comes to making those decisions.
Analytics can help us address these challenges by ensuring that we make informed decisions based on data rather than gut feel.
On a day-to-day basis what kind of analytics tools and techniques are you using?
Any analytics task is always about what you’re putting in to your model, as well as about understanding what the business challenge is and what’s within the remit of the business to change. The starting point for the data analyst is firstly to understand the business question being asked, and then to understand what data do I have that could enable me to help the business make the best decision?”
One of the key challenges for us is that we have several different databases across our different businesses and brands. We’ve got retail and online, then we might have some other datasets around wages or shop catchment as well as demographic information, and the ability to blend all that together is one of the key things that people in analytics have to be good at. Tools like SPSS Modeler we find incredibly efficient for doing that, rather than writing traditional database code which is more cumbersome and doesn’t allow you to as easily merge data sets across completely different data warehouses that may exist in different physical locations.
In terms of the modelling output, it’s critical to keep it commercial and put it in business context. I’ve seen lots of examples of very smart data analysts with PhDs who are incredibly intelligent but approach modelling as a purely theoretical exercise. Doing it this way means that sometimes it can take so long to hone or develop a model, that by the time it’s ready it’s no longer timely so can’t be used by the business, or the model can’t be applied perfectly in real life in terms of what it actually means in the commercial environment.
How have you seen the analytics space change during the course of your career?
I think where there’s a lot of innovation at the moment is on the visualisation side of things, trying to make data more accessible to non-analytical people. It’s quite exciting what some of the visualisation tools like Tableau and Power BI can do.
There’s also been an explosion in data volumes over recent years – we have much more data now that we did a few years ago. The challenge is, for analytics, how you can make all that data accessible if it exists in different formats and different databases.
Have you got a project that you’ve worked on that you particularly enjoyed or found particularly interesting or rewarding?
Any project where the business has been engaged with it and ready to take on the recommendations is always rewarding. In my previous role, I was working with a large retailer in the United States to build their customer segmentation into the business, to enable them to use it consistently in terms of price, promotion, range and shelf base arranging. That was really exciting as they were really up for it, in terms of being willing to pay the money to build these things in the first place and then the readiness to the business to try to work with the output and do it consistently. We helped them design an entirely new smaller shop format, using the customer segmentation to ensure they could get exactly the right range in those shops.
If you had an unlimited budget and you could do anything you wanted to, are there other opportunities you could see in the organisation, ways you could be getting more value out of analytics?
Budgets always seem tight in retailers in terms of properly embedding analytics across the whole business. Ideally if you were starting a retail business from scratch, you’d make sure that your shop and your call centres, and your head office and any other warehouse or distribution systems that you might have, are all on the same platform and able to talk to each other in real time and have a closed feedback loop so that if somebody has a conversation with a colleague in a shop or with the customer service or something happens in the supply chain then that is immediately fed into that closed loop system decisions are made automatically off the back of that. That’s something I’ve not ever seen a business do quite properly, largely because of disparate teams and systems in different locations, as well as the lack of budget needed to do what would be ideal.
What advice would you give to people who are thinking about entering analytics?
It’s definitely an exciting area that’s continuing to grow and the business demand is still there. Predictive analytics is only getting bigger and more important for businesses. In terms of advice, I think it’s important that you really understand what the business is doing. Take the time to have the experience as a customer with the company you’re working with and others in the same industry. Make sure that you really understand the business you are working with, how decisions are made and who the key stakeholders are, rather than just sitting in front of the very exciting data sets. Get in front of them, show them how analytics and data can be used with clear simple examples. Even if the modelling itself is complex ,it is really important to keep it simple in terms of what it’s actually doing and how it can be applied in the business.