Martin Squires is a leader with extensive experience in customer insight, marketing analytics & data science. He has had senior roles with organisations that include: M&S Money, Walgreens Boots Alliance, HomeServe and Bradford & Bingley. Selected for the last 5 years as a member of the Data IQ Data 100, Martin has considerable experience helping organisations drive value from building a deeper understanding of their customers.
What was your route into technology, data and analytics?
I always liked maths and stats as a kid – I was then, and still am, very happy buried in a book of football statistics! I did an economics degree and then around three years in production planning and logistics types roles before starting as a junior economist for the old National and Provincial (N&P) Building Society.
“In terms of data and analytics it all started when I sneezed and slipped a disc shaving would you believe?”
It was the early 90s and I was working on the analytics and forecasting side of things at the N&P. When I hurt my back I was off work for about 6 weeks. On my return I was offered, and I accepted, a role on a project about customer data and segmentation. Four weeks later I asked if I could stay with that project because it was a lot more interesting to me than economics!
How would you define “modern” data management and the connection between that and business analytics?
It’s the classic garbage in garbage out, but it is even more important now as we move into the real drive towards machine learning and artificial intelligence. Fortunately more people today are taking data management seriously than 30 years ago – even if it is still regarded as a necessary evil by some. Particularly the data governance side is beginning to be seen as really important.
If I look back to building models 15-20 years ago, the human statistician was more in control – a decent statistician could certainly examine a model and be able to find any bad data.
“As we move into a more automated age – where things are delivered by black-box techniques, deep learning and AI – it is even more important for the data to be correct because otherwise it will impact the customer experience. If you don’t get the underlying master data right then all the AI and ML techniques will let you do is get things wrong faster!”
There is now a growing acceptance as well that cloud is not “evil” – much more acceptance of that and open source as well. Things are becoming more flexible and the challenge now is probably that there is lots of technology tools around and they have over-lapping functionality. It is difficult for people to keep track of what they are trying to deliver because of all the different technologies they have to choose from.
What are your top 3 tips or resources to share for aspiring modern data masters?
- Data strategy shouldn’t be separate from business strategy and it must be fed from that. People still tend to regard data strategy as a silver bullet – they think you just need to build a data lake and everything will be ok. For me it doesn’t work that way – know what you want to deliver as a business and then decide on the data and analytics to support that strategy.
- Keep the focus on the value you want to deliver via that business strategy. Don’t get distracted by the tools and shiny new technology toys. If simple technology will do then use simple technology.
- Keep learning. One of the things I love about what I do is that nothing stands still.
“There are always changes on the technology side but also on the use case as well. You need to stay informed and build on your knowledge and expertise and also understand what other people are doing.”
You have a lot of experience and success in analytics of all types to drive deeper customer insights – can you tell us about a time when you have engaged with business users to successfully derive and apply new insights and what the outcomes were?
One example is a project I did at Boots who are famous for 3-for-2 promotions. We had one which is a good example of something I really believe in which is that you need to get out into the stores and not just sit in front of a computer. We had a promotion of 3- for-2 on a particular type of nappies which was only working in some, but not all of the stores. To find out more we needed to make a field trip.
We went into one of our large stores in Nottingham and what we saw was a young mum with a child in a buggy. She went to the nappies, put one packet in the buggy, gave one to the child to hold and put another under her arm before paying and leaving the store to go to a bus stop. We approached her and helped her onto the bus – having shown our badges so she knew we were Boots employees! We then went for a coffee and had a chat and we realised that what we should have been considering was how someone could get 3 packets of nappies home if they didn’t have a car.
“When we went back to the office and overlaid car park location data we then had an “ah hah” moment because the promotion only worked where there was a car park nearby. So you’ve always got to factor in all sorts of things that are never going to be neatly stored in a database.”
Then think about other ways to get around the problem – may be run the promotion on line with free delivery for example.
What do you think are the essential skills to be a good analyst?
“I still thing the four biggest skills are creativity, curiosity, communication and common sense – if you’ve got those then you can be taught technical skills.”
If you haven’t got the imagination to look at the data and almost forensically solve the mystery then you won’t make the best analyst. Just like in the TV programmes like CSI where the analysts solve crimes by a combination of analysis and field work, there is no better way of understanding what is going on than going into the real-world for example visiting shops or listening into call-centre conversations.
I remember years ago I sat in a mortgage call-centre and watched the reps tab through the non-compulsory data fields. They were selecting option 1 every time because they were targeted on speed per call and not on data quality.
“You only learn that kind of thing by getting out of the office. You’ve got to build your experience and remember that if anything in the data looks too good to be true then it probably is!”
The industry makes itself seem more complicated than it is even at the recruitment stage. We put people off and make them think they need a MSc in stats or Computer Science. If you are curious about data then there are plenty of roles for you.
“Some roles do require a PhD but mostly you just need that forensic interest in solving the problem – curiosity is much more important than loads of qualifications.”
Could you also please share an example of where things have not gone so well and what you learned from the experience?
There are plenty of those and many of them relate back to the previous example of needing to understand the business context and real-life situation. For example as a young analyst at N&P I built a model for Home and Contents Insurance in order to better understand our retention strategies. I was really excited because I was looking at the key variables and found a real link between clients closing their mortgage and closing their insurance policy. I went to tell my manager all about my latest discovery and he said “well they would do because we can’t sell insurance separately so if they close the mortgage then we automatically close the insurance policy!”
“Most of the things that have gone less well have been to do with not spotting errors in the data or not spending enough time getting the brief right up front.”
You might be asked a question but you learn over the years to drill into a bit more because a request for a “simple” piece of analysis is very often not straightforward. You need to understand the context and the big picture about what is driving the request so you can decide what analysis is really going to help.
What trends or changes do you predict to the data management and analytics arena in the next few years?
I think some of the consolidation in the industry is really interesting – for example the Salesforce acquisition of Tableau – and that there is a lot more consolidation still to come. I think this is generally going to be a good thing long term in terms of simplicity even if it means a bit of short-term pain for customers – with any merger they often end up with 2-3 account managers coming to see them instead of one!
I think there will also continue to be a drive to improve BI capabilities and also their useability for non-technical people. That can be a double-edged sword because I think you still need to have a certain statistical knowledge to use some of these things.
“As NLP gets better we may be able to have a conversation with a computer and have it decide how to do the analysis we need. It would be interesting to know how far we are from the Star Trek computer!”
What do you like to do outside of work?
The stats addiction still comes through a little bit and I probably take Fantasy Football a little bit more seriously than I should do! I managed to win one of the mini-leagues I was involved in last year.
“Stick me in front of any science-fiction movie and I’m generally happy. I love sitting watching things like the Walking Dead with my kids – don’t worry they are in their twenties now so it’s fine!”
I particularly like Neil Gaiman and so I am currently binge-watching Good Omens and a favourite book is Neverwhere. It’s just a brilliant mix of fantasy and real-world locations – it’s very difficult to get on a Tube in London and think about it in the same way if you’ve read that!
What’s next for your career?
I haven’t decided yet as I’m lucky enough to be in the position to consider my next move carefully.
“If there are organisations out there who want someone that will build their team, get their data right and get s*** done then I’m open to having a chat!”