Jocelyn Paulley: Hello, my name is Jocelyn Paulley and I am a Principal Associate at Wragge Lawrence Graham & Co. I am a member of our First Tech team, I'm here today with Rupert Naylor the Senior Vice President of Applied Predictors Technologies or APT, they are the world's largest cloud based predictive analytics firm. They work with over 100 leading global businesses in a range of sectors using their big data to accurately measure the impact of a wide range of business decisions. Hello Rupert.
Rupert Naylor: Hello.
Jocelyn: So what is driving customers to use analytics and predictive technology?
Rupert: That is a good question, I think what we have seen in our client bases a few years ago there was this big debate about what data have we got, we have got to have data in shareholders and they would say what is your big data strategy? We have gone away from that people now have data the question is now is what do I do with it. So really there is a lot of demand for our clients and beyond to use that data to drive insight, particularly in markets such as the UK, the States and France and Germany where it is incredibly competitive and one of your leavers to beat your competition is to understand better the environment in which you are operating through data.
Jocelyn: So it is essentially finance driven, looking for optimisations, market advantages.
Rupert: Yes certainly at the end it's got to get to the bottom line, you know it's not just pretty analytics for the sake of it and I think that visualisation of that data is important but actually if people don't see that it's going make them make better decisions then they are not going to invest in changing their decisions through that data.
Jocelyn: So the return on investment point is quite important.
Rupert: Super important, you know if you think about a business they are doing so many things. Retailers say they have got to buy the product, get the product to the shelves, they have got to price them correctly, they have got to put them in the right places and a whole lot of people involved in that process and if you just make each step of that process slightly better all the time you can actually build up quite a lot of money.
Jocelyn: Do people see the gains that they are going to make or is it a bit of challenge or some slight disbelief that just the data has the answers.
Rupert: Yes, no that's a great question. It does, you know it's a journey, you think, we have always said lets be as transparent as possible about what is going on within the software so that is one thing and then you just tend to talk to people and as you're doing each initiative you have to engage with people who are actually going to be doing the initiative. As you are analysing it you are showing them this is what the data is showing and then you have a proper discussion about that. It's not like you are living in an ivory tower and the software comes down and says the answer is 42, it looks like in these four stores something isn't quite working in the way it perhaps is in these 170 so let's look into that. It looks like these are the drivers and then you engage in that sort of discussion with you know real things for real people that they can see.
Jocelyn: Is a lot of the skill in using this and getting the most out of it is actually asking the right questions because you say the software doesn't magically produce answers or information.
Rupert: Yes, that is a good point as well and it is part of the discipline of using your data and analytics is say what am I looking for. If you just sort of go fishing and try and do some sort of correlations well that's fine but actually what you, one of the differences we bring to our clients is, what is it you're wanting to answer, just to say it in a very simple one sentence ten words, what is it you are trying to answer about a particular initiative just we believe this initiative drives sales and does not result in the decline in transactions or something. Fine well you can then test that and that is where we come in we help our client by moving from the vague well let's have a look to being very specific and then analysing that.
Jocelyn: Although there has been movement if they are talking about big data, talking about right sized data.
Rupert: Yes, yes.
Jocelyn: So I can imagine companies think they must need to have lots of data to look at, is there also a way in which to identify the relevant bits?
Rupert: Well that's, yes and certainly the way we think about it is that the set back and the ability to process data is driven by a smaller source and that is growing very quickly over 18 months, the ability to store data is driven by Kryder's law which is growing faster than that, it's one of the huge growths faster than that so you actually store in more data than you are able to process. Now the question is then, I have got all this data which data is actually the useful bits and if you think about the data sets, say one of our client's transaction logs everything goes through the tills and through the online sales engine say, you know maybe that's 50 billion rows of data a day but if you think of putting a CCTV camera in a store, the footage over several months and then over a 1000 stores that's really big so which part of it is actually useful and which part of it is not. Which is just describing where people are going but it's actually giving you something you can use. So we talk about kind of big data, which is you know very big data sets but not necessarily the massive data sets of structured data were very large data that are really where people are generating business value at the moment.
Jocelyn: Okay.
Rupert: But over time those other data stats will, you will be able to derive structured data from those other data sets and build them into what you are doing in an automated way.
Jocelyn: Okay. So is there a challenge as well around making sure the data you are using is accurate and giving you the correct picture. Are people having to change their underlying process and the way they are working to actually catch the data so in a way that is then helpful for this kind of analytics?
Rupert: Good question yes, I mean certainly we spend a lot of time checking that the data we are getting is accurate and ties up and you know does the total equal all the bits from a very granular way, but you know, the other day we were working with a client around a neighbour programme, they wanted to know where they should put more people in stores and where they should take people out, over staffed and it turned out, you know, we said well lets loads up the labour log into the system, we'd assumed that they would have, who was working on a particular day, what their level of experience was, you know, what hours, when they were clocking on, clocking off, that sort of thing. And they didn't. They could tell us by week how many people had worked you know, that week, total. They couldn't tell us their level but they could tell the number of hours worked, so that's all the data they have and so with that we can do a lot, but we said to them at the same time, well look now, why don't you start capturing all this on the data, don't worry that you haven't got it any more but you're realising that a lot of things that you may've deleted you know, over written files and so on. You no longer have to do that, you can keep all that data, so it's sort of helping people understand that to be a little bit I guess more forward thinking about the day to day, they don't always know what's useful and not useful at this stage.
Jocelyn: And how do you deal with issues around personal data? I mean there you're mentioning employees and who's working where at what time and I think a lot of people as soon as you mention big data, the privacy alarm rings, certainly in a lawyers head. So how do you work with those kinds of issues?
Rupert: Yea, no, it's a great point, privacy is a very important thing. Now we get round it by not having any personally identifiable data in our systems. For us we don't need it because we're talking about what you can see from the aggregation of how much data's fed and how that can... what that can imply about individuals, but not... it doesn't really matter that a particular individual is doing something verses another person. So, we... for example on our employee data set everything would be anonymised and the programme that people have to anonymise on that basis, it's well thought through to make it non-traceable backwards...
Jocelyn: Yes.
Rupert: ...and when we load a, say a transaction on there's no way that it can be traced back to an individual and that way we get round it but it is difficult but in Germany for example, even that isn't always enough and you have to do extra things to ensure that the data is in no way traceable to an individual.
Jocelyn Paulley: Yes so when you take data from different countries as well that brings its own problems as well...
Rupert Naylor: Yes
Jocelyn Paulley: You say with additional requirements
Rupert Naylor: That's true.
Jocelyn Paulley: ..those countries may have. So I think my final question was around I know you've been acquired by MasterCard recently, does that open up new possibility for what you as a business might do?
Rupert Naylor: Yes, well it's very exciting, I think what I hadn't realised is the size of the data set that MasterCard has because we already have in the US actually have 1 in 4 retail transactions are in our system so pretty big but MasterCard have 2.2 billion cardholders, debit cards, credit cards operating, they have this incredible data set about what people are spending on and where they're going to spend and that sort of thing and that's very valuable...anonymised obviously and MasterCard's like us, very strict around data privacy...anonymised are an incredibly rich source of information around spending habits and how initiatives may be a business for taking, are changing the way they are spending their time shopping and changing what they are spending their money on.
Jocelyn Paulley: Sounds very fascinating
Rupert Naylor: Yes, yes, its early days but it's been a very exciting set up.
Jocelyn Paulley: Excellent. Well thank you very much.
Rupert Naylor: Thank you.
Jocelyn Paulley: Thank you for listening, I hope you've found that interesting and useful and if you have any more questions about anything to do with big data please don't hesitate to get in contact with any of our tech team.