Big data – what to do with it
01 January 1970
You can also listen to these podcasts directly from the Old Mutual app, which is available here.
Hello and welcome to another edition of Old Mutual Live Business, my name is Chris Gibbons. Here’s a question, does your business need big data? It may be something you already have inside the company, but aren’t utilising to the max. Or it may be something you’re unaware of that is just sitting on your computer. Could it be that you’re not even certain what big data is? Is it something to do with this Cloud thing you keep hearing about?
Fair enough, let’s get both an introduction and an expert’s overview, Kerry Chipp is a Senior Lecturer at GIBS, one of South Africa’s leading business schools, also a member of the South African Market Research Association. Kerry, welcome to Old Mutual Live Business, thank you for joining us. Let’s start at the very beginning. When we say ‘big data’, what do we actually mean?
What is meant by Big data
Kerry Chipp: Well, the most widely accepted definition is the four V’s. It’s a lot of data. So there’s Volume of data, in a Variety of formats, coming at you at speed, Velocity. Then the next question is, what is its Voracity, how good is it. There’s a massive amount of data of all sorts of types coming at you at speed and you need to ascertain the signal from the noise.
CG: If I genuinely don’t know what the concept ‘big data’ means or I didn’t until a moment ago, how do I find out whether my company has got it or not?
KC: All companies have digitised data assets, whether they realise it or not. If you look, Discovery is a lovely example cause they used to cache, what they said, two megabytes on their customers. They now capture something like a gigabyte. Their customers always had behaviour.
Discovery has upped the ante on what part of that behavior is captured and digitised. In life, in general, regardless of what your operations are, more and more things are becoming digitised. Even if you look outside this office. There’s a security video, that is data, it’s coming in a different format. It’s not ones and zeroes, it’s a large, what they call a large visual file unit.
You now have the technology to actually analyse that. So they can see facial recognition of people that are supposed to be here, of how many people are here, how often, what’s their movement. If you take this to a shop floor, you can see what is your foot traffic.
If you take it to your manufacturing wing, you can see what happens in your processing area. You have the security camera, it’s recording data, what you’re doing with it. If you’re putting it into a pool, what’s called a ‘data pond’, you’re capturing it. If it’s just getting viewed by a security officer and then dumped after a regulatory period of time, you had data, you just don’t have it anymore.
Knowing what to keep and what to discard
CG: How do I know what is valuable and should be mined or fished out of the pond or whatever the phrase it and what should be discarded?
KC: There’s two different approaches. The first one is know what data you can capture. Know what your data pool potential is, what you can pool. Then you’ve got to have a plan on how do you manage it. A lot of people refer to this data pond as a toxic pond.
Because if you just go and aggregate everything and you don’t do anything useful with it, it’s like a flea market. There’s a whole bunch of stuff in there. There might be something valuable, but the effort that it takes to find it is ridiculous. So have a plan of managing data.
There’s a counter-argument that says, if your company is not evidence based, having lots more data to ignore is not really; if you’re not operating from an evidence based environment, what’s the point of getting more data. Because it’s more evidence.
Then the last step is what can you commercialise. So it’s capture, manage, commercialise. Commercialise is to actually make money out of this. There are two approaches, one of which I think is much easier. When Osama Fayed was here, he was the first big data officer, they say, in the world. Because he was head of Yahoo and he is now the Chief Data Officer of Barclays, London.
He was here and he was saying, you could look in this data pond, put somebody very bright in front of it and they can find value. Or you can say, what are our pressing business questions? Tell somebody very bright, find a solution to this problem.
The one is being driven by business problems and the other one is, we know there’s value. Let’s put somebody bright to extract the value. The second bit is much easier. So if at Discovery they’re saying: we want to be able to assess our risk more accurately.
If we had people’s driving behaviour as they drive, in real time, so it’s velocity, coming at you. A volume, so every single movement on a car, and there’s also some sensitivity to the phone. We can actually accurately measure our risk exposure due to you. We can’t do it from all the other cars on the road, but we can actually do it for you. It’s to solve a business problem.
Discovery actually have said that they want to be in the position that as soon as you walk in a door they can say you have a heart problem. You have this, you have that, you need this, you need that. So the data itself from your FitBit, from your phone app, from your Smart Car, from your Smart House, will tell them how to respond to you.
Handling data is very specialised
CG: Is this something that should sit on the Chief Executive’s desk or the Chief Marketing Officer or do we really need a Chief Data Officer?
KC: You need a Chief Data Officer that sit in C-suite. But you also need a Chief Executive that understands that data is a resource. If you don’t understand it as a resource, you will never unlock its value. I think any C-suite, the problem with technology in general, you’ll have like a CIO or a CTO, Chief Technology Officer.
They are seen as an infrastructural software kind of person. They’re not seen as actually strategic. Data and infrastructure is seen as an end in itself rather than as end in the business. You can’t have somebody who thinks purely in terms of infrastructure and software and what can be captured. You need somebody who is strategic.
If you look at the internet of things, where they say everything is coming online, if you ignore this, what’s going to happen? Your competitor is going to understand there’s the internet of things, where all your processes are fed into a data storage device. You can do that and lots of people do, but do something with it and then you can make money out of it.
Smaller business’ are using data more effectively
CG: Isn’t this a very expensive process?
KC: You know, we had two pieces of research that we undertook last year. One was to see the maturity model of big corporates and the other one was to look at what small business has done with big data. What was fascinating is if you ask all the corporates, they say, we have clever people. The mining bunch, they have geologists, they’ve got number crunchers. The financial corporates will say, we’ve got statisticians, we’ve got people that can do statistical modelling.
Then you ask them are you mature and they say no. Because the reason why they say they’re not mature is because they’re not making money out of it. If you look at the small companies who operate for big business, they’re making money out of it, interestingly enough. But they’re not actually doing it for themselves.
A big corporate will say, we have X problem, the small company will come along and say: okay, we can analyse your data. Give you a solution to your problem and here’s our invoice. They are actually agile, they are entrepreneurial.
Because what came out of that, there needs to be entrepreneurship. There needs to be within the corporate some push for innovation and change using data. Which is why Discovery has been so successful, because that’s part of their DNA.
CG: They are the shining example of how to use big data well?
KC: In South Africa yes, and it’s part of their DNA, it’s how they’ve grown. If you look at like Airbnb, they use data very well. If you look at the car, Uber, they use data very well, even to the point of pricing. Amazon changes its pricing over 2.5 million times per hour, so it’s completely in response to people’s searches, requests. Whether they buy stuff and you actually can even put an alert on Amazon to see if what you’re after falls within your price range, dynamic pricing.
CG: It’s not a question of should I be doing this, it’s a question of, if I’m not doing it, I’m in trouble.
KC: Yes, absolutely.
CG: Kerry Chipp, Senior Lecturer at GIBS, thank you for joining me on Old Mutual Live Business.