How important is speed of insight when it comes to analytics?
People are jumping into technology and big data solutions without knowing the business problem. That will define what is the nature of real time or what kind of speed of response or insight they need.
There are different levels. You could have a website where consumers come in and you have a recommendation engine responding in real time to create a profile of what the user wants.
There’s also shades of grey based on real time, near real time, to things that aren’t real time but that have a lot of complexity.
Most big data is unstructured and so harder to analyse. What tolerance for inaccurate data should companies have?
Anything is an improvement on today, if people are not doing anything then that’s an issue. But with better insight you are going to improve the business. It’s really as simple as that. People are using very static batch data or subsets of data rather than using the full pile of what is there.
There are solutions for people with no skills, low skills and high skills. High skills are data scientists who need the deep, complex capabilities. But depending on how deep you want your insight to be there is a solution for you.
When firms can’t see all their data it creates the problem of dark data. How much of a burden is this ‘invisible’ information?
Dark data is nothing but latent data that people haven’t had the time or energy or inclination or to do it in a cost-effective way. Storing it, managing it, curating and leveraging it takes time and effort. But that’s where some of the simplification we are doing with the HANA platform tackles that complexity and using things like predictive analytics allows you to leverage more of that data.
How relevant can dark data still be if it’s got quite old?
It can be. You don’t know what insights you have missed if you don’t do something with it and try to exploit that information. There could be huge opportunity or at the very least you find out that type of data you are looking at is not valuable. Knowing what value or not you have from it is valuable to optimising the efficiency of the business.
What is SAP doing to take advantage of the Internet of Things?
A large part of the IoT is machine to machine (M2M) and that’s where you do need predictive and analytics capabilities. What you’re seeing is we’re working with communication vendor that is trying to get their operators to make sense of one petabyte of data per day. M2m is your Playstations, and things like your communications devices, it’s becoming more and more the case that people don’t know how to mine that data.
Instead of doing it reactively, we want to proactively manage these things. M2M is a huge problem. It’s all the digital marketing going on, those are all real time interactions with users.
How will companies’ responsibilities to customer privacy change with so many data points on those customers becoming available?
I subscribe to a lot of applications they all ask do you want to make yourself aware locationally. I have that choice and companies need to make it clear it’s a choice of whether you give some of your privacy for the benefits of a certain application.
Where the world is going to is really personalised one on one marketing and really knowing me as a customer and not putting me in a certain age category living in a certain location. Instead it knows my behaviour, my likes and my dislikes: firms need to go beyond segmentation. Businesses need to use data at a level of granularity beyond simply segments of data: what age bracket a customer falls into. Real time isn’t just something you’ve created – maybe I get married: suddenly my profile and preferences will change, and at that point in time it’s important to know how I’m adapting to something that’s different.
