BlogsRSS

Data is now big and analytics has become advanced

As featured in DisruptiveViews

“People are utilizing huge amounts of data and massive processing power as if they don’t have huge amounts of data and massive processing power,” said Prudencio Pedrosa, associate partner in McKinsey & Company’s Madrid office and core member of the Advance Analytics Practice during a presentation at WeDo Technologies WDC event in Miami last week.

“They are still using the traditional regression they learnt in school. What’s the use of having a fast car and driving at 40kph?”

But how do we change the way we are working today to make the best of use of all that data, all that computing power and cloud services?

Firstly, according to Pedrosa, we have to stop thinking of advanced analytics as some form of magic. We hire extremely bright guys from university that have studied for years and we assume they can do magic with the data. That’s a myth and completely wrong. If you are going to use advanced analytics you need to know what you want to know.

Second, what things do you have to actually predict with? Advanced analytics is about using data to predict something. If that something has never happened before then what’s the point of using advanced analytics – you’re probably better off with a crystal ball.

Thirdly, you need to predict something that is useful. Sometimes data scientists spend four months working on something but when it is presented to the business they say things like – ‘I’m sure this is great but I don’t know how to use it!’

Advanced analytics is about making the right business decisions. We are becoming obsessed with having real-time information about monitoring customer activity and making them an offer that we think they will be attracted to - but do you really need terabytes of data to achieve this?

Business owners should be worried about the impact of analytics and not about the process – they should leave that to the data scientists. And we shouldn’t assume that one data scientist has all the skills required. You need people that know about the business, they need to be able to abstract value for data, and they need to be able to use mathematics in an extremely sophisticated way. That means you have three different profiles and the chances of finding all of them and one person are pretty slim.

Fantastic analytical models alone are useless. If you don’t provide impact by changing the way the business operates then the exercise is wasted. When we’re talking about advanced analytics the least important things are the algorithms and the machine learning. The most important things are workflow integration and change management. Many businesses are simply unwilling to change the way they operate.

No project succeeds without having people on board, especially the last mile. The whole company has to be aligned and there is no guarantee it will work first time. You have to be prepared to fail. The data collected at this phase can then be analyzed to improve the next model.

While advanced analytics has existed for over 20 years, big data has accelerated interest in the market and its position in the business. Rather than being the domain of a few select groups (for example, marketing, risk), many more business functions now have a legitimate interest in this capability to help foster better decision making and improved business outcomes.

Whilst everything Pedrosa put forward made sense one gets the feeling that even McKinsey has problems convincing businesses to put their faith, and their future, in the hands of advanced analytics.

In the case of telcos, radical change is an anathema and moving from tried and tested processes and business models must be proving quite a challenge. Everyone is talking about it but how many have managed to get buy-in from all parts of the company – a key element to making it a success?