From a railway tweet to a social analytics roadmap

3 July 2014

Jan van Ansem

Jan van Ansem

Principal consultant SAP Data Warehousing

It is Monday morning at Paddington station during rush hour, and I am trying to get on the train with my folding bike, two large bags and a cup of tea. I know it is going to be busy, so I’ve given myself some time so I can get on the train, secure a seat and store my luggage before everyone else. 

The information display confirms the train is on time, but doesn’t tell me which platform. This is annoying, because the platform might be far away, which means by the time it is  announced, I might have to rush across the entire station, along with hundreds of other people trying to catch the same train and thousands of others going different ways. The prospect of finding a seat is diminishing with every minute I wait. I then remember that when I checked my journey earlier on the phone, the platform was specified. . So I get my phone out, tap in my destination and there it is: Platform 8. I make my way to the platform, enter an empty carriage and sit down with a satisfied grin on my face.

And then I tweet. I ask First Great Western why my mobile app knows where I should go but the information display doesn’t? Do they need help with managing their information processes, I ask wittingly, as I am after all an ‘Information Manager’ by trade.

What happens next exceeds my wildest expectations. Within seconds I have an interesting response from First Great Western. A few tweet exchanges later I have a satisfying answer, which I’ll come to further down. This fast, high quality response makes me feel valued as a customer (they listen to me) and I think better of them (clearly they have efficient processes in place for dealing with inquiries, so other processes might run as smoothly too).

Social analytics

I was intrigued by the whole experience and as I have a professional interest in data management I decided to update my knowledge about ‘social analytics’.  Several years ago I was first introduced to unstructured text processing engines and I was blown away by what I saw.  What I saw then was a bunch of emails which were fed into the engine and as a result an overview was produced showing the sentiment of different users for various products. Back then, a lot of research had gone into developing the engines so they could recognise entities (customers, products) and emotions. Even then they worked well enough to impress me. The difficulty back then was that tuning these engines for a specific data set and for a specific use case. Specialist knowledge about the workings of the engines was required to get useful results. This obviously greatly reduced user adoption. If you would only allow people to drive a car if they can assemble the engine themselves, it would be very quiet on our roads.

What strikes me know is that compared to five years ago the actual engines might perhaps not have improved so much, but the ways to apply an engine to unstructured has vastly improved. Social media analytics has started to become a commodity product for which you choose from a wide range of competitive tools. In addition to this, some functionality is now also often embedded in traditional toolsets, specifically in Customer Relationship Management tools and content management systems.

The roadmap to social analytics

It is clear to most that businesses engaging effectively with social media have a competitive edge. By using social analytics you reach customers they would not reach otherwise and gain a better understanding of existing customers; it helps to strengthen brand reputation and provide insights that could not be gleamed otherwise. Although Gartner predicted in July 2013 that through 2015 85% of Fortune 500 organisations will not be able to exploit ‘big data’ for competitive advantage, most people will agree that this figure will decrease significantly in the next few years (Gartner Hype Cycle for Content and Social Analytics, July 2013)

Organisations should at least be able to provide a roadmap for social analytics otherwise the marketing departments, customer services teams and product developers will come up with their own point solution. Stakeholders need to be asked about their priorities for social analytics. A brand manager might be ready to start with brand reputation analysis but the sales manager might only just start to think about engaging with customers through social media. There are different use case scenarios and these will require different functions within the same tool or different tools. Sooner or later there will be a need to pair up social analytics results with the ‘back office’ so it is wise to put that on the roadmap as well and maybe that ability should weight more than the ability to produce out of the box pretty graphs about usage stats. Developing a roadmap will be a significant investment, but it is better to spend the money on this now than ending up with disparate tools as point solutions for isolated use cases.

Twitter

Now you know you should have a social analytics roadmap, but one important question is still not answered: Why can information displays at the station not tell you which platform to go to, while this information is available on a mobile app? Well, the first thing I learned from my Twitter feed was that the information screens at Paddington are not controlled by First Great Western (@FGW) but by Network Rail (@NetworkRailPAD). The latter explained that some mobile apps show the ‘booked platform’ rather than the actual platforms. Why the information screens still not show the actual platform when the train is there, ready for me to board, was something @NetworkRailPAD did not explain. But perhaps they have state-of-the art social analytics so they see this blog and leave an answer on this very page?

 

View comments

Comments

Blog post currently doesn't have any comments.

Bluefin and SAP S/4HANA - welcome to the one horse race