How to Mine Gold: Handling the Communication Channel Shift to Social Media
Data being the gold of the 21st century is a given fact by now. Still, most companies do not have a strategy on how to handle social media data, even though it has become the main channel of client communication. To tackle this problem, we recently carried out a ti&m garage with a major Swiss bank.
Data being the gold of the 21st century is a given fact by now. Still, most companies do not have a strategy on how to handle social media data, even though it has become the main channel of client communication. To tackle this problem, we recently carried out a ti&m garage with a major Swiss bank. The vision of the project was to initiate an on-premise social media data archive and monitoring tool as financial regulations ban any cloud based solution (no client data must be stored cross border).
For the garage, we focused on three areas: data archive, customer care / monitoring and alerting. The technological foundation of the project was the ti&m Analytics Suite and the Analytics Dashboard, which made it possible to start storing data and visualize results within the first two days. The ti&m Analytics Suite is our big data approach, it is built on the Hortonworks Data Platform using the Lambda Architecture, the industry best practice for big data analytics.
Aggregating social media data and storing it was complex. Each network is different in the way data is published - completely public as seen on Twitter or completely private like Facebook’s user timelines or direct messages. Each network has to be connected individually through its API and the consumed raw data has to be transformed into a common data model for further processing. This can be simplified by using data aggregators like GNIP, but unfortunately, they only work with large ventures and 10 million dialogs per month is the required minimum.
The main reason behind our decision to develop our own solution rather than using existing monitoring tools, was the ability to archive raw data. As mentioned earlier, data and specially meta data is a very valuable asset. On the senior management level, the awareness to archive such information is there, however, the willingness to act is small as the current gain seems marginal.
Unfortunately, the collected information only starts to become valuable after a longer period of time, leading to a hen and egg problem.
Our answer to this is to embrace a small investment early on instead of counting on a possible but much larger investment involving the buying of historical data in the future. There are also immediate benefits! The monitoring functionality and know-how sharing within a team increases the quality of customer care and community management. On top of that, social media alerting is an additional, very interesting use case.
Social media reactions on outages have shown to be faster and more accurate than internal monitoring. To make use of this, we implemented a trigger that sends emails and notifications to internal systems based on tweets. Additionally, recording social media data on outages and incidents is also an important information source during the aftermath of such events.
While history has shown that communication channels change and evolve constantly, customer dialogues have always been an important data source. Handling social media data is more challenging as it includes third parties, most of which are cross country, very diverse legal issues such as data privacy, financial regulations and third party terms and agreements. Within six weeks we delivered a system able to aggregate data from Twitter, Facebook and Instagram feeds, connect and enrich user profiles as well as analyse and visualize the recorded data on the ti&m Analytics Dashboard.
If you want to take a closer look, let me know in the comments or send a message to firstname.lastname@example.org.
Nachdem das Jahr 2015 aus unserer Sicht als das Jahr des digitalen Portemonnaies zu Buche schlug, überlegten wir uns, wie wir die Akzeptanz des digitalen, bargeld- und kartenlosen Bezahlens an öffentlichen Verkaufspunkten fördern könnten. Geboren war die Idee einer virtuellen Registrierkasse, welche jegliche Art von digitalen Geldbörsen unterstützen würde.Mehr erfahren
Small businesses need to change their general big data approach. The question they should ask themselves is not “what can we gain from the collected data?” but “what do we need to grow the business and how can data analytics help reach that goal?”.Mehr erfahren