previous article
In his presentation at the BI Congress in Munich on 3 May 2018 organized by b.telligent, Roman Egli will reveal how his department has succeeded in generating significant added value for Swisscom TV. In an interview carried out in the run-up to the event, he discusses the factors needed to successfully embed a data DNA across a business.
Swisscom TV is one of the most innovative and user-friendly IPTV solutions in the German-speaking countries. Within a few years of entering the TV business in late 2006, Swisscom grew to become market leader with a 33% share of the Swiss market. As well as a focus on innovation and user-friendliness, the measurement of Quality of Experience (QoE) was a top priority for the business from the outset. Swisscom TV has already gained numerous awards for its market-leading user experience and performance quality. For example, users can search for content, switch channels and navigate through the menu via voice commands in Swiss-German and other Swiss dialects. Swisscom is regarded as a technology driver for TV solutions both in Switzerland and internationally.
In order to continuously measure, analyze and improve the quality of experience and end-to-end service quality throughout the Swisscom network, senior data analyst Roman Egli established and became product owner of his own technical data warehouse for Swisscom's TV division in 2007. He has worked at Swisscom TV Business Operations, where he is responsible for the TV Analytics division, since 2013. Thanks to data-driven quality assurance, optimized business processes and multi-vendor management, his department has so far managed to save the company several million Swiss francs. Not only that, but the analyses of anonymized TV usage data have proven valuable in helping the company to develop and optimize its TV products.
You created Swisscom IPTV Analytics using a small number of staff from your own department. How did you go about this and what is your success formula?
Roman Egli: In 2007, we made a virtue out of necessity when we set up "Swisscom TV Data Mart" with staff from our department. There was nothing like it in the company at the time. So we created our own DWH and chose to link it to the IPTV service right from the beginning. There was a fairly long project phase in which we defined the initial data architecture; after that, we began constructing the database piece by piece. We were also able to continually refine Swisscom TV Data Mart while we were developing the service.
We wanted to quickly locate and resolve the root causes of quality problems with the aid of TV analytics and feedback from customers. After all, we were dealing with a completely new technology on the copper-based phone network. Today, our data solution also serves as the basis for refining and developing our TV product and optimizing our data-driven customer support. We combine the technical data generated by the box with that from a number of other sources. The speed at which reports can be made available to call center agents is getting closer to real-time availability all the time.
What were the data science and big data challenges -for you and the organization as a whole?
Roman Egli: The key task of the Swisscom TV Data Mart is to improve the quality of our TV products. High customer satisfaction, quality and consistency are our top priorities. A product that doesn't work very well results in excessively high support costs and a poor NPS (Net Promoter Score). If data analysis can help us improve product quality and at the same time reduce the weekly call rate by 0.1 percent per week, we will achieve cost savings of well over CHF 1 million annually, based on 1.4 million customers, while simultaneously improving customer satisfaction. We haven't stopped tweaking these levers since Swisscom TV was launched; the leverage has since become considerable, allowing us improve quality even faster today.
What were the data science and big data challenges -for you and the organization as a whole?
Roman Egli: The key task of the Swisscom TV Data Mart is to improve the quality of our TV products. High customer satisfaction, quality and consistency are our top priorities. A product that doesn't work very well results in excessively high support costs and a poor NPS (Net Promoter Score). If data analysis can help us improve product quality and at the same time reduce the weekly call rate by 0.1 percent per week, we will achieve cost savings of well over CHF 1 million annually, based on 1.4 million customers, while simultaneously improving customer satisfaction. We haven't stopped tweaking these levers since Swisscom TV was launched; the leverage has since become considerable, allowing us improve quality even faster today.
Can you imagine implementing this project without the contribution of the Data Mart analysis?
Roman Egli: The Swisscom TV Data Mart combines data from the IPTV box with data from the software. It runs across Swisscom's entire managed network and includes the entire end-to-end supply chain. We have broken down the data silos so that we can better identify the origin of supply chain problems. This is what makes Data Mart so successful. We insisted on high quality standards at Swisscom TV from the very beginning. Without Data Markt, I don't think we would have had this success and could certainly not have grown so quickly. The success of Swisscom TV is very closely linked to Swisscom's Data Mart.
In your presentation you will reveal the "data DNA" success factors for companies. Could you give us an example of one such essential factor now?
Roman Egli: The managers at Swisscom TV believed in the project from the beginning. That was our biggest advantage. Developing a DNA and introducing it into an organization takes time. And we had that. But we were also able to demonstrate concrete successes with the Swisscom TV Data Mart right away.
Last year, you replaced Swisscom TV Data Mart with a new high-performance solution. What challenges did you have to deal with?
Roman Egli: In 2017, we migrated our database to another manufacturer. When we looked into how best to handle this kind of migration, we had two options. One was to carry out a detailed analysis of all load processes and completely redesign the data architecture. This would have eliminated unwanted legacy encumbrances and optimized the performance of the new system at a stroke. The second option was to carry out a 1:1 migration and make any necessary adjustments afterwards, which would allow us to maintain our existing procedures and architectures. As both time and costs were of the essence, we opted for the second option. This was certainly the better choice for us, as the process was more consistent with our usual iterative working methods.
What advice would you give to other data analysts thinking about building a data solution from within their own department, as you successfully managed to do?
Roman Egli: At that particular time, Swisscom had neither the infrastructure nor the know-how needed. Apart from that, the success of Swisscom TV depended on being able to quickly bring about improvements in product quality. Nowadays, it is not uncommon to find very good infrastructures that you can and should build upon. This is why I would expect to work with existing infrastructures today. However, I am strongly convinced that analytics should be driven out of the domain and that all the necessary resources should be integrated into the team as an interdisciplinary unit. Anyone wanting to deliver goal-driven, value-creating analytics must first be able to identify with the product and understand the daily problems of the market, product development and business operations. The task of IT as an enabler, in my view, is to provide the infrastructures so that the departments involved can build their own analytics capabilities.
b.telligent BI Congress 2018
Date:
Thursday, 3rd May 2018
Location:
Hilton Munich Park Hotel
Am Tucherpark 7
80538 München
Secure your tickets now:
http://www.bikongress.de/anmeldung.
Further Information about the Agenda, Speakers and Sponsors:
