Deutsche Telekom: Creating a Customer Portal Using PaaS

Deutsche Telekom is one of the world’s leading integrated telecommunications companies with over 236,000 employees and 2011 revenue exceeding €58 billion ($77 billion). Headquartered in Bonn, Germany, Deutsche Telekom provides fixed-network/broadband, mobile communications, Internet, and IPTV products and services for consumers, and information and communication technology solutions for business and corporate customers.

Learn more about Accenture's Cloud Services.

Deutsche Telekom wanted to create a portal that would inspire potential customers with travel destination ideas and connect them with travel providers. The goal of the website, called Tripdiscover, was to offer visual, innovative navigation options powered by an intelligent, learning database that generates user-relevant recommendations. Filter and search functions, combined with high-quality image content and user analytics, are used to present targeted travel ideas to customers.

Key objectives of the solution included:

• Inspire users through professional, high-quality picture content.

• Recommend relevant travel destinations.

• Enable a unique user experience by means of an innovative and intuitive user interface.

• Generate recommendations based on posting of travel wishes and cross postings.

• Enable user-generated content to keep the page interactive and fresh.

• Connect multiple channels, business partners and affiliates on one platform.

Accenture was able to offer the precise skills needed to design and implement the first phase of the portal within three months. This fast timeline was made possible by leveraging an agile team structure: a development team with Google App Engine skills in Accenture’s India Delivery Center was led by a local team lead and architect. Agile methodologies and corresponding cloud technologies were used to deliver features with the highest priority and value iteratively in seven sprints. A second release involved the Accenture Multimedia Agency to shape the solution into an eye-catcher.

The architecture of the portal included multiple Google technologies, including App Engine, Cloud Storage, Prediction application programming interface and Maps application programming interface. Over 2,500 high resolution images were purchased for the site and stored on Google Cloud Storage, utilizing compression technology and the Google Application Engine datastore blob. Back-end development was done in Java, while the front end was written in Javascript, HTML and Cascading Style Sheets. Other technologies used on the site included Github, Facebook, Mashape, Rome2Rio, jQuery, REST/JSON and Panoramio.

The solution will evolve and content moderators will identify new picture and content material with the support of learning and recommendation algorithms.

The initial phase of Tripdiscover took 12 weeks to deliver, from project set-up to its release date as a minimal marketable product. The cost savings and speed-to-market of a platform-as-a service (PaaS) ecosystem like Google Application Engine made the accelerated 12-week timeline possible. A second, six-week development release achieved the full-fledged product vision, including user-friendly test feedback. Tripdiscover is taking advantage of the flexibility and real-time scalability that comes with the cloud, serving up to 24 million expected page impressions annually. The publicly available site is accessible at www.tripdiscover.de.