© fotolia - katafree

Start-up in the Group

ZERO.ONE.DATA – Big data for everyone

06/2016 - ZERO.ONE.DATA is the first start-up venture pioneered by DB Systel and was launched to forge new paths in the big data jungle. To this end, the project transcends fixed, rigid corporate structures and challenges everyone else to think outside the box and get actively involved in a new culture of data exchange and data usage.

If you would like to understand what distinguishes the small ZERO.ONE.DATA (ZOD) start-up from other companies in the DB Group, all you have to do is take a look at the think tank behind the team. In the offices on the twenty-seventh floor of the Silberturm in Frankfurt, nothing is reminiscent of the plain, down-to-earth working environment of one of the biggest players in the country. If there were any wallpaper there, you would never be able to see or recognise it. Memos, leaflets, post-its, notes and various plans are stuck everywhere. Welcome to ZERO.ONE.DATA, where everything is just a little bit different – just what you’d expect from a start-up.

As you can imagine for rather young teams, , one of the buzzwords circulating around in the last few years has been on people’s lips here as well – that of “big data”. Mountains of data that for ordinary people would initially seem boring represent nothing less than the lifeblood that drives exciting projects. “Our staff members are quite inquisitive folk”. Team leader Lars Freund cannot help but to grin. Whereby it must be said that he is not very fond of this title. It goes without saying that somebody has to step up to the plate and formally accept responsibility. “More importantly, however, we promote an atmosphere of mutual trust. To this end, we employ the principle of dual control. When two individuals from the team believe that they have reached a solid decision, this is then viewed as being okay.” Voting rounds in large groups are only considered if they are absolutely necessary.

© DB Systel

We do not devise elaborate concepts in advance. We also do not have any strategy or policy documents collecting dust in some cabinet because we are too young for that. Rather, our work is based on the principle of trial and error

Dr. Lars Freund (DB Systel GmbH)

Day-to-day work at ZERO.ONE.DATA strays just a little from the norm, you might say. In this context, the team strives to reduce as many cross-disciplinary functions as possible. Intermediary units such as portfolio management and business development, and strict separation of the sales team from the delivery or supply unit, are non-existent. After all, because the team comprises only 15 individuals, everyone must be immediately involved in adding value in terms of acquiring, analysing and preparing data. Agile project management and streamlined processes are the key to all of this and are fully in line with the spirit of a proper start-up.

Big responsibility – small team

ZERO.ONE.DATA enjoys a level of freedom that other colleagues would envy. Whereas the Group typically maintains long-term strategies and therefore promotes lead times, the extremely effective business unit can adopt a different approach as it assumes end-to-end responsibility for the topics covered. ZOD thus decides directly what can help the customer realise success and simply tries this out. “We do not devise elaborate concepts in advance. We also do not have any strategy or policy documents collecting dust in some cabinet because we are too young for that. Rather, our work is based on the principle of trial and error”, explains Freund. And this is exactly the right approach when it comes to data analysis: “Only in this way can you really work in an open-minded and cost-effective manner and not rule out any analysis results. In the end, sometimes all it takes is a quick glimpse at the data to determine whether a gold mine is buried in it.”

Another difference to working in traditional DB Systel departments is that the start-up engages in preparatory work for potential customers. This is an option that conventional operative units frequently cannot take advantage of as other key performance indicators are often used there for control and monitoring purposes.

Staying agile makes it possible to realise so-called minimum viable products, or MVP for short. In the case of ZOD, these are analytical structures or algorithms. MVPs are plain and simple and initially have only as many functions as are absolutely required. It is on this basis that one can then test and further develop analytical procedures. This is a typical start-up method, since an MVP test is typically far more cost effective than developing a complex product in advance. If a solution approach turns out to be ineffective, it is simply rejected and the team deliberates on a new strategy. No bureaucracy, no red tape. No time wasted. Work is carried out effectively and in close collaboration with the customer.

Big data all inclusive

The uncomplicated approach represents only the first part of the ZOD method. Big data is an exciting field in which it is also very easy to get on the wrong track, however, since the quantity of data available has now reached a point where it is almost impossible to be completely on top of. One of the main tasks ZOD pursues is therefore to advise customers accordingly and make recommendations for specific analytic approaches.

Portfolio of ZERO.ONE.DATA (ZOD)

To gain a better overview, it is a good idea to examine the portfolio of ZOD in greater detail, which breaks down into three areas in which ZOD offers a range of different services:

  1. Infrastructure as a Service (IaaS): ZOD implements the environment (so-called Hadoop memory) in order to securely store data. In the process, ZOD proceeds in accordance with the data protection requirements outlined by the DB Group. Currently, the system is being operated in the Berlin data centre maintained by DB Systel while ZOD evaluates a cloud-based variant of the information storage programme. ZOD also collaborated with the “Infrastructure 4.0” initiative to realise the free databox, which serves as an exchange platform and virtual window for data provided by companies of the DB Group.
  2. Platform as a Service (PaaS): ZOD provisions a fully functional big data analytics environment: A special IT system that can be compared to something of a toolbox for analysing data. On this platform, customers can work fully independently as well as utilise the services offered by ZOD. The basic service package consists of applications that are either free of charge or are reasonably priced. High-end software solutions with complex analytical functions can also be offered if so required.
  3. Consulting: The actual value added takes place via customer-specific applications through which the data can be used in a target-oriented manner. ZOD works together with the customer to devise data usage strategies, consults and advises on selecting the correct data sources and helps safeguard the level of data quality required. By collaborating with the customer’s experts, ZOD develops matching mathematical models and realises the application environment. The data and corresponding analytical results are then prepared for visual presentation and made available to the customer for further examination.

Clairvoyant janitors

ZOD partnered with DB Energie in a design thinking process with different departments and external experts to identify potential areas in which processes can be optimised and operating costs reduced. ZOD not only analyses data, but also consults and advises partners as well as develops targeted strategies together.

Dr. Stefan Manke (DB Energie GmbH)
© Deutsche Bahn AG

In the case of DB Energie, for example, it was discovered that expensive and sometimes unnecessary maintenance work is performed on substations. Anticipating such work used to be something of a crapshoot as electrical parts frequently do not reveal any clear signs of physical wear. The data experts, however, found out that the ageing process of specific insulating oils could be tracked and pinpointed using special sensors. A potential breakthrough! The next step was to analyse exactly how a full battery of sensors could help devise wear or needs-based maintenance schedules to improve efficiency.

Ways of making improvements with respect to purchasing power and customer service were also investigated. Dr. Stefan Manke from DB Energie was satisfied overall. Although no immediate financial benefit came out of the projects, many potential improvement areas could be identified. Different methods of working also promoted new perspectives: “For the first time, we learned that the value of some of the data we maintain was not even on our radar screen. This data was also only available in paper form and took a great deal of time to prepare accordingly.”

The trove in the Silver Tower

Every department of Deutsche Bahn should be able to profit from this data expertise in future. The only real requirement to ensure that this happens is that of active involvement on the part of the DB Group companies. As suppliers, converters and compilers of data, the experts behind the scenes are absolutely irreplaceable; without them, there would be no high-quality data and the ZOD team would have no starting point. Everything can be quite interesting when you think about it: How often are trains repaired? What faults or malfunctions are present? Data that leads to a directly tangible benefit or value is especially coveted. This is not a prerequisite, however, since the data on the insulating oils used by DB Energie were also beneficial although the customer and team initially had to “dig deep”.

The ZOD team must also continually work on honing its expertise and start-up culture. “With every day that passes, we learn new things as far as data analysis, algorithms and synergies between various different data records are concerned as well as deepen our self-image and awareness”, attests Freund, who himself has since grasped that “working in start-up cultures primarily means showing self-initiative and team spirit as absolutely nothing happens in their absence.”

To ensure that progress is made even more quickly in future, the databox started a data exchange portal for DB Group companies in May. After having convinced those in charge, many departments can now make data records available and profit from those of other company units. Previously, there was simply no suitable platform in existence for preparing and exchanging this data. Data is now the new black gold; it conceals great treasures, and by working together with the customer, ZERO.ONE.DATA plans to exploit them.