Corporate Data Meets Startup – Lessons Learned from the 34th TDWI Roundtable Berlin

TDWI as a networking and knowledge platform established its roundtables to let experts share their knowledge, experiences and latest inside information with other experts and a broader community. 

This 34th TDWI Roundtable Berlin focused on various insights into startup and corporate experiences around the collection, processing and analysis of data. Marco Szeidenleder, Carsten Minnecker and Andreas Wiener hosted the event and subsequent discussion.

Christian Gust, Director of Data Platform & Analytics at Pleo

Christian started off with an important reminder: To be successful, a company’s data strategy should always take into account the preexisting corporate culture. Using the “Competing Values Framework” by Quinn and Rohrbaugh, he explained that sometimes there are vast differences between corporate cultures, e.g. between a collaborative culture in a decentralized business and a control-type culture in a more centralized model. While each corporate culture comes with advantages or disadvantages, Christian stressed the fact that it is important to know which kind of culture your company houses and that it takes quite some time to implement changes. Contrary to a mechanistic model, this strategy is part of a more organic approach that pays attention to values, principles and rituals within a company to establish changes through new positive incentives and experiences. As a result, it creates intrinsic motivation within the employees to experience the benefits of a new data strategy and consequently, it generates the best conditions for an increase in data literacy.

Leo Marose, CEO and Founder of StackFuel

Next, Leo talked about ways to establish such profound data literacy in companies. According to him, one of the biggest challenges for many companies is not only the lack of much-needed data experts but also the absence of general data expertise among most employees. To enable companies to make use of this dormant potential and to meet the tremendous demand in this field in the coming years, a few necessary steps need to be taken:

1. The organization needs to develop a custom-tailored data strategy.
2. New data talents have to be employed.
3. The whole personnel needs to receive (advanced) training in order to improve their data literacy skills.

Leo stressed consideration of the last component, highlighting the importance of developing an adequate training strategy. Not every employee needs to be a data engineer or analyst. Some might need to undergo reskilling courses – others only have to expand their existing knowledge (upskilling). However, to distribute new knowledge within the company more efficiently, executive personnel should be among the first to receive such training.

Marc Roulet, Director of Analytics and User Research at XING

Marc talked about the benefits of combining user research and data analytics in his presentation. One key insight was that data analytics can provide answers to the questions:

How are users behaving when using my services and, therefore, what does the user journey look like?

On the other hand, when it comes to the users’ motivation or attitude, so explaining why they behave the way they do, analytics reaches its limits. Here, on the other side of the coin, user research is needed in order to receive a more satisfying answer and to create a better insights value. Mentioning Gartner’s analytics maturity model, Marc pointed out that the data collected by user research can inform business decisions at every step of the company’s analytics maturity stage: at the descriptive, diagnostic and predictive and prescriptive stage. He stressed the necessity of treating user research as one of the main pillars of data-driven insights and product improvement within companies next to data science, testing, and analytics.

Michael Stein, Senior Business Analyst at Scout24

The final presentation by Michael cast a spotlight on data collection and usage within the framework of real estate search engines. He asked the question:

Is search engine data corporate data? How can you integrate such data into your own data stack?

Michael used his own company’s digital ecosystem as a hands-on example for a proprietary solution. For many users, the Scout24 (IS24) platform is perceived as a search tool despite offering more features, like a communication service. This is important in order to understand the user funnel and which kind of data gets produced when and why. The IS24 user often enters the real estate search via Google’s search engine, a process that can partially be traced or indexed. This data combined with the IS24’s own search engine data makes for a raw treasure trove. Its usefulness is maximized by cleaning it from bots and spiders and joining it with consent-based analytics data and other information (e.g. object information, transaction-oriented conversions). 

IS24’s data landscape is designed so that searches – or more precisely – distinct IDs of these are inherited down the funnel into several data sets. This process creates a more coherent data stack which can later be analyzed and, in return, creates the opportunity to optimize search engines. While regarding all of these different data sources, Michael mentioned that it is important to find a “Data North Star”, meaning: Being aware of the information which is worth tracking since tracking a lot equals an increase in the maintenance load.   

Key Discussion Takeaways

Which data problems are typical for corporate and startup?

1. Startups: The dependence on key employees might create a bottleneck – if one key expert leaves the company the whole data analytics infrastructure might cease to exist.
2. Corporate: Even though consulting agencies often take a closer look at accounting, the quality of the databases of the company is seen as less important –  this might cause trouble especially after several mergers.
3. Startups: The lack of a coherent data strategy (and specific targets) from early on and missing expertise when it comes to tools and data literacy.
4. Corporate: A perfection over prototyping attitude is often a problem in bigger companies. It may disregard the necessity to create a prototype in order to measure its specific impact and instead focuses on generating a perfect product while collecting as much data as possible.

Which are the advantages of corporate and startup companies when dealing with data?

1. Startup: Due to lower complexity regarding the company’s structures, a faster working pace and the implementation of new tools/processes is simpler. 
2. Corporate: Often, bigger companies have already implemented a more thorough data strategy while adhering to useful structures such as regular strategy meetings to distribute concentrated knowledge.

Which aspects come to mind when considering the cooperation between startups and bigger businesses?

1. Data privacy is often an important topic in bigger businesses: Legal departments can facilitate any related cooperation whereas work councils might slow down the exchange of important data. Startups are often quicker when it comes to exchanging information and are not limited by multi-national legal regulations but frequently need external legal advisors.
2. Wayra, Telefonica’s startup accelerator, was cited as a positive example since they offered a clear focus on business strategy instead of innovation management, and provided transparent communication when collaborating with startups: First the basic problems and goals were defined in a contract – afterward financial investments were made.
3. Zalando was another positive example since the company created several spin-offs which profited from preexisting contacts and communicative structures within their former parent company.

More businesses focus on using central coordinative units, often called Business Intelligence Competency Center (BICC). What are some of the key aspects which constitute such a center?

1. It’s a good mechanism to establish company-wide standards (e.g. when it comes to tools) and to implement data strategies faster, mostly in larger businesses.
2. It combines the expertise of data and business intelligence analysts. 
3. It’s often seen as a service center supporting other departments, e.g. by providing necessary infrastructure, providing insights, facilitating interaction between other units and increasing the transfer of knowledge (e.g. through coachings and training). That also means that the data ownership remains within the according department while the BICC governs and analyzes the different data according to the company’s business strategy and advises the other units on its findings. 
4. The range and variety of its tasks and its name can differ from company to company – in some it might be called Center of Excellence (CoE).

* Der Beitrag spiegelt die Meinung der Autoren wider und ist keine allgemeingültige Meinung des TDWI. Als TDWI bieten wir die Plattform, alle Themen und Sichtweisen zu diskutieren. *