![\"Illustration](\"https:\/\/www.centigrade.de\/wordpress\/wp-content\/uploads\/Bild25.jpg\")
<\/p>\n<\/p>\n
When people talk about ZDF in both large industrial corporations and medium-sized companies, they rarely mean the broadcasting company based in Mainz. Far more often, these three letters are used to promote the pursuit of the \u201cnew gold\u201d of digitally networked ecosystems: numbers, data, facts.<\/p>\n
In many companies, however, it is generally apparent that although reports, measurements, and analyses are available in abundance, the willingness and competence to consistently translate them into real improvements is often lacking.<\/p>\n
This is precisely where the central challenge begins: How can data-based decisions be used in complex, heterogeneous organizations in such a way that they enable a noticeably better user experience for the end user?<\/p>\n
In the manufacturing industry in particular\u2014and especially in the automotive sector\u2014data volumes are generated every day that are difficult to imagine. Highly automated test vehicles deliver up to 44 terabytes of data\u2014per day. Combined with qualitative and quantitative customer feedback, this data pool would actually be the ideal breeding ground for well-founded, user-oriented decisions. A project with vehicle safety system manufacturer Humanetics<\/a> demonstrates the enormous potential of such data: we developed a software solution for this global company that uses virtually all available vehicle data to test and validate driver assistance systems under real-world conditions.<\/p>\n Companies rarely suffer from a lack of data. The challenge lies more in getting this information to the right places so that decisions can be made based on it. In most cases, data is isolated in individual areas of expertise and rarely makes its way out of departmental boundaries, the so-called \u201cknowledge silos.\u201d The reasons for this are as numerous as they are varied: systems are fragmented, relevant contacts are unknown, or the necessary communication channels simply do not exist.<\/p>\n But even when relevant results or information have found their way to decision-makers, another challenge arises: internal conflicts of interest mean that data is weighted differently or interpreted differently, and thus offers little guidance. For the strategic use of the data obtained, a committee is needed that has the authority to interpret, classify relevant information neutrally and objectively, and make the results available to other departments in a form adapted to their purposes.<\/p>\n A Center of Competence (CoC) brings together knowledge, data, and user perspectives\u2014and incorporates them into decision-making processes independently of product lines or budgetary interests. This neutral but comprehensive role is crucial when decisions have to be made in the area of conflict between costs, technology, and user needs. A CoC thus acts as a panel of experts that provides objective guidance throughout the entire product life cycle.<\/p>\n It also ensures that existing data is interpreted in the right context.\u00a0 Instead of individual opinions, what counts here are comprehensible, transparent facts.\u00a0 This makes the CoC an internal mouthpiece for users \u2013 a \u201cuser’s advocate,\u201d so to speak.<\/p>\n The following examples, which I have encountered in the automotive sector, show how this can be achieved:<\/p>\n Inductive charging modules are now standard equipment in modern vehicles. Before the Qi2 charging standard was introduced, however, smartphones had to be positioned precisely in order to charge reliably. Even small deviations led to sporadic charging \u2013 and corresponding customer frustration. The data revealed two options: clearer labeling of the optimal position or an upgrade to the Qi2 standard, which automatically aligns the phone correctly using a magnetic coil.<\/p>\n The decisive lever that can be used with \u201cZDF\u201d \u2013 i.e., numbers, data, facts \u2013 is thus to highlight potential for increasing customer satisfaction and to derive concrete recommendations for action to improve product design.<\/p>\n Two examples from the automotive sector clearly illustrate how important the right context is for interpreting data:<\/p>\n The typical new car smell is one of the most common customer complaints in China, while in Europe it is perceived so positively that it is available as an air freshener at every gas station.\u00a0 Feedback on beverage holders is similarly contrasting: while hardly an issue in Europe, the size and arrangement of these practical everyday aids is a constant point of criticism in the US \u2013 after all, a 64oz Stanley cup needs its space.<\/p>\n Quality studies such as the American Initial Quality Study (IQS) or J.D. Power’s Automotive Performance, Execution and Layout (APEAL) can reveal such patterns. But the essence remains the same: data is only half the battle \u2013 it is the appropriate cultural and situational context that makes it interpretable.<\/p>\n This is precisely where AI and automation offer valuable support. They help to evaluate large, heterogeneous data sets at a speed and quality that traditional manual analyses cannot achieve. At the same time, AI-supported systems \u2013 such as chatbots \u2013 can prepare information on specific topics in a way that is tailored to the target group and make it usable for different stakeholders.<\/p>\n Technology does not completely replace human interpretation, but it creates a contextualized and cognitively quickly comprehensible basis on which people can make informed decisions. After all, making decisions means taking responsibility, and AI will not be able to take responsibility for a wrong decision, either today or in the future.<\/p>\n You often hear people say that it’s \u201ctoo late for feedback\u201d or \u201ctoo early for reliable insights.\u201d The truth is, there’s no such thing as a bad time: customer feedback is valuable at any time\u2014only the nature and scope of possible actions change.<\/p>\n Even if, for example, hardware components can no longer be adapted or modified shortly before the \u201cstart of production,\u201d powerful measures can still be taken to improve customer satisfaction. These can include, for example:<\/p>\n Such measures can reduce frustration and have an immediate effect, even if they do not change the hardware. Usable feedback is therefore not a one-off event, but a continuous stream of data \u2013 and each phase offers a different scope for action. The advantage of so-called permanent sensors is that trends and errors are immediately noticeable as soon as the data situation allows \u2013 not only when asked. This means that both hardware and software-related results from different markets become visible immediately after product release. In addition to faulty components or incomprehensible comfort functions, software issues are of particular importance here. From the number of black screens to the frequency of use of individual HMI functions, all conceivable data can be collected, evaluated, and thus utilized. Tech companies Tesla and Xiaomi in particular demonstrate the role that software plays in vehicles and how vehicles can be perfectly integrated into the private digital ecosystem.<\/p>\n Data alone does not change anything. It is raw material, not a result. Data is not even information. It only becomes information when people understand it, take it seriously, and can translate it into a common framework for action. At Centigrade, we help companies prepare internal data and user research insights in a way that enables well-founded, human-centered conclusions. We do this using tools, metrics, and services that we combine with AI automation throughout the entire digital product development process to get the big picture. This also includes predictively determining relevant UX metrics such as time on task as early as the concept phase: in our freely available AI automation<\/a>, UI elements are extracted from a Figma screen flow and converted into a logical operating sequence using computer vision models. GOMS analysis is then used to determine approximately how much time a user needs to achieve their goal with the given screen flow using the recognized UI elements. This provides meaningful simulation data early in the design process, showing whether certain use cases are too complex or too time-consuming.<\/p>\nWhy data rarely has an impact \u2013 and how it can be made effective<\/h2>\n
The role of the \u201cCenter of Competence\u201d<\/h2>\n
![\"Infografik](\"https:\/\/www.centigrade.de\/wordpress\/wp-content\/uploads\/Bild35.jpg\")
<\/p>\nSupport options through AI & automation<\/h2>\n
![\"Illustration](\"https:\/\/www.centigrade.de\/wordpress\/wp-content\/uploads\/Bild45.jpg\")
<\/p>\nThe time for customer feedback is always right\u2014only the scope for action changes.<\/h2>\n
\n
Conclusion<\/h2>\n