Safely tested on people

How long does it take a driver to put on his favourite music? Is it better to control the air-conditioning using a button or a touchscreen? Škoda tests all that and a lot more on simulators, but also in real cars and with real people. See what the tests look like.

26. 11. 2024 Škoda World

UX (User Experience) research analyses the behaviour of car users. It seeks to understand their needs and preferences, which are then factored into the development of cars and technical solutions. The team of specialists from the User Testing Department collects and evaluates data from various sources, ranging from existing research to organised quantitative and qualitative research with customers. UX research is not just about finding out if and how a part or function suits people, whether they understand its use and whether it is something they expect. Specific measurements are taken – typically, how long it takes a person to perform a particular task.

The testing seeks to find out how people actually use various functions or controls in the car, such as the boot opener.

“For example, in the beginning we measured how long it takes a driver to find his favourite station or a playlist in his library when he wants to listen to music. Later, we also focused on how distracting these tasks are in the car. And we gradually added other subjective measures, such as how people feel, how satisfied they are, and whether they are impressed by something or other, whether they think it’s new or great, for example. Or is it something that was an unpleasant surprise and something he doesn’t like? These days, we also quantify these subjective customer opinions in a number of areas,” says UX expert coordinator Jan Rudinský.

A team of data analysts is developing the collection and analysis of data from the cars themselves. Cars have the ability to detect and record which actions a person has performed, and this can be used to improve the functions of the car itself and the user experience. For example, whether a particular function should be controlled haptically, i.e. using physical elements – levers, buttons or rotary controls – or whether digital control via a display, which is simpler and cheaper to produce, is sufficient.

Do cars collect my data?

It might have occurred to you that your car collects a lot of information. But there is no need to worry - the data that is collected is set up to protect your privacy and is only used to optimise and improve the service. Data analysts collect the necessary information in a targeted manner, based on specific requests. This information comes from vehicles in test fleets and from users who have voluntarily joined the program and provided their consent to limited data collection. Your sensitive data remains secure.

“One real example of user testing driving innovation in a car under development is the air-conditioning control in the latest-generation Kodiaq and Superb. We learned that when we want to lower the temperature, using the display takes significantly longer and is more distracting than using a rotary control. That’s why we took the opposite route to most of our competitors and put rotary controls in the cars,” Rudinský explains.

Here the test technician asked the participant to change the function of a rotary dial.

His team also worked on the design of the new steering wheel, i.e. how the buttons on it should be arranged. Here, too, the data they collected helped – which controls people use often, where the various elements should be positioned on the steering wheel etc. One feature is used ten times in a journey, another a few times in a hundred journeys. “So we know how important each element is, and we were able to design the steering wheel accordingly,” Rudinský adds.

AI assistance

Artificial intelligence has been part of our lives for several years now, and data analytics in user testing is no exception. It would take a human months just to sort through the massive amount of data that is collected in this field. For artificial intelligence it’s a matter of seconds. What’s more, this can be fed into the language model in natural language, so there’s no need to do any coding or programming.

The participant changes the volume of the radio using Smart Dials.

“The system also comes back with an answer in ordinary language. We’re currently experimenting how to use these systems effectively to process the large amounts of data we already have from customers, primarily from the results of various studies. Artificial intelligence is wonderful in that it can identify relevant sources, sift through them and see how a driver is using a particular functionality,” says data analyst Ivana Bohuslavová.

On top of that, AI can put information into context. Let’s say you’re interested in how cruise control is used. The system will go through a huge amount of operational data. It can determine how long a journey took, how many times the driver used the cruise control during the journey and at what speeds. And the answer doesn’t come out as a jumble of numbers, but as clear and structured information. “We are still at the beginning, but we can already see that the future possibilities in this field are extremely broad,” adds data analyst Štěpán Pícha.

Here the participant is changing the function of a rotary dial.

People watching

The second part of user experience research involves tests with real users and customers, i.e. the interaction between the user and the car and how various functions are operated. “Our world is the car interior and what happens in it. In the course of developing and improving the air-conditioning control I mentioned earlier, or other elements, one or more concepts will emerge that need to be compared and verified. Unless there’s a risk of a conflict of interest, we often use company employees for our tests,” says researcher David Rejzl.

Put simply, if you’re checking the functionality of an air-conditioning controller, it’s fine if the company accountants take part. In other cases, though, the user testing team brings in people from outside, and has its own criteria for them, whether these are general attitudes such as their views on sustainability or family values, or maybe specific socio-demographic factors. “We invite them in for an interview, and if we have a prototype car, we talk right beside it. The respondents try out the concepts being tested and give their opinions on their pros and cons, what worked for them and what they didn’t like much. At the same time, we observe the respondents to see how they work with the concepts and how they perform the tasks,” adds coordinator Petr Filip.

Various actions, like radio station selection, are tested while driving.

Testing in a simulator that has a steering wheel, pedals and dashboard is also common. And simulated driving takes place using screens. The driver steers, accelerates, turns, slows down. And performs tasks. He’s perfectly safe – he’s not going to crash. “We observe the test subjects and measure a range of variables, such as deviation from the optimal driving line. This indicates whether a person was under more stress than necessary and the task was more distracting than it should have been. This can give us an indication that a concept is probably not ideal, and that there’s a high probability that it could pose a risk if it was used in real life,” describes user testing expert coordinator Maxim Savelov.

A second faster

The most advanced testing takes place in a real car, usually on a closed polygon for maximum safety. This is testing where the driver feels the real car, sensing speed, acceleration, deceleration, lateral forces and so on. The drivers perform various tasks in cars equipped with the various concepts and the experts watch how the drivers get to their defined destination.

“Again we observe the drivers, and then we talk to them about how easy or difficult it was, what was convenient for them and what slowed them down. We collect the results from all the participants, evaluate them and pass them on to the teams working on each concept. They process this feedback and decide which concept is better and has more potential for further development,” Savelov explains.

Testing temperature control while driving. The participant was asked to described how easy it was to perform the operation and if she knows exactly what temperature she set.

This is how the aforesaid air-conditioning control that eventually made its way into the current Kodiaq and Superb as Smart Dials was tested. “When we asked the test participants to turn the temperature down by one degree while driving, we discovered that the touchscreen control took them a second longer to adjust than the rotary controls – a second during which their eyes weren’t on the road. This is measured with the help of special goggles that use cameras to track the test subject’s eyes and can determine where their gaze is directed at any given moment and how long it stays fixed on that spot. A second may not seem like much, but at motorway speeds that’s 36 metres of driving without looking where you’re going. That was one of our arguments when we were pushing for the haptic controls solution. And our arguments were heard!”

User Testing Department

The department is split into two teams: data analysts and researchers. The former processes extensive data from existing research, but also data from vehicles in operation, and can therefore accurately determine how specific elements (buttons, multi-function controls, displays etc.) are used in cars. The research team focuses on studies that involve actual users, or respondents. The respondents’ real experiences are examined using questionnaires, moderated interviews or by observing their behaviour in a simulator or while driving. “We are currently expanding our activities to regional specifics, especially with regard to the strategic Indian market and the ASEAN region,” adds Petr Filip.