Car Dashboard Research & Design

While I use automobile with some of the research material, I am going to simplify to ‘Car’ for this write-up and discussion, it’s easier to write and more easily understood.  Finally, with all research it should be understood that I am asking for people, regular people, to share their insights and thoughts.  As such, some of these answers might be wrong or incorrect, no judgments will be made except as evidence that the conversation we have as designers is lacking.

Introduction

Cars are interesting machines.  They’ve changed a lot in the last 100 plus years but each step forward is built upon the base.  Watching this growth unfold over that time span and you’re faced with the modern day Ship of Theseus or the grandfather’s ax if that’s more familiar.  Outside of the machine, we have our ever changing society holding onto the philosophical concept of a car with an iron grip of rules and laws.

It is only within the last decade that any true change has been introduced into that cycle of just replacing bits and bobs with a newer version.   The reason for that is not important in this discussion, but has the change truly addressed the issues which are holding the industry back?  Have we gone outside the problem and looked in to see where we can improve the experience?

As is always the case, it depends.  In some ways we have and others we haven’t.  I was thinking about the dashboard of our cars and just how unfriendly they are.  I wanted to design something more modern and responsive to our needs as drivers and passengers but before that step, we need to do research.

Link to the Conclusion

The Problem

Cars are legacy devices trapped within the past and the monumental companies that build them.  As we have watched in the last decade, new companies are moving into existing markets to disrupt the status quo: Uber, Lyft, Tesla, Google’s Autonomous Cars and so on.  Now that there is a true pressure to compete and innovate, new designs and interactions are being introduced.  The problem is they are still stuck within the existing mental model of what a car is.

There are new designs, interactions, mental models, and paradigms waiting to be tested and used in this space.  First though, there is always the research.

How do people use the dashboard of their car with a specific focus on the usefulness of the RPM (revolutions per minute) Gauge/Tachometer.

The Background

How do you use your car’s dashboard while you are driving? First do you have an automatic (or similar varieties) or a manual transmission? If you have an automatic, what do you use the RPM gauge for?

Audi built a better headlight which the NHTSA would not allow on American Cars and Google built a driverless car that has jumped through hoops to be allowed on public roads. To be fair, testing and regulation is not an enemy but at what level does it become excessive or purely a bureaucratic nightmare?  With all this technology and sophistication, why does my car tell me RPMs or Oil Temperature? I don’t know how to use this information in a reliable manner, do you? Does the average driver know?

A caveat. There are reasons to keep the RPM gauge on some cars and some use cases. It is useful when driving a vehicle with a manual transmission or when you have a trailer attached.  This is discussed in the conclusions.

Finally, I like to ask about Why’s of a situation.  I’ll follow up numerous questions asking for text responses to elaborate on the previous question, to understand methods and thought processes.

The Process

Knowing what I didn’t know, the use of the RPM gauge and finer use of the car dashboard only provided the beginning to this journey. If I am going to recruit people to take a survey, I should get answers to a range of questions that will provide information for later in the process. I view survey construction to be analogous to how one would write a story.

To this end, I wrote down my assumptions about dashboard use and then on the other side I wrote down current and future technologies which can make an impact or a vague plot outline. The next step is constructing questions to address and incorporate those two things. This usually results in a giant blob of questions with no underlying structure.

Now I address the thematic organization of the survey, moving questions into a rough grouping where similar questions are near to each other or chapters of a book. When this is finished, it is easier to see if there are duplicate questions or gaps within those I do have.

From there I organize the questions into an order which allows the survey takers to read them as a story. One question leads to the next and information is built on previous answers. To try and avoid memory chunking when possible and in this situation, keep the questions light. When respondents try and guess your motive, they can damage the integrity of the results you get.

Finally, I test the survey. I give it to some people and check responses, ask follow up questions, and make sure everything is working to my satisfaction. And then you let it loose in the wild.

The Generic Data

The survey collected responses from 102 people. I had to remove 1 response from the results. It was clear when reading their responses that they had an agenda for their answers.  These did not answer the question asked, provide informational at all, or were aggressive towards the survey creator.

The survey consisted of 26 questions and was created using Google Forms. Respondents were solicited using several Facebook accounts and groups with an additional website. Individuals identifying as female represented 61% of responses, 35% were male, and 4% chose not to identify.

The age breakdown is as follows:

chart_indi_16

The Habits of Drivers

If you’ve never spent time driving you might not know that drivers tend to have a specific style of driving.  How much of it is learned or built off our own personalities is a fun thought experiment for another day.  I know, personally, that the car I drive impacts the speed I tend to maintain and how relaxed I am to aggressive drivers.

The following questions work to explore the habits of drivers.

Cars are extremely complex devices that weigh thousands of pounds, have more moving parts than we can probably count, and have internal logic systems that we don’t understand.

One of my personal complaints when driving is the ambiguity of the error messages. It isn’t solely a lack of understanding of what the icon means, design of the notification system also contributes to the user having to work harder to maintain the vehicle. With the two cars I have driven recently, a 2008 and a 2011, can alert the driver that tire pressure is low. Unfortunately, it is unable to provide which tire triggered the sensor even though both cars have a dot matrix screen available to display this information.  And more importantly, the car has sensors which alert it to which tire is low it just isn’t designed to tell me.

There are, no doubt, explanations for this behavior which could be contributed to cost or something more complicated. It would be interesting to interview designers at the Automobile manufacturers about the thought process behind this decision but that isn’t relevant here other than to say it is a poor omission.

With all that…

Do you know what all the icons on your dashboard represent?

chart_indi_11

The results of this question surprised me.  With 92% of respondents indicating that they know most of the icons on their dash, I don’t know if I can trust the data for the question and the question itself.  I knew before this survey that this area had the potential to be spun off into its own research project.

Thinking through this question, one can see potential pitfalls.  When doing research you want to avoid ambiguity (usually) so that you can get answers that help you but you also want to be concise and to the point.  The ambiguity here is not in what an icon is but rather what is the scope of the question.

For instance: you know what the gas gauge means but do you know what the following icons mean?

car_dash_4 car_dash_2 car_dash_1 car_dash_3

 

 

Going even further, if the icon only shows up once a year do you consider it within this question? I have a hunch that this question was answered as the following, “Do you know what the icons you see on your dash mean?”

But even then, people might not be totally truthful. The following two responses were for people who marked ‘I know most of them’ and then their responses to a question about the rpm gauge.

  • Speed of the motor itself?
  • I would have to be in my car looking at the dash to even know what that is…

Which sort of lends itself to the idea that they don’t know most of them.

Icons aren’t the only way to get an understanding of overall knowledge of a person and their car.  When I was younger car maintenance was talked about more, as a skill that was needed but this isn’t something I hear about in my current social circles.

These questions serve to enlighten us about how people view their cars, inferences can be made when combined with other questions and answers.  Like I said at the beginning, we’re building a story here.

How do people treat their cars?  This is a broad question but we can ask a specific aspect of it, to get a feeling for people’s actions.  This also builds on some concepts later in the survey.

chart_indi_15

Do you perform your own maintenance on your car/vehicle?

chart_indi_14

But what does maintenance really mean in this scenario?  Do you consider it maintenance to put windshield wiper fluid in your car?  Or to the fill the gas tank?  Perhaps you think of it as someone who can change the oil? Or the brakes?  To understand the extent of the maintenance, there was a follow up question about what maintenance they did.  I developed an affinity diagram with this information.

chart_indi_13

The affinity diagram categories are:

  • None:Users indicated they performed no maintenance on their vehicle.
  • Minor:Users said they performed minor maintenance on their vehicle: tire pressure, wiper fluid, and check fluids.
  • Intermediate:User performed more complicated maintenance on their vehicle. This was primarily noticed by the addition of changing the oil in answers but also included tire rotation or brakes.
  • In Depth:These users mentioned complex maintenance in their answers that included more specialized knowledge of the vehicle: suspension, engine work, gaskets, and spark plugs.
  • Unknown:There were three responses here. Two users spoke about what their spouse was capable of and not themselves. The other person here responded with “All the time” for the “Do you perform your own maintenance?” But then answered “everything” for what they do. While it could be argued that they would fit into one of other categories, I left them out due to the ambiguity.

A healthy discussion could be had about the exact impact that someone who performs maintenance on their car has on their ability to use their dashboard.  While the relation isn’t perfect, this researcher is putting forth that people who do Intermediate or In Depth maintenance have at least a slightly better understanding of how their car functions as they would need that information to perform the maintenance.

Finally we can explore the subject of the RPM gauge and I can finally learn if this is antiquated unused information or a display that everyone uses while driving.  This was the question that inspired this research.

There were 85 responses for individuals with Automatic and related transmissions (CVT, DVT, and electric) and 16 responses for Manual transmissions.

chart_indi_12

Without further ado, “Do you use the RPM gauge while driving?”

chart_indi_10

This includes responses from all of the participants.

Automatic only responses:

chart_indi_09

Once again, we need to understand deeper into the why of this question.

“What do you use the RPM gauge for?”

chart_indi_08

Below are the 5 categories I condensed responses to.  With the initial question about use of the RPM gauge, it is possible that some responses were skewed by the Social Desirability Bias.

  • None: User does not use the RPM Gauge, possibly doesn’t even know what it is for or where it is in the car.
  • Informational: User consults the RPM Gauge for information on how the car is running but does not use the information in driving the vehicle.
  • Purposeful: User consults the RPM Gauge and uses it in making decisions towards operating the vehicle.
  • Other: User might not have the RPM Gauge on their vehicle or use is indeterminate.
  • Manual: User drives a manual transmission wherein the RPM Gauge could be considered mandatory.

The distinction between Informational and Purposeful can be razor thin.

A side note about the affinity diagram which made this chart.  In the course of making it a bat broke into the home and spent a noticeable amount of time looking over the work.  There was a distinct communication barrier but I think they approved.

A couple responses from this question and the group they belonged to.

  • Group None:I would have to be in my car looking at the dash to even know what that is…
  • Group None:It’s an automatic transmission. As far as I know, there’s no reason to use it at all.
  • Group None:Ride mileage per gallon?
  • Group Informational:Just to make sure it’s not revving or idling too fast
  • Group Informational:I glance at it when on highway and gears shift to make sure it returns to normal range.
  • Group Informational:Watch acceleration, try and estimate fuel usage, see if car is idling ‘correct’
  • Group Purposeful:Since I drive an auto, it isn’t as important. I still use it to know if it’s going to stall in really cold weather or if the engine is idling wrong.
  • Group Other:Keeping below a certain threshold as I try to keep gas usage down and engine usage low.
  • Group Manual:Just for shifting really and starting and stoping [sic]. I check the speedo less than my rpms. I’d like to get good mpg and I try to get low rpms while driving.

It can be easy to assume that only people who drive manual vehicles use the RPM gauge but there were some responses of manual drivers that indicate it isn’t a necessity.

  • Group Informational:Checking if the engine has warmed up
  • Group Purposeful:It’s a standard so its super useful. If i have music going and cant hear my engine it helps a ton, especially with the cars short gears.

So we’ve covered the use of icons on the dashboard, how people deal with maintenance and the information it provides, and their thoughts about the RPM gauge.  I would consider these topics to be the big legacy concerns when designing a car’s dashboard.  But there are lots of changes on the horizon, how do people feel about this?

New Technologies as the Future Approaches

With each new model, technologies that were once science fiction are finding their way into more cars and significantly decreasing costs.  Is this something that the common driver wants with their car or something that a designer feels we should want?

I asked a couple questions about Voice Control, Cruise Control, and an overall sense of Trust with the car.  These technologies are seeing extensive development from every company that does business in this field.  Cruise control is changing from the static of ‘go this speed’ to a device that maintains distance between vehicles, assists with lane changes, and in some cases drives.  Audi, BMW, Ford, Volvo, Google, Tesla, and so on are working on the autonomous car.

Does your car/vehicle support voice commands?

chart_indi_07

Do you want your car/vehicle (or next) to support voice commands?

chart_indi_06

While a little over 25% of respondents have cars that support voice controls, it something that almost 60% of people want their next car to support.  While not asked, it would be interesting to see what features people wanted from the voice controls.

While autonomous vehicles aren’t here yet, the technology is improving and having a rough idea about users and cruise control adds support to detailing their style of driving.

Do you use Cruise Control?

chart_indi_05

As you can see, a respectable 65% of respondents use cruise control at least ‘Sometimes.’  How much is Sometimes though?  Or more to the important point, what makes someone use cruise control?

Please explain how you decide when to use cruise control?

chart_indi_04

For the most part, the groups are self-explanatory but let me elaborate on a couple.

As we can see, a lot of responses were concerned with Open Roads and Long Trips. I’d be interested to see the statistics from Tesla’s Autopilot concerning when it is used and if it matches those sentiments. I have seen reports that people are using it incorrectly (sleeping and driving?) but what does that truly mean? Haphazard? Innovative ways? Unexpected? Contrary to how it was designed?

There were a couple responses that were marked as ‘Sometimes’ used but their reasoning was about the danger that cruise control represents. I placed these people in the ‘Negative’ group. The question references a positive effect and these answers explore a negative effect. I do believe that the information on why they don’t use it is helpful and can be incorporated into future designs and work; it just didn’t fit within the question as asked.

Only nine people mentioned Weather specifically but it could be argued that some responses hinted at it, these were not included though.  Just as some people might have been subtle in talking about the weather, only one person spoke about using cruise control as a method to maintain speed within the flow of traffic.

When we make decisions we balance multiple inputs to come to a conclusion or decision.  If someone is asked about a decision they made, they’ll choose an answer that best fits the scenario or the biggest factor of the decision.  It doesn’t mean that it is the only reason.  What this means is that some of these reasons may cross over to multiple people.  We don’t know exact percentages for each reason, just that these were the most popular.

Let us get into areas that tie into a more automated driving experience.  This doesn’t have to mean where the car is the driver, just that it is more self-aware of its status and able to inform us.

Would you trust your car/vehicle to remind/prompt you about maintenance?

chart_indi_03

Please explain the above answer.

chart_indi_02

The affinity diagram categories are:

  • Unknown:User’s response to the question was lacking content or vague.
  • Lack of Trust:User did not feel they could trust the car to report information in an accurate format whether this was due to their own preferences or lack of sophistication in the display.
  • Lack of Info:Information provided to the user does not have enough context for them to form a plan of action with or it is believed that the fidelity of the report is missing from critical areas such as tire wear or brakes.
  • Trust:These users felt that the car would not be able to match their knowledge of the car and any information provided would be wasted or on the other side, that they were not qualified to manage the car and only a mechanic was sufficient.
  • Indifference:The responses in this category felt that this functionality
  • Frequency:Users were worried about the frequency of prompts or obtrusiveness.
  • Minimal:Users mentioned that they used one or two prompts but also that this wasn’t something they cared about.
  • Useful:Users found the prompts to be helpful in managing their car or something that they would use if it was currently lacking.

It should be noted that some answers had multiple applicable groups. As such, there are more responses than people who took the survey.  Also, that there is a section for Lack of Trust and a section for Trust.  Lack of Trust is mostly directed towards the car while Trust reflects the other side or that there is a more reputable source.

Would you like more information provided about the car/vehicle?

chart_indi_01

As someone who enjoys data, even if I don’t get to apply it sufficiently beyond spreadsheets here and there, I was quite happy to see that over 50% of people wanted more data from their vehicles. This leads to questions about what the user would do with the data and how it would be adequately communicated. I think this is a great area to have an app that is able to report the data in a visual manner. This would also allow some remote cases that could provide a better user experience; how much gas do I have left, what is my oil life at, or how many miles since my last ‘x’ without having to be in the vehicle.

Conclusions

A recap on why hasn’t anyone ever done research here, and they might have, you might be asking yourself?  Cars provide a captive audience where the user is unable to go somewhere else.  It behaves like a monopoly.  As people know, in such a situation, competition is usually a token gesture.  It is only in the last couple years that new features are starting to change the landscape.

There is a lot of information here about a lot of aspects of what is a complicated system.  So what did we learn and how do we apply this going forward?

The results for the Icon question were disappointing, frankly, but even in that there were areas of success.  While I wouldn’t say for sure that the results were impacted by the Social Desirability Bias, it is a possible answer.  The Social Desirability Bias has to do with how a person will answer a question in a way that does not embarrass them in front of others.  It can be easy so see the response as, ‘of course I know what these icons mean,’ instead of asking others what they mean.  This area could benefit from dedicated research.  I briefly began work on a test that would have a grid of icons that matched the model of car they drove and asked for the description of it.  It would be a lot of work and what would it prove?  As designers we know we need to always design comprehensible error messages.  Assuming we use digital displays, we can move away from icons and use plain text for errors.  Left Front Tire Low or Oil Change Needed with more serious errors having a greater change on the dashboard.

As we saw from the questions about the RPM gauge, that there is a lack of use and an attempt to use the information for other means e.g. an indicator for some other problem.  Which is one of the age old problems designers’ face, what information does the user want and how do I give it to them.  But has anyone ever asked/answered this question with this domain?

This is one of the fundamental problems on dashboard use, the lack of understanding on what the user wants, how to give it to them, and adjusting for changes in mental models.

There is a drawback to going down this path though, beyond just the changing of something that has remained static for so long.  It is the concept of Trust.  You’ll remember that this came up when discussing the car prompting for issues with two different answer groups.  It is a big concern when doing design work on Artificial Intelligence.  We tend to understand humans better which is most likely because we are one and share empathy.  The user needs to be able to trust the machine, to understand the path it takes to reach its conclusion.  My gas light comes on at 50 miles to empty but what if I couldn’t find this statistic in the manual or that the amount was random, would I trust this indicator?  Would you?  This is probably one of the biggest reasons that Google is so forthcoming with data from its autonomous car, so that we can begin to trust that the technology is safe.  This information should be available to the user, it would remove the problem solving that takes place on the road and place it on your couch or at the restaurant.

More reading on trust: FastCo.design or Information Week

As I wind this down, there were places for more extensive research along these paths.  This whole subject could be explored, which would be exciting.  As cars become more automated the cruise control aspect should be researched extensively and also from the other side, as drivers who have to drive with autonomous cars.  I’ve spoken about the work with the icons a lot but that could use plenty of study.  I think it is synonymous towards the work that has gone into the hamburger menu on desktop devices.

The end takeaway from this research is that the future is changing and we need to adapt our response to it, tacking on new features on top of old technologies makes for complex and muddy interactions.  Some of the biggest tech devices lately have been for simple concepts executed cleanly.

But do we need an Rpm gauge? No, not really unless you do.  But really, no.

Future Designs

There is a duality to designing for the future.  More information is required but it can be difficult to cram a lot of information into a UI with a small number of physical buttons that already serve other purposes.  I think a good example is how hard it sometimes is to change the time in a car.  My grandmother’s car is so difficult that no one in the family is able to change the time for daylight savings – including me, the expert.  Having a touchscreen available to the user would allow a changing UI/UX for the information that was needed but also being able to pair that information with an App would be so practical.  Below just gives some ideas for the dashboard/car while I plan to do design work on the app in the future.

So we have done all this research, asked questions of everyone, followed up with interviews on some people, and have a stack of data but where do we go from here?

I think it is obvious that the current standard is lacking in many ways.  In designing for the future, there is one technology that I think is mandatory if we want to escape the designs of yesteryear: A digital display.  Including a digital display allows us to escape the static designs of the past and give us the ability to change the information that is displayed.

This leads me to the next important change, something that is minimally used today but could be used to greater extent: Progressive disclosure.

For those who don’t know what this means, it has to deal with displaying only the information that is needed to reduce the mental load of the user.  An example that is currently used in (most of) our cars: when your gas tank approaches empty there is a light that turns on or sound that beeps to alert you that you have X miles to empty.  My car has a threshold of 50 miles but I know my friend’s is 30 miles.  The light isn’t needed until this rule is triggered, so it remains off.  Actually, a lot of the error notifications work this way and is something that is well understood.

A car is a legacy system, as we’ve discussed.  It is built upon over the years, new layers placed upon the old layer and change becomes very difficult because it is assumed that everything needs to remain within the design.  Or more so that users will leave if the experience changes to greatly.  Cars are a bit different, they represent a greater investment and changing cars is more difficult once one is built.

But this thinking can trick you in this instance, we’re not getting rid of anything we’re just moving it or changing how it is communicated.  Let’s jump in.

What do you need to drive your car?  On the dashboard, what information is absolutely imperative 100% of the time?

  1. State of Car (Park/drive/on/off/etc…)
  2. Speed of Car
  3. Turn Signals

You’re probably trying to picture your dashboard right now to argue that I have forgotten something.  The two that come to mind are the RPM gauge and the gas tank gauge.  We’ve discussed the RPM gauge but odds are you don’t need it to drive, so why distract yourself with it?

But what about the gas tank?  This is an interesting question but how often do you check that gauge?  How does it impact your driving?  Think about every time you’ve gotten gas, what was the reason for it?  Now how often was that when the tank was of 3/4ths full?  If the tank was that full, do you need to know that?  Or would it make more sense for the tank to only display at some user set level?  Like 50% or 30%?  This ignores the dynamics of how we drive and the reasons we drive.  So we could tell the car that we like cheap gas, or gas from this brand of station, or that this is a long trip (gps) and we’ll need to never be further than 100 miles from a station.  And when the criteria are met, it alerts us.  All of these things are things that cars do in pieces already.  The problem with something like this is designing the interaction to be smooth and easy.

I went driving with someone, nothing too taxing, and asked them to imagine this setup.  They had an issue with a gauge missing off of their dashboard.  Specifically the Engine Temperature gauge.  I asked more to understand why this would bother them.  They told me that a previous car was always overheating and so they rely on that gauge now.  It is a different car but it is still a concern for them.  I asked if they would trust the car to alert them to danger with this specific case and they said no.

This isn’t the full design, this is to get you thinking about the problems, solutions, testing, iterations and where to go.  There is so much more here but I will spin it off to another project.

If you want to talk more about this subject or your feelings on my results, drop me a message: Email