Drive-Thru

Experience Redesign

Redesigning the physical experience of drive-thru about fast food restaurants.  

Overview

My Role & Contribution​

UX researcher - conducted online and field studies, interviews and user testing to gather user insights

Service designer - brainstormed different solutions and went through the human-centered design process to sort out  final solution and assisted with redesign prototyping

Finding the Right Problem

 

 

 

Following the double diamond model, the first thing that we decided to do was to find the right problem.  In order to find the right problem,  we started with a bottom-up approach. We first identified our stakeholders, those who are affected by the drive-thru.

Stakeholders

Primary Stakeholders: Orderers (Customers)

 

Needs: An experience that is hassle-free, easy, and without error

Secondary Stakeholders:  Order-takers

 

Needs: 

· Performance is based on correctly serving customers efficiently

· Enables them to provide better satisfaction to the customers

User Research & Data Collecting 

After identifying our stakeholders we had a guiding direction for our data collection. We used various methods of data collection, but the most essential method of identifying our design problem was the interviews we conducted with customers while they were using the drive-thru. The interview questions were around the experience they had just been through and their opinion about their past experience. By doing this, we were able to receive key insights, make observations, and fundamentally understand our topic of focus in a new light. We went to 5 different fast food restaurants in total.  

Pictures from field interview and observation

To supplement our interviews we also conducted survey and secondary research in order to get more data and a wider perspective, which took the form of a Google survey form, which we posted to Piazza and Reddit, and the 2017 QSR Magazine Drive-Thru Performance Study, a study conducted by the UX company SeeLevel.

 

The data we collected, both from our interviews and the secondary research, assured us that the drive-thru is far from perfect and we could do much to improve it. Our interviewobservation and survey data were collected and transformed to an affinity diagram which categorized and consolidated all the data we had. This is our digitized affinity diagram.

The physical affinity diagram, made with stickers.

Major Problems We Found: 
  • Time - people felt like they were waiting too long or they felt a time pressure to order faster to not make other customers behind them wait. This issue was often related to problems involving the menu, communication, or long lines.

  • Physical layout - customers had trouble navigating the drive-thru due to unclearness of lane merging or the lanes being too small.

  • Order confirmation - interviewees felt frustrated over the fact that sometimes orders are not done wrong and that this could be improved by having a better order confirmation system.

  • Communication - sometimes the customer couldn’t hear the order-taker and vice versa, which then led to a delayed experience and frustration from both sides.

  • Prior knowledge/experience - some interviewees related their experience of drive-thrus with their own prior knowledge

We then went through each of those categories and threw out whatever idea we had to fix that particular problem. The most feasible and advantageous ones stuck, while the ones who would give little benefit were thrown out. In the end, we came up with the following main potential design ideas:

○ Lanes

○ Physical Layout

○ Menu 

Finding the Right Solution

Initial Design Problem

" Components of the ordering process of a drive-thru create unnecessary conflict for stakeholders, which negatively impacts customer satisfaction." 

Exploring our data

After creating the affinity diagram and redefining our area of focus, we created additional design models that would allow us explore our design problem from various perspectives. 

The Sequence Model allowed us to visualize the standard procedure that customers undergo in the drive-thru.  By doing so, we were able to identify particular points of the drive-thru that could cause problems and points that could be improved. 

The Collaboration Model served a similar purpose, but this time focusing on the interactions between the customer and the order-taker. 

Personas
Initial Ideating

Our design ideas essentially address the problems of the lanes, physical layout, and the menu. Our goal is to have customers be able to navigate the drive-thru as quickly and easily as possible. For the lanes, we came up with the idea of having two separate lanes for two different purposes. The lane that hugs the restaurant’s building would be the regular or “leisure” lane. This lane would be meant for people who are in no hurry and who don’t know what to order and need a longer look at the menu.  

For the physical layout of the drive-thru, we wanted to have two lanes which would merge to the pay and pick-up windows. To avoid confusion, we would also add a stoplight that would inform cars from each lane when to move forward.

 Finally, we wanted to have two menus displayed. One would be the “pre-menu” which would be placed a few cars behind the speakerphone and above the whole drive-thru. It would essentially only display the most popular items (determined by rating or volume), so this menu would be less cluttered and confusing. The second menu would be more extensive and specific and would be placed by the speakerphones. 

Re-scoping

Finally, from the received feedback, we found that the menu was a big source of the time pressure and confusion that a lot of people felt in their drive-thru experience. Since redesigning the size and physical layout of the lane was out of our reach and completely dependent on the restaurants themselves, we decided to solely focus on the time factor, which mostly caused by the menu.

The storyboard based on the idea

To attempt to give customers more preparedness when ordering, we wanted to have two menus displayed. One would be a “pre-menu”, which would be billboard style and placed a few cars behind the speakerphone and above the drive-thru lanes. This menu would essentially only display the most popular items (determined by rating or by volume), which would also address the confusion problem by having it being less clustered than the main menu. The second and main menu would remain the same - it would be more extensive and specific than the first menu, and it would stay placed next to the speakerphones.

New Design Problem​​​​​​​

Prototyping Process

In order to test out our solution, we build a paper prototype with several components. The size of the prototype was pretty big. We intended to create a sense of real drive-thru by having the large menus, but it was still a bit small comparing to real life situation.

First Iteration

The Extensive Menu

The Billboard Menu

For the first iteration, our design idea was to simply have a billboard menu that was a few cars in front of the main speakerphone while there would be a more extensive menu at the speakerphone. The billboard menu would display the restaurant’s popular items (based on rating or volume) and it will be less clustered and confusing. This should help customers pick their items based on popular choices.

Prototype Testing 

Since prototyping for an actual drive-thru was not possible, we were not able to test at a drive-thru. Instead, we decided to simulate the drive-thru experience by using a long hallway as the ordering lane and wheeled chairs as the cars. The billboard menu was placed above the doorway that led into the hall we were going to use as the drive-thru lane, since it was the place that made the most sense. As for the extensive/main menu, we counted three “cars” after the billboard and place it there, which made it at about a fourth inside the hallway. 

For the actual testing, we brought back stakeholders that we have previously interviewed in the first stages of the project. These stakeholders and new interviewees are all people who regularly go to drive-thrus. Some members of the group tested with multiple people at a time to simulate real traffic, while others tested with one person at a time, but still simulated traffic with empty chairs. We advanced each “car” about every 30 seconds - longer felt too long and awkward and shorter wouldn’t be a reasonable amount of time compared to real life waiting. 

1st Round Testing Feedback

○ Forgot what was on the billboard

○ Distinguishing certain items from others

○ The order in which they were in line mattered: People in

   the front felt rushed/People in the back were more prepared

Second Iteration

After our first round of testing, interviewees mentioned that they had trouble recalling the items from the billboard menu once they reached the ordering window. Using these recurring data, we decided to add a smaller version of the billboard menu above the extensive/main menu, which would also be the same length as the main menu.

Since our titles were color-coded, another suggestion was made to make the color coding even more distinctive, so we outlined the boxes with the color of their titles. Finally, it was also suggested to add a “featured” item to showcase the less popular items but still high rated items. We decided to shrink and split the sides box and we added the featured item box on top of it. We also color-coded it red to make it more salient.

Interviewees thought that it was an important factor to have in case customers forgot what the popular choices are or if they didn’t have enough time to look at the first billboard menu in the first place. However, they weren’t too fond of its location. They suggested having it on the left instead, so the menu could still be read left to right. They also thought that it is on top of the main menu would be too difficult to see. 

Final Iteration

We were able to create high-fidelity models of our prototypes. This includes a digitized version of both menus where all the item pictures and names are shown more clearly and cleaner. We moved the second billboard to the left of the main menu as suggested. ​​​​​​​

The Billboard Menu

The Extensive Menu

Finally, a physical model of the whole drive-thru was also made to give a better visual representation of how the menus will be laid out in an actual drive-thru. This was made with cardboard, hot glue, paint, and hot wheels.

Reflection

One of the biggest challenges we had was finding stakeholders that are willing to be interviewed while they are at a drive-thru. Although In order to gather contextual data from our interview, it is optimal to observe interviewees while they are in the car at a drive-thru, it is also hard to get permission to enter people’s car without proper incentives, especially at the drive-thru where most people are in a hurry, and they come and go really fast.

Another roadblock was at our testing stage. Due to the nature of our project, it is almost impossible for us to build a prototype with all the components of drive-thru in the time given. We were also struggling about how and where should we set up our prototype, and in what way should we test our prototype with our stakeholders. 

Overall, this project is a good stepping stone for knowing more about the HCD field. We followed the basic rules of doing HCD, and finished this project in 10 weeks. We could do better in the prototype testing section since we were kind of rash and out of time in the end. We should get more feedback on our prototypes and our final design solution. So always start early and start often. HCD is not a linear process, we always need to go back to previous sections like user research and testing to validate a new idea or new scope.

We were able to create high-fidelity models of our prototypes. This includes a digitized version of both menus where all the item pictures and names are shown more clearly and cleaner. We moved the second billboard to the left of the main menu as suggested.

Special thanks to my teammates: 

John Yoon 

Jiashi Feng 

Cecile Pham

Jonathan Wang

Copyright © 2020 by Janice Lixiang Zhou