Our group wanted to create a system that could make the interaction between customers, the kitchen, and servers smoother and more efficient. In our first prototype test, we hope to see whether our new interface is intuitive to use and actually improves efficiency, or whether it introduces too much information to servers and confuses them.

Mission Statement:

Our goal for this project is to create a system that can aid the work of servers in a restaurant by easily providing them information on the state of the tables they are waiting. This information could help them make better decisions of how to order the tasks they complete, and complete those tasks more efficiently.

Avneesh, Kuni, and Yaared created the  rough prototypes.

Joe and John reviewed and improved the prototypes. They also wrote the task descriptions.

All members participated in documenting the prototypes and writing up the discussion.

PROTOTYPE DOCUMENTATION

The Motherboard displays all the statuses of the tables in a given section of the restaurant. The tables are arranged in the floorplan of the section, where each table has indicators for cup statuses, a two column list table of all the orders, and a timer for how long since the order has been placed. For cup statuses, we have 3 lights–green yellow and red–that correspond to full, half-full, and empty, respectively. A number beneath each of these lights indicated how many cups are in each state. Our table of orders highlights each order as either green or red, depending on whether the item has been prepared or not, respectively. When the item has been delivered, the entry becomes a white box. Finally, there is a timer for every table that notifies how much time has passed since the orders were placed. If a table has had all green items for over 5 minutes, the table itself turns red, indicating that the food has been sitting for a while. Our coasters were simply cardboard squares for the moment.

The start screen for entering a new order

The screen to enter the details of an order.

The screen to enter comments for a particular item ordered

The confirmation screen for changing an existing order

The confirmation screen for canceling and order

The confirmation screen for completing an order and sending it to the kitchen.

The is the device through which servers input orders. It has three functionalities – making new orders, changing existing orders, and canceling existing orders. The interface is a touch-display screen. The home screen simply allows users to pick a table and then either “Make a new order” or “Change the existing order” for that table. The order screens respectively display the “Current order” (+ any comments), “Menu”, and the “Make new/change order” and “Cancel order” buttons. For a server to add something to an order, he or she must simply ‘flick’ an item from the menu to the left (this will propel the item to the left and add it to the current order, for more than one simply flick again). To delete an item from the current order, simply ‘flick’ it to the left again so it is out of the order (if there are x2 of an item in an order flick left twice to get rid of both items). Servers can also add comments (e.g. well done, spicy) by simply pressing the comment box next to each item in the current order, which navigates to a comment screen with a keyboard in which you can attach or cancel comments. Once the order is done press “Make/change order”. If an order wants to be canceled simply press “Cancel order”. serverCenter is set up so that we don’t run into consistency issues with information in the kitchen center/mother board.

The help button

The help button after calling for help.

TASKS

EASY – Calling for Help

Yaared is serving a customer and see’s that another one needs to be served.

Yaared calls for help from another waiter.

Avneesh sees the call for help on the Motherboard.

Avneesh helps Yaared’s other customers

Job well done team!

Yaared calls for help from another waiter.

Avneesh sees the call for help on the Motherboard.

Avneesh helps Yaared’s other customers

Job well done team!

When extremely busy and unable to ask another waiter for assistance, a waiter may request help by pressing a button on his or her handheld device.  The other waiters then receive a notification on their handheld device that the given waiter needs help.  If the solution to the problem is obvious, a nearby waiter can address it directly.  Alternatively, they could consult the motherboard to see if anything is amiss with the busy waiter’s tables.  If the problem is less obvious, they can at least come to the vicinity of the troubled waiter so that they can receive instructions directly.  The original waiter can then clear the help requested status when the problem has been resolved.

MEDIUM – Checking customer status

The motherboard shows the “state” of each table so the server can infer customer impatience from this low level data.

A medium-level task that a waiter may have to perform is checking to see which tables are waiting on orders, have drinks that need refilling, or may be requesting attention directly. Currently serving staff in restaurants needs to physically go over to the area in question, survey all tables in detail, and then report back to the kitchen for what they need. The prototype app makes this task nearly trivial. Each table’s order data and drink levels are displayed on the prototype screen, along with the amount of time the group has been at the table. The user can see easily the number of drinks that need refilling by looking at the red and yellow lights. If a group has been waiting at a table for a significant amount of time without ordering food, this will also be clearly visible because the table will change colors.  All of this data can be easily monitored from a central location instead of by manual survey.

HARD – Determining task order

Determining task order is made much easier when there is information to see what issues are urgent.

Perhaps the hardest task of the waiting staff is just to determine in what order to complete the various other tasks to be completed. Different tasks take different amounts of time, and it is often difficult to complete them in an order that leaves no patron waiting for too long. The most time-consuming of these is actually bringing out the food – since a waiter has to estimate how long it will take for the food to be prepared and may waste time standing around to wait for it – or, in the opposite case, miss it when it comes out since he or she is performing a different task. In the prototype, the presence of red and green highlighting underneath the food ordered at each table shows whether or not it is ready. Using this system, the user will no longer have to waste time going back to the kitchen to check, as the info is right in front of them on the prototype. This data, combined with the easing of checking customer status, should provide the user with an easier way to determine a task order, and provide a greater margin of error for a sub-optimal task order.

DISCUSSION

We created the layout for the motherboard on Adobe Illustrator, and all other parts were constructed from paper (the coasters were cardboard squares). Using Illustrator was a new technique, but it didn’t take too much time since we had a group member who knew how to use it. The most difficult part of our prototyping was figuring out a good interface for the input of the ordered items. We did not want to make the interface complicated and slow down the wait staff, but we wanted to be able to log enough information so that the Motherboard would be effective. We also did not want the waitstaff to have to record orders twice, once at the table and again inputting into the system. To solve this, our system could use a dedicated employee to take the waitstaff’s order notes and then enter them into the system. This way, the waitstaff have no “down time” where they can’t be serving or on the floor, and the pattern breaking activity of order entry for all waitstaff is concentrated in a single employee. We thought organizing the table information on the floor plan worked well for the motherboard, since we are presenting the information in a layout that people are already used to. Color coding various signals also provided a very simple way to signal certain information. We feel mix between textual and color information prevents the motherboard from becoming too cluttered with text.

OBSERVATIONS

1. Student 1 was observed in Frist, sitting in a long (>10 minutes) gap between classes. He was browsing the internet on his laptop. When asked what he generally did between class in short periods (10 minutes), he said he only had time to walk to classes. When asked what he did during long periods, he said he generally tried to get homework done.
2. Student 2 was observed in Frist, sitting in a long (>10 minutes) gap between classes. She was reviewing class notes. When asked about activity during long breaks, she said she tried to do classwork. When asked about short breaks, she said she didn’t have time to do much, but would sometimes try to look over notes before a precept.
3. Student 3 was observed in a long break before classes. He seemed to be walking around looking for people to talk to. When asked about what he did during long breaks, he generally said he went back to his room to either work or hang out. When asked about short breaks, he said he only ever had time to go to class.
4. Student 4 was observed in a long break in Frist. He was reading over notes for a class. When asked about long breaks, he said he generally tried to get school work done. During short breaks, he said he didn’t have time to do anything except get from class to class. He also added that during “mid-length” breaks (~30 minutes), he struggled to find anything productive to do.

All students, when asked about what they did during the short period, seemed to be mildly surprised by the idea of actively trying to get any errand or task done during the short interclass period. All also mentioned that it was sometimes a struggle just to get from class to class, much less attempt anything during that time. During long breaks, all students observed seemed to have a plan to be productive, but during short breaks, there was no attempt. All students said that concentration was an issue in the long breaks. The salient problem for long breaks seemed to be keeping students productive and focused within that limited time frame, while the salient problem during short breaks seemed to be simply getting students from class to class on time while allowing them to do anything else.

BRAINSTORMING

1. An app that counts the number of people in the lecture hall of your next class, letting you know if you have to rush to get a seat.
2. An app that lets you know how far your professor is from the lecture hall, letting you know whether you need to run to class (helpful in small classes/seminars).
3. An app that, given two classes and a destination that must be reached in between, gives the shortest path.
4. An app that allows you to access the notes for your next class with a single button click, and updates itself without user action between classes.
5. An app that displays your upcoming homework due dates graphically
6. An app that shows you the number of people waiting at different coffee vendors around campus.
7. An app that lets you know whether your friends are in your next class or not (so you can skip and get their notes)
8. An app that, given a syllabus at the start of the semester, updates you on the readings/lecture topics each day.
9. An app that lets mail services know you are coming for a package so they can find it for you (sort of like a fast-track line)
10. An app with games to warm up your fingers on cold days.
11. An app that maps your activities for the whole day (classes, meals, extra curriculars)
12. An app with an “on the way” signal that is sent to you professor when you are close to the lecture hall (so the professor can decide whether to wait for stragglers).
13. An app that gives you a spot to meet friends so you can walk to your next class together.
14. An app with an onscreen arrow that points in the direction of your next class.
15. An app that signals when you pass close to a coffee vendor so you don’t forget to get a cup in a rush.

PROTOTYPES

QuickNote – This idea was chosen because 10 minutes is plenty of time to review notes on the way to a class, but current tools don’t make note-viewing on the go instantly accessible.

The schedule editing screen.

The schedule editing screen (accessible on phone or web) allows users to enter their class schedule into the app. Since the app knows when the users classes are, it can provide the notes for the next class without user prompting, ideal for note access on the go.

Note entry screen

The note entry screen (accessible by phone or web) allows you to add notes to the app to display later. Accessibility by computer allows you to easily take notes on the computer and copy and paste them into the app. Extensions for a desktop application could be made that would automatically read a directory, parse the files contained in the subdirectories, and update the notes on the app.

Users first select the class they want to add notes for, create a new note and give it a date,  and enter the note. They can also view older notes for a class and edit them.

Note viewing screen.

The note viewing screen (the default screen on the mobile app) automatically updates to show the notes for the upcoming class. Since the app knows your schedule, opening the app from a phone will automatically bring you to the notes for your next class first, thus minimizing use time. Users can also select notes by class and date if they do not want to view the notes of the upcoming class.

——————————

PathFinder – This app combined the various map based ideas from brainstorming to create a tool that could allow students to get from class to class quicker, while giving them the information that would allow them to efficiently perform simple errands (such as stopping by a different building, getting coffee or picking up packages) in between.

Home screen

The home screen allows the user to pick what service they want. They can

1. Edit their schedule
2. find the shortest path from class to class while detouring through another location
3. find the fastest place to get coffee
4. see wait times for picking up a package and join the “fastpass” package line
5. and find the shortest path from class to class normally.

Schedule editor

This screen lets you edit your schedule.

Detour editor

This screen lets you add a new stop to your detour path, and the app can reorder the stops to have the shortest path. It also allows you clear the detour path you have set. If you navigate away from the screen and come back, you can see what detour path you are currently on. After reaching a class, the detour path automatically clears itself.

Detour map

On the top of the screen is an arrow that gives you the straight line direction to your next destination. This feature is for those people who like to go off the normal paths and find shortcuts through campus.

Coffee finder

The coffee finder screen takes two things into account when finding the fastest coffee location. The deviation from the path to your next class, as well as the number of people in line at the coffee vendor. Taking into account these two variables, the coffee finder function provides the location that will provide the minimum delay. It also has the shortcut arrow at the top of the screen.

Package Picker Upper

Like the coffee finder, the package picker upper calculates the time you will be delayed in picking up a package by looking at the deviation from the path to your next class and the number of other people using the app that have said they are picking up packages. If you decide to pick up your package, you can notify the mail staff and they can take your package out of the stacks ahead of time, and you could be served by an express line. It also displays the shortcut arrow to your destination.

Next class map

The simplest feature, it shows you the path (and shortcut arrow) to your next class without any distractions.

TESTING AND FEEDBACK

User testing was conducted in Frist. Temperature and unwillingness of test subjects prevented the tests from occurring outside while walking from place to place. Instead testing took place in Frist, and I attempted to recreate the situation of moving from class to class by giving situational “events” (i.e. “you’re leaving your class. Do you want to get coffee?” or “Is there anything in another building you need to pick up.”)

The user felt confused by the choice of the word “detour”

The user found the arrow functionality helpful, preferring to follow it instead of the map.

The user suggested coffee should have a wait time display.

Five users were tested. All said without prompting that if such an app were to exist, they would use it, with three encouraging me to either make the app or propose it to the USG.

Insights:

Users differed in whether they were more likely to follow the arrow or the map, with 3 users preferring the map and 2 preferring the arrow. Care should be taken to present the arrow and map as complementary, so that users do not feel like the two are presenting conflicting information. In an app specifically designed to minimize decision making in stressful times, forcing an unexpected choice (whether to follow the arrow or map) introduces a huge inefficiency.

The choice of the word “detour” was confusing to some users, with one thinking it was a function that would provide a pastime to users with time to kill. The “add detour” wording was also confusing to some users, with some users being unsure whether pressing add would add a new stop along the path or create a new path entirely. In an app designed with a minimal interface for maximum intuitiveness/inefficiency, a few extra words could go a long way towards clearing up use (i.e. “add new stop to current path” instead of “add new stop”.)

User suggestions:

The class/time schedule, with locations coming from the registrar, could have problems when class locations are changed. Instead, the schedule should be location/time, so class changes could be easily fixed, and the app could generalize to all activities for the entire day instead of just classes.

Instead of just having a shortcut for just coffee and packages, a “common detours” menu should be available that would also include libraries and food vendors (dining halls, convenience stores, etc).

A time estimate display similar to that on the package menu should be added for the coffee finder.

An explicit back button and help/instruction menu should be added.