ContextualInquiry-Group:Whoo-Whoo...iz liek an alarm clock
From CS160 User Interfaces Fa06
Contents |
Contributors
Heung Tai
Anton Mikhailov
Target Users
Person A: Student at UCSC, philosophy/linguistics major. Relaxed and easygoing.
Priority: Pleasant experience.
Likes: Human attention. Choices. Human interaction. Understandability. Long walks on the beach.
Dislikes: Technology (considers unreliable and confusing).
Person B: Restaurant chef. Previously worked at another place which did not have a computer system installed. Complained about bad handwriting of waiters there. Noted that often chefs get blamed because of miscommunication between waiter/cashier and customer.
Priority: Understandable orders.
Likes: Clear text. Easy orders.
Dislikes: Bad monitor quality. Bad waiter handwriting. Picky customers.
Person C: Working mother, often goes to fast food to pick up food for children. Children are often disagreeing on what they want.
Priorities: Keep kids happy, and get out fast. Safety.
Likes: Quick service. Consistency. Easy ordering. Simplicity.
Dislikes: Confusing technology. Complexity added to simple tasks.
Problem and Solution Overview
Fast food restaurants often have unpredictable customer flow/wait times as well as poorly trained cashier staff. We aim to solve the problem by introducing an automated, staff independent ordering system using the Anoto pen. Customers order their food by filling out a form, and pay the order with a self-help pay station like in some grocery stores today.
Contextual Inquiry - Interview Descriptions
We chose fast food restaurants as our target for analysis of the system. Customers there are often waiting in line and have time to talk about their thoughts on the process. At first we went to McDonalds, but found that there were too few people and the staff was not too helpful. We then turned to Jack in the Box where we got our data.
Right away, we noticed (as expected) many people waiting in line without looking at the menu, which implies they have already picked their item. The ordering procedure was very systematic and repeatable since most customers chose by numbers, although there were many requests for omitting condiments. The order process itself was pretty fast, and the slowest part of the interaction was the change given by the cashier.
All customers had roughly the same ordering pattern. They first picked a number, followed by the items they wanted to add or omit. In the case of multiple orders, things usually got confusing for both cashier and customer. The feedback from cashier is slow and unrecognizable so the customer often feels the need to clarify the order. This slowed down the queue dramatically, when the order was even a little bit wrong, since the cashier had to manually re-enter the order.
During the customer interviews, we started off by asking if the ordering process was pleasant to the customer. We then asked how often they come to this restaurant and place orders here. We asked about traffic, delay times, and the quality of the staff. We then asked if they had any suggestions on improving the process, only then introduced a vague idea of our system. We then asked the customer again, if they could improve the process but now with the pen as an extra tool they can use.
For the chef interview, we simply asked the chef to go over what he felt was important, and then presented him with our idea. Because of limited time (he was on break) we could not talk much longer before we introduced the pen for critique.
Person A Person A was not a common customer at fast food restaurants. She said that the wait times do not bother her, as she is usually not in a rush. If the restaurant is too packed, however, she will most likely go elsewhere. She had no suggestions on improving the current system, since she didn't seem to dislike it. When told about the Anoto pen and the automated menu, she found the idea terrible. She had a very strong opinion about this, notably about the fact that it eliminated human interaction, as well as reduced available jobs.
Person B Jack in the Box did not let us into the kitchen because it is a health hazard by regulation. We observed the chefs from behind the counter and were able to talk to one while he was on break.
Person B noted that computer displays were much better for displaying orders than handwritten ones. He mentioned that cashiers still get the orders wrong, but its less frequent and he doesn't get blamed as much. The computer system did take a while to get used to. He felt that the idea was interesting, but found it more of a novelty than useful. After we mentioned that it would give more precise orders he seemed to respond positively.
Person C Person C seemed nervous. She was preoccupied with thought when we contacted her. Notably, she mentioned that she did not have much time and that we would have to be quick. This showed that she is trying to get out of the restaurant fast. She said often times she would have to write down orders from her kids which were long and complicated, and it would be easier if she could just give the cashier the paper with the orders. When we told her about the pen concept, she felt that it was a great idea and asked if she could somehow get menus to give to her kids at home. We said that it is feasible, yet complicated if kids were at home, but reasonable for drive through or in restaurant. She was interested in this idea, but would like to see something she can fill out before she comes to the restaurant, and which she can turn in for faster ordering. Like an "Express Line".
Task Analysis Questions
1. The main users of the system are cooks, customers and cashiers.
2. Customers choose the food and pay for it. Cashiers take the order, translate it to cook language, take payment, and give change. Cooks interpret the order, and make the food.
3. Automatically count coins given by customers. "From home" ordering system.
4. Customers learn by experience how to order by number. Chefs learn from experience and training how to interpret this restaurant's ordering abbreviations. Cashiers learn from experience and training which order options are available, and how to ring them up. Items often not found on the menu have to be learned from more senior employees.
5. Customers order at the counter. Cashiers work at the cash register. Chefs work in the kitchen.
6. The user has the data, and must transfer it through the cashier to the chef. The user is very familiar with the food they want and what number it corresponds to.
7. Maybe paper. Maybe pen. Maybe keys. Finger to point.
8. For multiple orders, people scream incoherently at each other deciding what they want and who will order it. The process is chaotic and confusing, which slows the line down dramatically causing further commotion in people standing behind them.
9. Typical order is roughly 1 per 2 minutes. These are performed daily, at variable frequencies throughout the day.
10. There is no "real" constraints aside from customer frustration and food going bad (very long).
11. Screaming. Supervisor is usually called. We did not witness this (unfortunately). Food is rejected and most likely thrown away.
Analysis of Tasks
1. Choose the food. (MODERATE)
2. Order the food. (HARD)
3. Pay for the food. (HARD)
4. Cook the food. (MODERATE)
5. Pick up the food. (EASY)
6. Eat the food. (EASY)
Interface Design
The design must use the users previous knowledge from ordering off common menus. We must also retain the customizability present today so users don't abandon the system. This is why we decided to keep the large menu present in fast food restaurants and create a smaller version, an order card, for the user. Each order card would come with a semi unique (say modulo 100) number which will be used to pay at one of the payment machines. Once the payment is accepted, the order is cooked and the number is called for pickup. The order paper may also be used to double as your tray paper to conserve resources and recycle.
Order Card Prototype
Note the global information such as order number and For Here/To Go option.
The top portion of the menu consists of combos listed on the big overhead menu, while the lower one has part of the dollar menu and other ala carte items you may order. To get a big mac ala carte, you request 1 Big Mac and circle NO for the Combo section. One issue with this system is that as it stands there is no clear way of indicating 2 Big Macs, 1 with cheese, one without. This is a design issue. If we give a more general form, one where the person fills out the number of the item they want (say 1 for Big Mac, 32 for cheese burger), we must place the Options on each line. This may get confusing, since all lines must be identical, a customer might think that if they do not cross off the pickles on a fish sandwich, they will get pickles. This would be ambiguous, are the "modifiers" only applied to what is actually on the sandwich or are they a global sandwich configuration? The way it is listed now, the ambiguity is gone, since the options are unique for every item. However, this limits the generality of the entire system. This decision is still under debate, and we will have to ask more people and take further data to see how the customers want it resolved.
Small and large fries etc were separated out to avoid confusion and also add functionality. If a person say wanted 3 large fries, and 2 small fries, we could not simply do this by having a type/quantity/size line for "Fries". In this setup, you can easily order the above. This issue is similar to the one described above.
In this menu, I am requesting 1 Medium Big Mac combo, no pickles, no cheese.
2 Fish sandwiches, alone, no cheese.
An extra small fries.
2 large fries.
A large soda.
And an ice cream.
Analysis of Approach
The main affordance of the Anoto system over others in this context is its durability and simplicity. The pen has to be durable since we are catering to large masses. TabletPC or PDA's would be too expensive and fragile for our audience. These are also much more prone to theft. The Anoto pen comes with a digital signature, which we can perhaps rig to only work for restaurants, making it useless outside (disable user upload?). We can also chain the pen down much easier than a PDA and it is much less desirable of an object overall. The pen is much less prone to system failure, than a PDA or a TabletPC. Smart boards are not versatile enough for our cause, and would produce incoherent orders because of the amount of handwriting recognition needed.
