Pass 2 (Assignment 4B) AN ONLINE TRIP PLANNING SYSTEMcourses.ece.ubc.ca/.../EECE418_Assignment_4B_Team_7.pdfEECE 418-201 GROUP # 7 Pass 2 (Assignment 4B) AN ONLINE TRIP PLANNING SYSTEM

EECE 418-201 GROUP # 7 Pass 2 (Assignment 4B) AN ONLINE TRIP PLANNING SYSTEM

Project Team: Rocky Chan Gurpreet Virdi Hootan Ziyaeimatin

Submitted to: Dr. S. S. Fels

Instructor for EECE 418-201 Electrical and Computer Engineering University of British Columbia

Issued Date: March 29, 2006

Version: 1.0

1

A1: Redesign Rationale • What did you learn from Pass 1

Pass1 was a great exercise in heading towards our final design decisions. We learnt two key things, with regards to our design, as follows:- o Our current interface is supporting a very broad category of users. Trip planning

is a topic that is heavily influenced by two factors that are the type of user and the purpose of the trip. Therefore, at this stage it is very important for us to limit our users and make a custom website for the one major type of user like a businessman or a tourist. One website cannot cater to all. This is a very important lesson that we have learnt and this was one factor that was making design decisions very difficult and complicated.

o The current interface seems to give a final schedule that is completely inflexible. One thing that we learnt from performing a lot of evaluations, formal heuristic evaluation and informal walkthroughs was that users don�t like to be confined. They always need flexibility. Therefore the next version will focus more on flexibility besides a schedule oriented fixed result.

• What is good and worth preserving about the current approach o We think that our current approach of designing this software is good. Trip

planning is a very broad topic and we have been able to reach this stage of finalizing our design and having a clear idea of requirements by frequently involving users and brainstorming design decisions amongst us. We will continue this approach. Since evaluations were very central to our approach, a lot of key design aspects found support at several levels of evaluations: ! The first page seemed to have a unanimous vote of clarity and the idea of

having a row of the main options seemed good. ! The idea of a wizard based form was very intuitive to the users. ! The feature of �Add to Cart� was another familiar approach to users who

knew how to use it and add/remove their options from the cart. ! The major concept of having a first set of options that would allow the user to

enter his/ her personal constraints, a second set of options generated by the software that would allow the user to choose from what was available and a final step of giving a detailed schedule seemed to be the intuitive way to plan the trip when a user was exploring a new place(s).

! Several cosmetic things like color coding, placement of links on the top right corner, screen partitioning (options on left and details on right) etc. followed the user�s intuition.

• What needs to change & is it a completely different direction The three major changes that need to be done at this stage are: o Flexibility in the result and final schedule. This is perhaps the most important

change needed in the current interface. The new interface should be able to support a flexible along with a rigid schedule.

o Some major cosmetic changes need to be done to the interface such as the placement and naming of buttons and links.

o The focus of this interface should be towards a limited target audience.

2

• Critical questions to resolve the uncertainly The two key questions that have come up with regard to our interface are: o 1. What type of result is the user looking for? A fixed schedule, a schedule he can

modify or just a tour guide of key places? o 2. Some basic cosmetic questions such as about using pop-up windows, scroll

bars, color coding, horizontal versus vertical panels/ divisions, naming conventions and placements of buttons etc.

• Types of evaluation to be used We are planning to use the following evaluations: Qualitative Analysis o �Think Aloud� observations & retrospective interviews on our new prototype:

This would help us answer critical question #1 and #2. Both of these critical questions involve the user either making a choice between completely different approaches (rigid versus flexible schedule) or thinking intensively about comforts with respect to the cosmetics of the interface (segregation through panelling or no panels at all). These kinds of questions involve a lot of thought process, and one of the best ways to analyze the thought process is by the �Think Aloud� observation.

o Questionnaires- This would help us answer critical question #2 about cosmetic choices because these kinds of questions usually have a yes/ no type of answer.

Quantitative Analysis o Controlled Experiments � This will be the quantitative evaluation that we will

perform. This will be used to compare the design alternatives we construct at this stage, the pass 2 semi functional prototype versus the Low-fi prototype we had and several performance measurements. More has been detailed out in Section B1. This analysis will critically help us in answering the critical question #1 on the type of result the user is looking for. We will show the user the old �inflexible� result of the Low-fi Prototype and the new flexible result of the pass 2 prototype and do a statistical analysis.

• Type and extent of prototype to be used During Pass1, due to the complexity of this project, it had become important for us to develop some concrete designs to help us visualize our project and we did it in C#. We will now focus on developing a completely interactive semi functional prototype lying between a medium and high fidelity prototype (does not have all features but has sufficient features/ cosmetics/ interaction). This also follows from the fact that we pretty much know our main requirements and would now like to focus on finer navigation details and some design changes. We are planning to use a mix of horizontal and vertical prototypes. The horizontal prototype part would show the user all the key options like inter city planning, wizard, cart, final result, and search. But these options shall only work for limited pre-defined values like: the search and trip planning would only work for a fixed set of one or two cities. This way the user would get to see all the possible ways he could use the interface yet being limited to fixed results which at this stage are not as important as being able to navigate the interface and get a feel for all the available options. Since a high level of interaction is needed and this is a web based interface we will use Microsoft Front Page website tool for our prototype. Also most of us in the team are familiar with this tool.

3

A2: Additional Analysis and Evaluation We carried out the following informal user evaluations, using qualitative analysis like �Think aloud� observation, retrospective interview etc., with three users on our Low-fi Prototype to help us judge the user�s mind and extract some more suggestions to our interface before we started work on the Pass 2 prototype. User 1: Sam • Sam started from the home page of the website. His reaction to the home page of the

website indicated that he had a good understanding of the intended functionalities of the website. However, he seemed a little bit unsure about the functionalities that the �Intra-city� option would have to offer.

• Sam proceeded by selecting the �Inter-city� option. • He was directed to the next page where he started getting a little bit confused. • Asking him to �think aloud�, he mentioned that by seeing the three steps provided at

the right side of the page he assumes that there should be three main steps for this task. However, he wasn�t sure if that meant he could skip any of the steps and jump to another while he has not yet finished for example the first step. If some steps are optional there was not clear for him which ones are optional and which are mandatory.

• He was not sure on what grounds we were asking the user to have at least $10 for the trip while the actual budget needed may be way more. Why not giving the total budget at the end let the user decide if the amount is within his or her budget.

• After filling out the necessary information, again Sam was not sure what will happen if he would press the �Finish� button as opposed to pressing the �Next� button. He would guess that pressing �Next� would take him to step 2, but he wasn�t sure what pressing �Finish� would do. Would it skip steps 2 and 3? Does that mean they are optional?

• Trying to explore other services provided by the system, Sam pressed �Next� so that he was directed to the next page (step2).

• He found it really strange that he had to exclude Landmarks/Categories. He expected to include them rather than exclude.

• Sam was asking why the system is only offering one option for flight. • Similarly, he was expecting more options for accommodations. • Sam was not sure what �$150 x 2 �meant under the �Rate� column. • After trying all this, Sam wanted to submit his request but he was absolutely confused

as whether he should press �Submit DEFAULT: Plan my trip� or �Submit: Plan my trip.

• He was not sure what each button does and what the difference is between the two. He tried to guess that by reading the line beside the �Submit DEFAULT: Plan my trip� button which says: �*Note: This will not submit your changes� but he got more confused.

• After doing a �post-test interview� (retrospective interview), Sam mentioned that the system seemed to have all functionalities needed for such a system but its features are not as clear.

4

User 2: Rob • Our second user, Rob had more or less the same problems as our first user Sam plus

he mentioned some additional useful points. • Rob found it more efficient if the system would ask the user of the purpose of the trip

so that different features could be offered based on the purpose of the trip. He reasoned that the needs of a businessman and a tourist are completely different even though they are taking the same trip.

• During system usage, Rob mentioned what if the user wants to go back and forth to make changes and/or to make corrections to the previous data entered. (no �Back� or �Previous� buttons).

• In addition he mentioned that the system does not actively show the path that the user has taken so far.

• He suggested that a path like the following could be helpful: Home--> Inter-city plan--> …etc…

• Rob was not sure if the help provided for the users was enough. He asked for example what happens if the user clicks on the �help� link. (i.e. what kind of help is provided?)

• Rob also suggested that it would be useful if the system could offer a �tour guide so that the user could actually print it out as use it as reference.

• When he was looking at the schedule offered by the system he noticed that the schedule was so tight/ rigid. What if the user is just a tourist and doesn�t want to adhere to such an inflexible schedule.

User 3: Craig • The very first thing that Craig pointed out was that it is highly inefficient to have

a drop-down menu containing an entire country's city selection. A suggestion he made was to break the list down according to different provinces/states.

• He pointed out that the system does not provide any feedback as what it will do when �Finish� is pressed prior to the last Wizard step.

• He also pointed out that certain buttons and words like SUBMIT DEFAULT were confusing and he didn�t really understand what �DEFAULT� meant.

• The system does not provide any feedback as what it will do when �COMMENT� is pressed. (i.e. what does COMMENT mean in this context?)

• Craig mentioned that why the system limits the set of options for the accommodation (i.e. no resorts and hostels options for users with limited budget)

• What functionalities does the SEARCH option offer? • In the middle of system usage Craig decided to make a phone call. Therefore he

decided to save the session and exit by selecting �Save and Exit�. However he correctly noticed that on the home page of the website there was no option to resume an already saved session.

• After playing around with the system for a while Craig decided to plan a multi-city trip, but there was no option for that. (i.e. he was restricted to only one city)

• He could not find the difference between �home� and �back to home� links? • He also mentioned that there was no option for users whose goal is to fly on a

specific class (i.e. economy, business, or first class)

5

A3: Prototype Illustrations The sequence of events for the figures is as below:- Figure 1 is the main home page that displays the two major options of inter city planning and individual search. Figure 2, 3, 4 is the three step wizard that follows after the user selects inter city planning from the main page. Figure 5, 6, 7, 8 is the summary of user�s choices and list of options for hotels, flights and attractions that the user sees after he completes the wizard. Figure 9 is the final rigid schedule that the user sees once he submits his options in the preceding step and clicks on the button labelled �Show Schedule�. Figure 10 is the tour guide which the user sees when he selects �Show Tour Guide� after step 9. Figure 11 is the search screen the user is taken to when she selects search from the main page in Figure 1. Figure 12 is the About iTravel screen which has information about the iTravel company Figure 13 is the Help screen in the form of an FAQ

6

Figure 1: Main Page

7

Figure 2: Inter-city Trip Planning � Wizard Step 1

8

Figure 3: Inter-city Trip Planning � Wizard Step 2

9

Figure 4: Inter-city Trip Planning � Wizard Step 3

10

Figure 5: Main options

11

Figure 6: Flight options

12

Figure 7: Hotel options

13

Figure 8: Attractions options

14

Figure 9: Rigid Schedule

15

Figure 10: Tour Guide

16

Figure 11: Search

17

Figure 12: About iTravel

18

Figure 13: Help

19

B1-Evaluation • Hypothesis

H1: Users can plan their trip on iTravel more efficiently than Expedia. H0: Users can plan their trip with equal efficiency on both iTravel and Expedia.

• Subjects: A set of 8 users who are experts at using internet but have never used an

online trip planning will be used for this experiment. Most has been detailed out in section C2.

• Dependant Variables: Time to plan a complete trip and number of clicks to plan the

trip. Time to plan a complete trip will measure speed and number of clicks will measure clarity and preciseness of interface. Speed and clarity/ preciseness are the two major components of efficiency. This gives rise to two sub hypothesis: Sub-Hypothesis 1: H1: iTravel takes less time than Expedia H0: There is no difference in time between iTravel and Expedia Sub-Hypothesis 2: H1: iTravel takes less # of clicks than Expedia H0: There is no difference in the # of clicks between iTravel and Expedia

• Independent variables: iTravel pass 2 prototype and Expedia (http://www.expedia.ca/daily/enc4105/packages/default.asp?rfrr=-47326&lnkcid=15) . Both the interfaces provide the two major functionalities of individual search and trip planning. The user will be asked to plan a trip on both interfaces. As part of trip planning, some users would want to find a destination before planning a trip to that destination. This would automatically test the individual search.

• Nuisance Variables

1) The individual differenced between people 2) Some users know their destination and other�s don�t thereby requiring that extra

step of search in some cases. 3) Questionnaires in between the tests could distract the user somehow. But they are

essential and cannot be eliminated. • Tasks:

1) Fill up a pre-questionnaire on years of experience with the internet, current methods used for trip planning, some personality trait questions (related to the demands for a rigid schedule versus tour guide), time the person wants to spend for trip planning etc. These questions will be used for nuisance variable evaluation later.

2) Then the user will be taken to the usability lab in which we will book 2 cubicles. One will have a computer setup with Expedia and the other computer will run iTravel.

3) 4 users will be taken first to Expedia cabin and the rest of the 4 users will be taken to the cubicle with iTravel each user using it at one time.

20

4) Then the users will be asked to plan a trip to a destination on both the interfaces. At this stage the user will not be prompted anymore and he/ she will use the interface from scratch.

5) The design team will stand a distance behind the user and make data recordings. 6) The user clicks and timing will be measured by us using a stopwatch and

observing the clicks respectively. 7) After all the 8 users have completed their tasks, the post questionnaires will be

given to them containing closed objective questions about their experience, confusions etc. These results will be quantified later. This questionnaire will also give the users a little break between using both the interfaces.

8) The users will be swapped i.e. Expedia users will now test iTravel and iTravel users will test Expedia. This will now allow us to collect data for within-subject evaluations. Because there are exactly half the users starting with Expedia first and exactly half the users starting with iTravel, the contamination of users will be negated. Some more on contamination will be mentioned in C2. Like before time and clicks will be measured again.

9) After the user has done and tested both the interfaces, the users will be given the same post questionnaire as before (but for this interface and not the old one) that will focus primarily on asking questions on their last used interface. Again these will be quantitative close ended questions that will be used in analysis of results

. • Analysis:

After we have collected the number of clicks, times and questionnaire we will perform a statistical analysis on the data. Within Subjects A set of double means, standard deviation and variance will be calculated for both time and # of clicks from the data collected in stage 8 for all the 8 users who would have all used both the interfaces by this time. We will be conducting the two tailed t-test for each set of data to calculate the confidence intervals for each set of data. We will be using 95% confidence interval which is reasonable high and most commonly used.

• Randomization:

The two major segregations that we have in our users are as follows: 1) Half the users we have chosen know their destination and the other half need to

search for their destination. That extra step of search for destinations might lead to differences in timing calculations.

2) The results might be influenced by which set of users use which interface first i.e. Expedia or iTravel.

In order to avoid #2, half the users will use Expedia first and the rest of the half will use iTravel first. In order to avoid #1, the users who know their destination will be split equally amongst these two groups. More will be detailed out in the summary of the results.

21

C1-Subjects: Our subject pool consists of expert internet users who have never used an online trip planning tool to plan their trip i.e. novices in trip planning. Since iTravel is intended for tourists, our pool of subjects will fall under the following major categories: • Mostly students within the same age, education and experience. • Self employed people looking for a short vacation planner. We will be using 8 users. 4 users will start by testing Expedia followed by iTravel and the rest will test iTravel first. We will be using �within-subject� methodology. More will be detailed out in C2. To make the results consistent users who know their destination and users who don�t know their destination will be split up on both systems. Here are the people we have tested:- 1) Tana is a 19 year old computer engineering student who is desperately planning her

summer vacations in May 2006. She likes to go on the mountains and loves snow boarding. She would like to plan a vacation somewhere in the mountains but wants to restrict herself to Canada and does not know where to go.

2) Suzy is a 21 year old electrical engineering student who is graduating this May. She has 10 days of vacation and a fixed budget of $1000 and she wants to plan a trip with four of her friends to Whistler. Till now she has been using tour guides for site seeing and has used her contacts to ask about hotels etc.

3) Kitty is a 21 year old electrical engineering student. Kitty wants to travel with her family. Kitty is a very methodical person and she likes a fixed output or result that she can rely on and follow blindly. Kitty is new to Canada and wants to plan a trip for her family somewhere near about. She needs to search for her destination first.

4) Jenny is a 20 year old girl who is in third year electrical engineering. Jenny is very picky by nature and very unpredictable. There are some days that she needs to know exactly everything hour by hour and there are days that she just wants to do everything her own way. Jenny wants to plan a trip to Montreal. She has been trying to plan this trip for months, but finds that maps, asking people for information takes too much time and wants to have a tool to plan her trip.

5) Betty is a 28 year old self employed architect. Betty has a one week holiday from May 1, 2006 to May 7, 2006. She wants to travel around BC.

6) John is a 26 year old self employed accountant. John wants to plan a trip to New York. He has a fixed budget and a three day long weekend to go around New-York. He has two restrictions. He likes to use Canadian dollars for all his budget (which he likes to manually calculate himself) and has already purchased his flight tickets.

7) Bob is a 24 year old young mechanical engineer, who has started a small garage for making small custom �smart-cars�. He has an interview with venture capitalists in San Jose. He needs a tool to plan his flights and his site seeing for the 2 days that he wants to go for.

8) Craig is a graduate student who works in the Semi Conductor Lab at university. He wants to plan a trip with 10 of his friends to somewhere in BC. He needs to make flight and hotel bookings along with getting a list of places for site seeing.

We have limited the scope of this project to tourists to make this project less complex. So the design and requirements have significantly changed. So our present set of users is a more focused subset of our initial of cast of characters.

22

C2-Evaluation Results: • Numerical Data and Results

Name Expedia Time (min) Expedia # of clicks iTravel Time(min)

iTravel # of clicks

Started with

Tana 16 48 6 23 iTravel Betty 19 68 6 44 Expedia John 10 35 4 18 Expedia Bob 14 40 5 25 iTravel Craig 24 56 7 20 iTravel Suzy 32 41 8 37 Expedia Kitty 15 31 4 22 Expedia Jenny 35 67 10 30 iTravel

Table 1: Data ANALYSIS of time account for individ. differences by removing subject mean H1: iTravel takes less time than Expedia H0: There is no difference in time between iTravel and Expedia Using 95% confidence interval N: 8 df: 7 (1) (2) (3) (4) (5) Subj. Pres. iTravel Expedia Subject iTravel Expedia # Order (min) (min) Mean (1)-(3) (2)-(3) Tana IE 6 16 11 -5 5 Betty EI 6 19 13 -7 7 John EI 4 10 7 -3 3 Bob IE 5 14 10 -5 5 Craig IE 7 24 16 -9 9 Suzy EI 8 32 20 -12 12 Kitty EI 4 15 10 -6 6 Jenny IE 10 35 23 -13 13 mean: -7.2 7.2 SS: 169 169 s2: 19 19 sd: 4.3 4.3 se=sd/sqrt(N): 1.53 1.53 t: -4.6966 4.6966 tdist(x, df, tails) : tdist(t,df, 2) : 0.11255560% Reject H0

Table 2: t-test for time to use both the interfaces

23

ANALYSIS of # of Clicks

account for individ. differences by removing subject mean

H1: iTravel takes less # of clicks than Expedia H0: There is no difference in the #of clicks between iTravel and Expedia Using 95% confidence interval

N: 8 df: 7 (1) (2) (3) (4) (5)

Subj. Pres. iTravel Expedia Subject iTravel Expedia # Order (# of clicks) (# of clicks) Mean (1)-(3) (2)-(3)

Tana IE 23 48 36 -13 13 Betty EI 44 68 56 -12 12 John EI 18 37 28 -10 10 Bob IE 25 40 33 -8 8 Craig IE 20 56 38 -18 18 Suzy EI 37 41 39 -2 2 Kitty EI 22 31 27 -5 5 Jenny IE 30 67 49 -19 19

mean: -10.6 10.6 SS: 423 423 s2: 47 47 sd: 6.9 6.9 se=sd/sqrt(N): 2.42 2.42 t: -4.3566 4.3566 tdist(x, df, tails) : tdist(t,df, 2) : 0.18325369% Reject H0

Table 3: t-test for # of clicks

iTravel

Name easy to locate main link?

Wizard contains all options ?

easy to choose hotel/ flight?

easy to book

hotel/flight?

use search option?

time to complete=expected

Tana 5 5 5 4 Yes 5Betty 5 4 5 3 Yes 4John 5 5 5 4 No 4Bob 5 5 4 5 No 5Craig 5 4 5 4 Yes 5Suzy 5 4 5 5 No 5Kitty 5 3 5 4 Yes 4Jenny 4 5 5 3 No 3

Table 4: Post Questionnaire data for iTravel

24

Expedia

Name easy to locate main link?

Wizard contains all options ?

easy to choose hotel/

flight?

easy to book

hotel/flight?

use search option?

time to complete=expected

Tana 5 5 3 4 Yes 4Betty 4 5 2 3 Yes 3John 4 5 4 4 No 4Bob 4 4 3 3 No 3Craig 4 5 2 3 Yes 3Suzy 5 5 2 3 Yes 3Kitty 3 5 3 2 No 2Jenny 3 4 1 1 No 1

Table 5: Post Questionnaire data for Expedia

• Summary of Evaluations and Findings

We have pinned down on the two-tailed test because either Expedia can be faster or iTravel can be faster. Also as can be seen from Table 2 &3, all of the values were normalized around each of the user�s mean. This was because the 8 users were repeated for both the experiments. So there weren�t really 2 different sets of users. As can be seen in Table 2 &3, iTravel is 7.2 minutes faster than Expedia and uses approximately 10 clicks less than Expedia. We have chosen the 95% confidence interval because that value is usually applied; it is most common and seems reasonable for this analysis. As can be seen in Table 2, tdist(t,df, 2) : 0.1125556-% <<< (100-95)% = 5%. So the null hypothesis that implies there is no difference in the time to use both the interfaces, Expedia and iTravel, can be rejected. As can be seen in Table 3, tdist(t,df, 2) : 0.18325369% <<< (100-95) % = 5%. So the null hypothesis that there is no difference in the number of clicks used in Expedia and iTravel can be rejected. So in summary,

We are 95% certain that iTravel is 7.2 minutes faster than Expedia and uses, on an average, 10 clicks less than Expedia.

As mentioned before, speed (time) and clarity/ preciseness (# of clicks) are the two major components of efficiency. Since iTravel is better in both timing and # of clicks, it can be concluded that,

We are 95% certain that iTravel is more efficient than Expedia. The questionnaires results in Table 4 &5 above also support the results (In the questionnaires a higher number means better/ more positive result). In summary the real difference between the two interfaces comes in the Columns 4 &5 in each of these tables. As can be see in Table 4 &5, people can choose and book flights/ hotels much more easily in iTravel than in Expedia. (iTravel had used the �cart� concept at this very stage.)

25

• Problems & Bugs with the evaluation itself Some major problems with the evaluation are: 1) Most of the students evaluated belong to the same faculty of engineering (had

same computer skills etc.) So this might be an insufficiently representative group. This can be seen from the pre-questionnaire data that was collected.

2) Due to varied schedules, most of the tests were ultimately performed in the study rooms. There was noise and some distractions for the people being evaluated.

3) Some people used �Search� in between planning their trip. This could have affected the overall time but exactly half the people used Search so it got kind of counterbalanced.

4) Noting down the number of clicks was very problematic considering that some people have the habit of randomly clicking mouse buttons while scrolling a page.

But overall, most of these errors affected both the interfaces. So the result is probably accurate but the numbers might have some experimental error.

C3-Final design rationale and discussion of the state of your design • Discuss the quality of your interface design?

At this last stage, after both formal and informal evaluations and controlled experiments, it seems that we have ended up with a pretty simple interface that is easy to navigate. The difference between Expedia and iTravel is very surprising. The 8 users seemed to have a much better idea of using a completely new interface than an old established Expedia. Besides being simple and easy to navigate, the interface does provide a complete set of core functionality. Looks wise, based on comments from users, the interface is reasonably professional to look at. So in summary, the quality can be considered averagely high with an easy to use interface, reasonably professional to look at, easy to navigate with reasonable functionality.

• What parts of the design work well and what still needs improvement?

The main page summarizing the two main options of trip planning and individual search worked extremely well with the users. Within 10 seconds or less, most of the users knew exactly which option to select and where to go. The idea of a �3-step� wizard was very popular as the user could specify all her constraints in the first step and then quickly glance over the optional Step 2-3. The whole idea of breaking the wizard into required and optional parts, and not clustering the form with too many fields kept the user active and happy throughout the form filling process. The last idea of having a cart and adding the choices seemed extremely intuitive. Most users knew exactly what to do and how to proceed from this step. This was perhaps the best part of the interface and the place where Expedia lacked the most. As can be seen from the questionnaire results, it is this stage that people start getting confused in Expedia and don�t know how to add their options. The final results need improvement. The schedule and tour guide seem incomplete to users and they need more information. This is one part that could have improved but could not be due to shortage of time. Also the search page (individual search) needs to exist on its own and somehow also needs to be merged with trip planning. For

26

example if a user does not know his destination, he should be allowed to search for it from within the wizard and somehow connect it back to the form fields. Presently the user is taken to a completely new page for search and then after the search, he needs to restart trip planning and enter the form fields. Some integration between trip planning and search needs to be done behind the scenes as well as on the interface to allow the user to navigate back and forth between the two. Lastly, the system needs to be enhanced in functionality before it gets shipped as a full product. Some example of this functionality include an option for rental cars, multi city trip planning, option to choose different transportations for different parts of the trip, different hotels for different parts of the trip etc.

• Do you really believe that the system would work well for your identified users and tasks? Our target users included students/ self employed people who were comfortable in using the internet and needed a tool for trip planning. We found a group of 8 diverse users who fell in this category. After the rigorous controlled experiment and questionnaires, the results clearly showed that the iTravel interface seemed simple and easy to navigate and much easier to use than the established Expedia.com. However, Expedia offers a lot for functionality which might be missing in iTravel, but the core trip planning which is the main thing was much more popular in iTravel than in Expedia. This has given us a lot of confidence that the basic underlying design structure and main functionality of iTravel is popular amongst the target group and this interface works well for the identified users and their tasks.

C4- Reflection on your design process • Whether anything has changed in your perspective?

This was a tough and long project. Initially we started coming up with designs and requirements on our own. But as time progressed things started getting complicated. Each of us had various different ideas and views. Finally we put all those ideas together and came up with our first prototype. We then showed it to some of the known people casually and with a lot of pride considering it to be the best interface design after many hours of effort. We faced a huge setback at that time. The people who saw it gave honest opinions and considered it confusing, unprofessional and highly cluttered. I think at that stage it was pretty clear, we needed people to be a part of this. Informally we started talking to known people and getting requirements etc. At the same time we were being taught about User Centered Design etc. in class. So practical setback experience and class knowledge guided us towards a completely user centered approach. I think we were one team which was really hard hit with a very poor interface to start with but we soon made a lot of progress with this new approach of involving users. So the approach has completely changed for most of us in the team and the importance of users is surmount and all the progress in this complex project has been achieved by involving users.

27

• What have you learned about user-centered design? Did it work for you?

As mentioned above, having a user centered design was perhaps the most drastic change for this project and the core part of the design process. We have learnt that users can be involved formally and informally. For example: when most of us are designing the interfaces, during break times, we usually invite our friends to just have a glimpse of the interface and make suggestions. At times this has helped us resolve complicated problems and also given great creative ideas because different users use different interfaces and software products. If all this knowledge that people have, can be combined we can come up with a great interface. Formal user centered approaches such as heuristic evaluation and interview has had immense benefits which will be highlighted in the next few sections. User centered approach completely worked for us.

• Did you actually see your design change under the influence of user

involvement? What were the biggest surprises for you � the things you learned from or about users would not have predicted based on your own experience and intuition?

Absolutely. This can be seen clearly in our progress from our prototype #1, Low-Fi, Medium Fi and the final one included in this document. The first one started with a cluttered interface to a final finished product that was clear, concise and professional looking. It still needs polishing and amendments but considering the time, the progress is clearly visible. The biggest surprises for us were that users are flexible people. Even if a user has been a long time Expedia user, he will not mind trying out a new thing like iTravel. Also users have the ability to mould themselves into the new interface and grasp the functionality. They are not as aggressive about using the old tried out interface with the very same placement, borders, colors etc. as we thought. Which means that people out there, are very willing to try out new interfaces and there is a lot of market for new products doing the same thing as the existing product but with a different interface. Another big surprise that we learnt from users is following the thought process as if you were talking to yourself and then designing based on that thought process. You could design as a sequence of steps for example the user will want to plan a trip and then maybe do a search Or you could design following the thought process for example the user will want to first do trip planning, in between lets assume she gets confused, she would want to open help, simultaneously search for a destination and then go back to where she was trip planning. This was something that we learnt after lots of brain-storming and later when we studied task-centered design in class, it just confirmed our design process. It was impossible to collect all the possible tasks, but with a streamlined process, we managed to collect some of the core tasks and the thought process of the user.

28

• Did the methods you chose for your evaluation and prototyping get at what you were looking for? In hindsight, would a different approach have been better? (We�re talking about process here, not specifics of your interface). With respect to evaluations, we are very satisfied with the heuristic evaluations, questionnaires, controlled experiments and observations. It really highlighted a lot of problems in the interface and suddenly an interface that looked perfect had all these defects. So evaluations were perhaps one of the best parts of the project. For prototyping we started with C#. C# is good for screenshots, but for complete website design, it�s pretty difficult. So after our Low-Fi and other sample prototypes in C#, we had to completely start from the beginning in Microsoft Front page to make a workable final model. Perhaps if we had started with Front Page from the beginning it would have speeded up things and we would not have to start from scratch for Pass 2. Also our approach was huge, we tried to aim at a lot of requirements and in the end when the interface was getting complex, we suddenly had to cut all our requirements to bare minimum. If we had decided on a small set of requirements in the beginning it would have made our design process simpler and less complicated.

• What were the most, and least, valuable among the activities we�ve asked you

either generally or specifically for your project? The most valuable activities included identifying users and tasks, heuristic evaluation, observation, questionnaires, controlled experiment and Low-Fi & Medium Fi Prototype. These activities really helped us in shape the goals of this project and achieve the goals in the best and most popular ways and also helped us in finding lots of bugs with the interface. The least valuable activities included cognitive walkthrough because with such a huge interface, a walkthrough was complicated and took too much time. The other methods of evaluation seemed to do the same as the walkthrough.

• Do you think that having gone through 418, you might approach your next

interface design project (whether for fun or work, large or small) differently? How?

As mentioned above, we learnt a lot about designing interfaces. And we have learnt many new and useful things which will certainly help us approach the project differently. Some of the key ones include user centered design, design based on tasks (sequence of actions), heuristic evaluation and controlled experiments. These gave us amazing insights into where our interface stands with the world and how it can be improved.

29

Appendix 1: Design alternatives Set 1: This has several wizard items optional

Figure 14: Step 1 of the wizard with optional second and third step and a submit button

Figure 15: Step 2 of the wizard with optional second and third step and a submit button

30

Set 2: This is the earlier draft for the main home page

Figure 16: An early draft of the main page Set 3: These are several alternatives for the options page with different cart, color, placement styles

Figure 17: Alternative one for the options page

31

Figure 18: Alternative three for the options page Set 4: An early alternative for the complete interface

Figure 19: Main Page

32

Figure 20: Inter-city Trip Planning � Wizard Step 1

Figure 21: Inter-city Trip Planning � Wizard Step 2

33

Figure 22: Inter-city Trip Planning � Wizard Step 3

Figure 23: Attractions options

34

Figure 24: Accommodation options

Figure 25: Rigid Schedule

35

Figure 26: Tour Guide � Page 1 (Upper Half)

Figure 27: Tour Guide � Page 2 (Lower Half)

36

Figure 28: Search

Figure 30: About iTravel

37

Appendix 2: Study Instruments Pre Questionnaire: Name: Age: 1) In your typical work day, do you use computers: O over 4 hrs a day O between 2 and 4 hrs daily O between 1 and 2 hrs daily O less than 1 hr a day 3) How do you usually plan your trip? Rank in order of your preference with the highest

number corresponding to the least preferable:

__ Tour Guide __ Web Search __ Travel Agents __ Ask friends/ family

4) When you plan a trip, how strictly do you follow your planned schedule

Don�t follow at all 1 2 3 4 5 strictly follow 5) How much time are you willing to spend to plan you trip

____

38

Post Questionnaire Name: Age: Interface used O iTravel O Expedia 1) Could you easily locate the option for trip planning

hard to locate 1 2 3 4 5 easy to locate 2) Did the wizard/ system allow you to specify all your constraints (time, budget etc.)?

not at all 1 2 3 4 5 absolutely

3) Were the hotel/ flight options easy to locate (i.e. visible)

not at all 1 2 3 4 5 absolutely

4) Did you know how to book your flights and hotels after results were provided?

not at all 1 2 3 4 5 absolutely 5) Did use the search option?

O Yes O No

6) Was the time it took you to do the trip planning within your expected time

unreasonable 1 2 3 4 5 reasonable

39

Instruction and Task Description handed out at the start of the controlled experiment

1) Please take a seat and face the computer. 2) Please read the consent form for the questionnaire lying on your right side. Your

consent will be assumed once you have filled in the questionnaire. 3) Please read and sign the �Consent Form for an experiment that does involve

video� which is the second from in the pile. This experiment does not involve video recordings.

4) Please fill in the pre-questionnaire, lying below the consent form, with a pen. 5) The monitor has now been turned on. 6) The screen shows you the main page of the interface of iTravel/ Expedia (only

one will be circled in the print out). 7) Your task is to plan a trip from once city to another. You will have to navigate the

interface and select options that allow you to do the same. I will not help you in anyway in navigating the interface.

8) At all times I will be standing on your left side and noting down your time to navigate the interface and the number of clicks you use to reach your result. You are free to think aloud or remain silent. Your loud comments will be recorded be me.

9) If at any point you feel distracted with my presence or are having technical problems or any problems, please notify me and the experiment will be stopped.

10) Once you have reached your final results and do not wish to use the interface anymore, please raise your right arm. The stop watch will be stopped and the experiment will end.

11) At this point you will have to fill in the post questionnaire, which is the last form in the pile, with a pen.

12) Please return all the forms to me. 13) You have a 7 minute break to grab pop and pizza at the end of the room. At the

end of the 7 minute break, please return to this computer to start the 2nd half of the experiment with the �other� interface and repeat Step 1-12.

40

Appendix 3 & 4: Raw Data & Consent Forms The completed questionnaires and consent forms have been attached.