NWR Trip Planner

Background Research

Before exploring solutions, we wanted to get acquainted with NWR and the hospitality industry in general. To start, we wanted to find answers to these 3 questions:

What’s NWR about?
What are some of the possible lacks of NWR?
What are the latest trends in hospitality and tech?

Background Research

To answer the questions we proposed, I planned out a series of secondary research activities for the team to complete together.

I conducted interviews with 3 past NWR guests to understand their purpose for visiting the resort and their experiences.
My team and I pretended to be prospective guests and booked a stay with NWR agents to experience the booking process from a real guest’s perspective.
My team members conduct market analysis on the services and technologies the hospitality industry is investing in to find design opportunities.

Findings

1. Regular guests tend to have a routine of activities that they stick to and take little recommendations from the resort on exploring outside their comfort zone.

2. It can be difficult for staff to remember all 70+ activities let alone promote them to guests.

3. Nemacolin’s biggest strength against their competitor is the large number of activities they offer.

4. There is a blue ocean in the hospitality industry for fun technology aimed at the pre-on-site stage of guest’s experience.

Quantitative Analysis Reveals Target Guest Group

While we were chucking along with the secondary research, our lead developer Kunal Bhuwalka also started analyzing the data we received from NWR on their past 3 years’ reservations. The data helped us get a basic understanding of their guest demographics and behavior patterns.

Findings

1. Though the older and wealthier demographic is the primary guest of NWR, younger guests (35-45) are spending more at the resort.

2. Guest spending on dining, SPA, and recreation generates the majority of Nemacolin revenue.

This analysis helped us identify the targeted guest group for our project:

35-45-year-old upper-middle-class individuals from the greater Pittsburgh area who generally has kids aged 6-15

Explorative Research

To identify the underlying cause of the low conversion rate and the key to attracting more guests, I led the team into doing 3-month explorative research where we merged ourselves into the hospitality industry and analyzed the guest booking process. In this explorative research phase, I planned several research activities:

Contextual inquiry at NWR with the resort staff.
Contextual inquiry with individuals from the targeted guest group.

Contextual Inquiry with NWR Staff

We were able to have a 2 day 1 night stay at NWR and I took the chance to do contextual inquiries with the staff member for us to better understand how the backstage of the hospitality industry works and observe how guests are served. The entities we observed included: Resort Concierge, Check-In Front Desk, Bell Hop, and Restaurant Manager.

Observed Entities: Resort Concierge, Check-In Front Desk, Bell Hop, and Restaurant Manager.

Findings

1. Reservations are key to accessing the best possible Nemacolin Experience. Guests’ lack of action before arrival has the potential to greatly impact their experience and impression of Nemacolin.

2. Nemacolin team doesn’t have a deep understanding of guest behavior during the pre-arrival phase.

Contextual Inquiry with Targeted Guest Group

From our onsite visit, we knew guests’ pre-arrival planning is key to a great experience at NWR but our client has little knowledge of guests’ behavior during this stage. To better understand how guests plan their trips and make decisions on reservations, I drafted a new round of research on the targeted guest segment.

Observation: ask participants to make a decision on resort to book and make reservations.
Interview: follow up with questions on their decision process.

Findings

From our research, I realized what differentiates guests is not where they are from, their occupation, or their age. Instead, I found out:

The primary distinguisher among guests are their motivations for going on to trip.
Guest motivations influences what they value during trip planning and how they make plans.

4 Motivations that Define Perceived Value

Value Dictates Planning Process

Additionally, through participants' interaction with NWR website, we found the following common influencing factors attributing to guests overall experience during trip planning:
1. Lack of specificity in hotel and room unique values prevents prospective guests from exploring their options.

2. High accessibility to resort information builds trust and makes guests feel valued.

3. Word of mouth from close social and personal influencers plays a big role in determining the value of a resort.

Identify Critical Variables Indicating Guest Behavior Through Machine Learning

We explored the guest behavior and spending pattern by analyzing existing data: reservation data, call center & CRM data, and guest feedback data. We wanted to identify the relevant variables in guests making a booking decision among the 4 hotel options NWR provides. Our lead developer utilized black-box machine learning to generate a CRM decision tree based on the actual guest data it’s trained with 43% accuracy in hotel booking prediction. From the algorithm, we were able to pinpoint the following variables as critical in hotel booking for NWR guests:

Reason for stay
Season
Stay month
Adults
Number of rooms
Week number
Day of the week
If there are children

Work of Kunal Bhuwalka.

Research Conclusion

With the insights we gathered through explorative research, we defined a 2-direction design approach that aims to better support prospective guests’ pre-arrival experience and enable NWR with providing more personalized services.

For Nemacolin Woodlands Resort

The Problem:

How might we leverage guest data to provide more personalized services and create conversions?

The Needs:

Increase booking
Gather more guest data
Provide more personalized services
Foster positive brand image

For Prospective Guests

The Problem:

How might we encourage prospective guests to make reservations prior to arrival?

The Needs:

An easier planning process
A sense of control in trip planning
A sense of luxury and quality

Iteration & Evaluation

Our 2 designers led a series of iterations on the design with me directing evaluation between each iteration, providing actionable insights. I was responsible for creating all the testing scripts and organizing the team to conduct them as well as assisting with designs.

Initial Explorative Concepts

We started the design process with 3 initial big ideas (shown below) and then brainstormed as a team through a crazy 8 activities to explore more possibilities. We developed the generated ideas into storyboards to help us 1) validate the needs and 2) narrow down on path forward through an evaluation method called speed dating. These concepts were tested with both prospective guests and NWR staff.

Work of Xiangzhu Chen.

Speed-dating recruitment plan Speed-dating script - Guest Speed-dating script - staff

Concept Validation Results

It became obvious after speed-dating our initial concept storyboards with users that the solution we provide need to support the entire guests reservation journey, from the moment they start learning about the resort options, to looking into the specific hotel and dinning options, to finally complete the payment.

Identify Information Critical for Decision Making

To understand the different pieces of information that are critical to the decision making process that users go through while planning for their trip, I led my team to conduct card sorting with 5 potential guests, asking them to categorize the information they’d like to see based on the stage they are in. Here are the results.

Iteration 1: Exploring Two Approaches to Encourage Reservation

We started the design with exploring how to increase reservations during the hotel booking process. We wanted to make the reservation process as smooth as possible but also deliver the tailored experience that prospective guests look for. To accomplish the goal, we wanted to build in an survey process during booking for us to

Make better recommendations to guests
Collect data to train recommendation algorithm

We decided to go with a mobile-first approach when designing because during our contextual inquiry, most participants reported that they do their research and planning using their phones.

Approach 1: Prompt Guests to Build Trip Itinerary During Booking

Group effort between Gillis Bernard, Xiangzhu Chen, and Melissa Shi.

Approach 2: Recommend Guest Pre-Built Itinerary During Booking

Group effort between Gillis Bernard, Xiangzhu Chen, and Melissa Shi.

Test Findings

With initial round of testing, my goal was to understand which of the two or parts of the workflows better matches with users expectation and habit. Additionally, I also wanted to confirm our decision with a mobile first approach with design by better understand the devices used for trip planning. To accomplish these goals and work around our limitations around participant recruiting, my test plan consisted of 3 parts: guerilla research at Wholefoods, 1 in-person test, and 3 virtual tests.
Through testing, we were able to receive some feedback on the two work flows.

Users desire more flexibility during the trip planning process and better affordance
Users want to see more detailed information regarding the options NWR offers up front
Users wish to know which days are the recommended options good for

Based on the feedback received, I was able to recommend the team to move forward with Flow 1.

Iteration 1 User testing plan

Iteration 2: Providing More Flexibility on Itinerary Building

With the second iteration, we improved on the design per feedback we received. In this design, we 1) integrated more details regarding the stay and experiences specialties, 2) allow users to manage itinerary easier, and 3) display itinerary based on day.

Test Findings

With this round of testing, I also decided to both guerilla and formal testing with guerilla research at the local farmers market and formal in-person testing. Additionally, we went through critique with our advising faculty members. During testing, we found out that we were lacking in the following area:

Not setting up proper expectations on what the tool is for before prospective guests start using it.
Not providing enough sense of urgency to promote reservation.
Not enough details were provided on NWR offers to demonstrate their unique values.
Need a way to come back to recommendations that guests are interested in.

Iteration 2 User testing plan

Guerilla paper prototype test at local farmers’ market.

Iteration 3: Build Step-by-Step Trip Planning Guide

From Iteration2 testing, it became clear that there is a lack of guidance for users to navigate from one step to the next, which prevents them from converting their trip interests to actual booking. In the new iteration, I helped built a better navigation for the planner.

Test Findings

Due to the limitation around recruiting testing participant from our targeted demographic, I decided to conduct round 3 of testing through two channels: guerilla research at local farmers market and virtual testing through usertesting.com. With the testing, I aimed to answer these questions:

If user understood what the tool was for
Did the overall process matches users’ expectations
Users’ confidence in the recommendations provided
Were there enough feedforward o guide users through the process

Iteration 3 virtual testing plan Iteration 3 Guerilla Testing Plan

Through testing, we were able to confirm that users are able to navigate the entire workflow presented smoothly with a clear understanding on what to expect from the tool. However, during this round of testing, we realized that there were other use cases that we haven’t consider yet. These use cases includes:

What happens if users want to come back to their itinerary later?
What happens if there are availability conflicts?
What happens if user only book room and not activities?

Aside from the above use cases, we were also debating whether we should implement an auto generate itinerary feature to boost the likelihood of guests making activity and dinning reservations. It was critical to better understand the sentiment towards auto-populate user trip itinerary.

Iteration 4: Refine Details

With this iteration, we were focused on refining the entire workflow and all the secondary use cases that we discovered during Iteration 3 testing. We built in the option to auto-populate itinerary based on users stared recommendations to assess users’ acceptance towards the idea. Additionally, we conducted a UI Exploration exercise to unleash our creativity on branding and uplift the over web app visual styles where we generated 63 different visual design layouts and asked peers to choose from based our intended branding image.

Test Findings

With this last round of testing, aside from evaluating whether we fulfilled the missing use cases discovered during iteration 3 we also wanted to assess users interest in having auto populated itinerary items and lastly understand users’ impression of the booking process and NWR.

When asked about their impression of the NWR trip-planning process, they reported that it was:
Professional
Modern
Efficient
Organized
Helpful
Clean
Inviting
Easy

We also looked into users sentiment towards the 2 ways of generating itinerary: auto-populate vs manual, and we found out that users do not have a strong preference toward auto-populate or manual itinerary creation.

Iteration 4 User testing plan

Categorize Activities through Card Sorting

Since the goal is to design a smart trip planning tool that encourages higher conversion through custom recommendations based on user motivations and perceived values, we wanted to categorize all 70+ activities that NWR offers into various trip style. To help us categorize, I asked my team to conduct card sorting with the rest of the cohort, asking the participants to group the activities based whether they are cultural, relaxing, romantic, family friendly or active/wellness.

Relaxation and Romance have a lot of cross-over. Depending on who you are travelling with, it could go either way.
Active and Adventure seem to have overlaps as well.
In general, more slow pace activity has the potential to be romantic.
Animal Care Center, Cigar & Bar, and Metabolism Testing seem out of place.
Adventure seems to be activities that requires to go offsite.

Final Design & Results

The final solution, the Nemacolin Woodlands Resort Trip Planner, is a machine learning based step-by-step trip planning guide that results higher booking and reservation conversion rate by providing tailored recommendations that with a streamline fashion that delivers the feeling of professionalism, modern, and hospitality, while helping the resort collect guest decision making data to further enhance the recommendation algorithm for better white glove services..

Try the final Prototype!

The 4-Step Process

Quiz

Recommendations

Itinerary

Booking

view Site Map

10 Design Principles

Digitizing white glove service

Curating for conversions

Accommodating for demand

Streamline the planning process

Match mental model for trip planning

Transparency and setting expectations

Offering flexibility within structure

Persuading with scarcity

Scannable and digestible design

Embodying Nemacolin brand

The Style Guide

We decided to follow follow the existing Nemacolin brand image and inherited the colors and fonts used by the current official website. I conducted a NWR website structure and style study at the start of the design phase, which we carried referenced during high-fidelity design.

The Back-End Decision Tree for Room and Activity Recommendations

The final recommendation logic is a method for filtering information to present users with only the most relevant items generated from 2 years' guest reservation, feedback, and call center data. Since our target segment are first-time guests, we do not have access to their past behavior or their preferences. As a result, we employ a Decision Tree to predict a guest preference fora hotel and then make a “recommendation”. Thus, the model that we use is not a truerecommender system but a classifier. A Machine Learning Classifier attempts to map agiven input to a category within a predefined set. For instance, we try to map the guestdata to the different hotels at the resort.

Work of Kunal Bhuwalka.

See Technical Memo

Our Achievements and Design Outcome

With every testing we did, I built in 5 questions at the end of the test session to gauge the success of the tool design.

How likely are you going to book a stay at NWR if price is not an issue?
How likely would you book activities and dining ahead of time if you are using this tool?
Do you think this tool will encourage you to book things early compared with what you typically do?
Three words to describe your impression of NWR after interacting with the tool.
Three words to describe your impression of the trip planner design.

For Nemacolin: Are guests more likely to book?

Increased potential for conversion

Before: 53% organic traffic comes from mobile with only 19% digital transactions completed on mobile.
After: All 25 testing participants from the 5 iterations that we did reported that they are likely to book a stay at NWR.

Encouraged earlier reservation by 42%

Before: 50% guests book ahead of time.
After: Out of the 25 testing participants, 92% said they are likely to book ahead and 42% said they will book earlier than typical.

For Prospective Guests: Is it an easier planning process?

Guest impression of trip-planning process turned positive

Enhanced NWR image