Haven
Website
MY ROLE
UX, UI, A/B Testing, Personalisation, User Testing & Research, Journey Mapping, Data-Driven Design, User-Centred Design, Wireframing, Visual Design, Graphics, Prototyping, Conceptualising, Mentoring Junior & Midweight designers
Overview
​
As a Senior UX/UI Designer, I was responsible for improving the top-of-the-funnel website experience for Haven, focusing on holiday and caravan sales pages. My objective was to create user-friendly journeys that aligned with business goals while delivering a seamless and personalised discovery experience for users.
​​​​
​
My approach
​
User-Centered Design​
-
Designed tailored user journeys for different user types, focusing on key stages in the holiday booking and caravan purchase processes. Developed "happy paths" to guide users smoothly from exploration to decision-making, ensuring a seamless and enjoyable experience at every touchpoint.
​
Data-Driven and Research-Centered Approach​
-
Adopted a data-driven design methodology, leveraging insights from various sources including Amplitude and Crazy Egg to analyse user behavior and optimise key touchpoints. Conducted A/B testing to validate hypotheses and refine the user experience, improving both holiday booking rates and caravan ownership leads.
-
Carried out additional user research, including online surveys and usability tests, to gather valuable feedback and understand guest and potential owner needs and pain points. This research enabled us to design with empathy and precision, ensuring our solutions addressed real user challenges.
-
Utilised guest insight data to inform design decisions and iteratively tested prototypes with users, incorporating their feedback to continuously enhance and optimise the user experience.
Feature design and personalisation
-
Designed new website features and implemented personalisation rules based on user behaviour, browsing history, web analytics, and research insights.
-
Collaborated with the SmartSeer team to integrate personalisation strategies for holidaymakers and potential caravan buyers.
​
Cross-functional collaboration
-
Worked closely with product managers, engineers, SEO specialists, and stakeholders to align strategies with user needs and business objectives.
-
Partnered with the branding and marketing team to ensure consistency in the newly developed Haven brand across top-of-the-funnel pages.
​
Key Outcomes
​
-
Improved User Engagement: Enhanced user journeys contributed to higher engagement rates and improved navigation of holiday and caravan sales pages and helped reach our business targets.
-
Increased Conversions: Personalisation strategies and A/B testing resulted in over 12% increase in holiday booking rates and a 18% rise in potential owner leads. The yearly holiday sales targets where exceeded each year.
-
Brand Consistency: Successfully implemented cohesive branding across key website pages, reinforcing the Haven identity.
Case study: Helping Users Find the Right Park
Case Study: Helping Users Find the Perfect Park
​
User research revealed a common challenge: users often struggled to identify the park that best suited their needs and found it difficult to compare options effectively.
​
Key user feedback included:
"I’d like to know which of your parks have a swimming pool. A nice swimming pool is very important to us. Our teens always want to see pictures of it."
"Which park is best for toddlers?"
"Are your parks mainly for families? Would there be much for us to do as a couple?"
"I would like to see more information about the park on your map. Oh, what just happened here? I thought I would see info of park by clicking on the park pin, but I'm taken to see the region instead."
Addressing these concerns became a top business priority for 2023 and 2024, focusing on empowering users to find the right park effortlessly.
​
Solutions suggestions included:
-
Enhancing park pages with richer content, more information, visual improvements, and improving their image galleries
-
Introducing advanced filtering options and improving map experience to help users narrow down their choices on 'Our parks' page
-
Personalising park cards by highlighting key features such as amenities, activities, and suitability for specific audiences (e.g., families, couples).
-
Exploring innovative AI-driven tools and interactive quizzes to guide users in discovering parks that best match their preferences.
​
Project: Designing AI Chatbot to find 'the right park'
​
Project Overview
​
To enhance the guest experience and simplify the process of selecting the perfect holiday park, we designed an AI chatbot tool for Haven. The chatbot aimed to assist users in finding the right park based on their preferences, motivations, and needs. This project was a collaboration with OneBeyond agency, which handled the development, while I focused on the UX, UI, and content design.
​
Objectives​
​
-
Help guests find the most suitable Haven park based on their holiday preferences.
-
Provide tailored recommendations by addressing diverse user motivations and needs.
-
Ensure the chatbot educates users of Haven proposition offering and elevates the Haven brand
-
Increase user engagement and booking rate
-
Gather user preference data to enable personalised experiences on the website and through email communications.
​
Key Design Decisions​
​
Flow structure and questions
​
Designed a conversation flow tailored to our key user segments, including those with dogs, accessibility needs, and specific holiday preferences. The flow accounted for user motivations, such as lifestyle choices, activities, entertainment, accommodation preferences, and park facility requirements. To enhance personalisation, questions about party composition were placed at the beginning, allowing subsequent questions, featuring hints and multiple-choice options,to be customised for each user group.​
Key user segments:
​
A) Families with young kids (up to 5 years)
B) Families with older kids (5+)
C) Guests with dogs
D) Adults only
​
User motivation groups:
​
A) Relaxation Seekers: Guests looking for a tranquil, laid-back experience, park entertainment, premium accommodation, self-catering - HIGH ON SITE FOCUS WITH GOOD ACCOMMODATION
B) Activity Enthusiasts: Guests looking for pre-booking lots of action packed and other type of activities, entertainment - HIGH ON SITE FOCUS
C) Explorers: Guests interested in perfect location, great ourdoors, expecially beaches, attractions - LOW ON SITE FOCUS
D) Entertainment Lovers: Guests interested in doing family activites on park tohether tather than off site, evening entertainment, eating out on park - HIGH ON SITE FOCUS
​​​​​
​
​Question Types
​​​​
Majority of the questions had multiple choice options with a text input option to enter their own response or open ended questions with hints. I made sure that the questions where appropriate would have a descriptive sentence or hints educating users on the Haven strong offering and make us stand out from the competitors.
​​​​​
​1. Open-ended questions with hints: Applied to questions with more than nine answer options or where users were likely to provide their own response. For example, in questions about activities and things to do on the park or nearby, we included a wide range of options while also capturing additional user interests, even those we don't currently offer, to report back to the business for future consideration. Questions about activities on the park and nearby were where we needed the most help from AI, as each segment group had very different preferences. Specifically, for nearby activities, we aimed to leverage AI's web-scanning capabilities to provide tailored recommendations near our parks.
​
On-site activities question template:
Brilliant! When it comes to on- site activities, Haven has a lot in store for you and your guests. Are you looking for a park that offers [AI generated short list of activities suitable for the user segment] Or something else?
​
On-site activities question tailored for family with younger and older kids example:
Brilliant! When it comes to on- site activities, Haven has a lot in store for you and your guests. Are you looking for a park that offers thrilling activities like a climbing wall, aerial adventure, or Nerf battles? Or perhaps one with a multi- sports court, soft play area, and a creative studio for crafts? Maybe you're interested in exciting water sports? Or something else?
​
​2. Multiple choice questions: Applied to questions with fewer than nine options, some of which included a type-in response option where we believed it could provide valuable business insights. I proposed adding images to multiple-choice selections to make the chat experience more visually engaging and inspiring, showcasing pictures that capture moments of joy.
Multiple choice swimming pools question example:​​​
Get set for a splashing great time. Every park has its own indoor pool, but some have even more water features. Is any of the below important to you? Options: A) Indoor pool B) Outdoor pool C) Outdoor splashzone D) Slides and flumes E) Lazy river
​
Data Integration
​
-
Fed the chatbot with Haven’s internal data and content about parks, activities, and facilities which also included curated user reviews to ensure accurate and reliable responses and minimise any risks of harming Haven brand.
-
Enabled the AI to pull web-based information for nearby activities and attractions, filling gaps in knowledge and enriching the user experience.​​​​
​Results presentation
​
The results screen needed to make sure that our users convert. Results content structure:
-
Recommended a park accompanied by a promotional video, leveraging data insights that showed a 4.5% increase in booking rates among users who engaged with the video on our park pages.
-
Park recommendation copy was tailored to each segment group, highlighting their selected preferences.
-
Option to visit the recommended park page to learn more.
-
Options to explore other similar parks.
-
Option book a break with leading prices taking users to the search journey.
​​​​
Design process
​​
User Research
Utilised existing user research on motivation groups and user segments to identify key motivations, preferences, and personalized content needed for the results screen.
Mapped the journey for different user types to ensure the flow met their needs effectively and that the results screen accurately reflected these insights.
​
Question Content and Answers
Created an initial draft flow on Miro to collaborate with the content team, refining questions through two workshops and obtaining approval from senior stakeholders.
​
UI and Visual Design
Designed a clean, intuitive interface, developing unique chatbot UI components that adhered to the Haven style guide.
​
AI Chatbot Development
Collaborated closely with the OneBeyond agency to fully understand the AI chatbot's capabilities, limitations, and potential risks. This collaboration ensured that the park finder AI chatbot aligned with our strategic business goals while delivering value to users. Provided detailed specifications for the chatbot’s flow, content, and functionality.
​
Project outcome
​
The project is still ongoing.
​​​​
Project: Park picker quiz
​
Background
​
Our AI chatbot project was a significant undertaking, requiring extensive time and effort to ensure it would be a robust tool with no negative impact on the brand. To address this challenge and generate actionable insights more quickly, our squad working on Haven Discovery decided to create a lightweight quiz-based park finder tool called Park Picker. This initiative allowed us to test key hypotheses, better understand our users, and lay the groundwork for the larger AI initiative.
​
Objectives
​
-
Validate Hypotheses: Use a lightweight tool to test assumptions that users are more likely to book when engaging with an interactive tool that provides that more information about Haven parks.
-
Generate Actionable Insights: Quickly gather data on what users value most in a park.
-
Build Foundations for AI: Develop a scalable framework for future AI-driven tool
Design Process
​
Leveraging existing work from the AI chatbot project and the data we had on park facilities, I developed a simple 6-question quiz for the Park Picker tool. Foundational work, such as user flows and data structures, had already been completed during the chatbot project, which allowed me to adapt and streamline the AI chatbot's flow to fit the Park Picker’s requirements. The focus was on delivering a quick Minimum Viable Product (MVP) using limited data available on our parks.
​
Limited Data Challenge : At the time, we did not have complete data on all 41 parks in terms of facilities and activities. Integration with Haven team tools holding this data required extensive development work. Our agreed solution to the problem was to group activities into broader categories and create a park score for each attribute, based on their knowledge and available data, on a scale of 0 to 3.
Sample question: What are the most important holiday facilities for you?​ Options: An amazing swimming pool, a variety of activities, plenty of places to eat and drink, playing golf, fishing, water sports, direct access to the beach.
​
Refining the Results Screen: During our team discussions, we realised we needed to iterate on how we presented the quiz results and we had differences in opinions, on the best design solution and the rules that we need to use to show the best results, so we set up workshop for Crazy 8's. Using the Crazy 8's design method, we quickly generated eight different ideas in eight minutes. This exercise allowed us to explore a broad range of solutions and narrow down what would work best for the initial release and find a joint agreement, while also identifying more aspiration features for future versions. ​​​​​
​​First Release Results
The launch of the Park Picker tool was a resounding success. The tool achieved an impressive 82% completion rate and a 5.4% booking conversion rate, indicating that users were highly engaged and the quiz was effective in helping them find the right park for their needs. These early results exceeded our expectations and laid the groundwork for further refinements and improvements.​
​​
User Feedback & Iterations
We added an Open Feedback input box from the start, allowing users to submit suggestions for improvement. Key feedback and desired features included:
-
The ability to select a preferred region instead of defining travel distance (which was initially categorised as under 2 hours, 2-4 hours, and over 4 hours).
-
A preference to see parks only within their specified travel distance, rather than being shown parks further away, which highlighted extra travel time.
-
Requests for more park recommendations and other personalised features.
Next Steps
​
Building on the success of the first release, we’re now focused on integrating user feedback and introducing new features to further improve the Park Picker experience.
-
I designed an additional feature that allows users to not only input their postcode and set a travel distance but also to select their desired regions.
-
Additionally, we updated the recommendation rules. Now, instead of showing the park with the highest score slightly outside the user’s set travel distance (max 30 minutes), we display parks that are within their specified distance.
-
I also revisited the original wireframes, exploring designs that show more park options in a stacked layout, making them more engaging and visually appealing. These updates aim to enhance the user experience by offering more relevant, personalised results.
​
Quiz Results Redesign
​
I designed a new quiz results screen featuring stacked cards to present multiple results in a more engaging and interactive way. This redesign also included an option for users to add their preferred parks to their wishlist. The redesign was implemented as part of an A/B test to evaluate the effectiveness of different content types within the cards.
The test included two variants:
-
A) Ken Burns Animation: A fading in-and-out effect applied to the park images. These images were personalised for each user segment, showcasing their selected preferences.
-
B) Looping 15-Second Video: A video highlighting the park’s surroundings, key facilities, and activities. This could be further personalised for each segment if proven effective in the A/B test.
Control Variant: The original quiz results screen (first release) served as the control for comparison.
Hypothesis:
By redesigning the quiz results screen with more visually engaging content and more results, we expect to see:
-
Increased interaction rates with the quiz results (e.g., clicks or scroll depth).
-
Increased booking rate.
Project: Our parks page redesign
​
As part of the initiative to help users find their ideal park, I redesigned the 'Our Parks' page, introducing an interactive map with intuitive interactions and advanced filtering options. This design was informed by extensive user research and testing, ensuring a seamless, user-focused exploration experience.
Project: Holidays homepage personalisation
Project overview​​​
​
Utilising user behaviour data, web analytics, and research insights, I developed a range of personalisation strategies for key pages tailored to our holidaymakers. This included testing variations of personalised content and adjusting its position based on the user's visit frequency, catering to different user segments. I also defined the personalisation rules to guide these implementations. These enhancements were rolled out in stages, delivering a seamless, research-driven experience that effectively met diverse user needs and boosted overall engagement.
Project: Accommodation Pages Redesign​
​
I redesigned Haven's holiday accommodation pages to create a more engaging and user-friendly experience. The new accommodation overview page provides a quick, easy-to-digest summary of all accommodation options and other useful information to the user during their discovery. The new design of accommodation detail page has clearly highlighted key features of that accommodation type, prominent video and dedicated sections for each room type, accompanied by motivational selling points offering a more visual and immersive walkthrough.
