Information Architecture: How usability improvements increased conversion
HOW IT BEGaN
Before delving into the execution of the AI project, it's important to understand why we embarked on this journey. It all began with Paddy Power Sports' new business strategy for 2021, which aimed to build a market-leading Acca Experience. An Acca, short for accumulator, consists of multiple selections from various events, where all selections must win for the bet to be successful (e.g., Football team A winning one match and Rugby team B winning another).
To achieve this goal, we assembled a diverse working group comprising individuals from various backgrounds and disciplines. Our mission was to assess the desired outcomes, objectives, and strategies necessary for success.
Team lineup
Head of Product
Director of Sports
Head of Football
Head of Insights and Research
Head of PP Design
Principal Designer (me)
Head of Promotions
Head of CMS
Senior Sports Trader
UX Research Manager Product Analyst
Identifying the Problem
PROACTIVE RESEARCH
We conducted proactive research to gain a better understanding of customer pain points and opportunities related to football Accas. We used customer insights, UX research, and customer journey data to identify the most significant pain points and areas of friction in the product.
NARROWING IN ON THE PROBLEM
When we triangulated the research we identifed 3 phases to the bet journey:
Researching a bet: this more commonly happens outside the product.
Bet building and placement: this consists of using the product to place a bet.
Bet tracking a settlement: this is a combination of watching or following the live event the bet is based on while interacting with the product.
We decided to focus on the Bet Building Placement phase for a number of reasons:
It’s an absolutely necessary step for placing a bet.
Too many uncontrollable variables with the other two phases.
We identified a lot of friction in the journey for building and placing a bet. The pain points included:
Inconsistent navigation
Difficulty in finding matches and markets
Excessive scrolling and pogo sticking in and out of different pages.
Lack of filtering options
A sense of missing out on valuable markets.
Planning
CUSTOMER AND METRICS WORKSHOP
We conducted a workshop to identify and align the team on our target customer cohorts and success metrics. It was important to define this before jumping into exploring what projects and features would help us be successful.
PRIORITY WORKSHOP
PRODUCT ROADMAP
Using the analysed output from the priority workshop, we created a product roadmap to break down the work into achievable deliverables and measure ongoing impact. The work was then assigned to the design team to pick up.
Information Architecture
Step 1: Competitor Analysis
Once we knew we had an Information Architecture problem we set out to make improvements. I lead the design team in conducting a thorough competitor analysis. To guide our evaluation, I created specific principles related to AI. I used these principles to train the designers on how to analyse each competitor. There were multiple design problems identified:
Lack of clarity and purpose of information
Excessive duplication of content
Long Pages of unused content
Inconsistent organisation and labelling
Step 2: Pupose Workshop
PROCESS
Due to the massive duplication of content across the football experience it was clear that it lacked purpose. I conducted workshop with the Product Manager, CMS team, Football Commercial Marketing Manager to agree on the purpose of each part of the football experience and create clarity
Step 3: Consolidation
PROCESS
I then consolidated the findings from both workshops to identify common themes and create a list of recommendations. The output was a list of 28 recommendations split into 4 themes:
Simplify IA
New functionality
Design Templates
System Capabilities.
I presented the recommendations to the Head of Product to prioritise against other roadmap initiatives.
Step 4: Execution
Working as team we split out the initiatives based on what was required to implement each recommendation and the effort involved. Recommendation like reducing the number of matches in a list, consolidating duplicated content into one and text label changes were delegated to the CMS team. Recommendations like including new functionality were deligated to the design team. Other recommendations such a vision for simplicity, tone of voice and tailored experience were ambitious so were added to a list to review later.
HOW IT PLAYED OUT
6 Quick Wins
There were 6 recommendations that were quick to implement and deliver value. These included changing the text labels of 2 tabs, combining 2 tabs into one, reducing the number of matches by 50% in one of the tabs, and updating 2 UI components to make it easier to scan information.
5 CMS improvements
5 recommendations were related to changes in the CMS. We shared these recommendations with the CMS team who had plans to run some AB tests to assess the value.
4 Ready to Start
4 recommendations were ready to start. They need to be assessed based on resources and time availability.
13 to be prioritised
Due to high number of recommendations, number of projects in flight, and resource limitation, some had not been prioritised.
Outcomes and key results
Multiples bets increased from 58% to 63%
Multiples placed from Accas tab increased from 38% to 65%
Conversion to bet increased by 3% from Accas tab
How we worked
User Story created in Target process
Working group assigned (PM, PO, DM Design)
Kick Offs
UX and UI
Design review/critique
User Testing
Design assets
Cross-functional collaboration
Reflection
The value of storytelling
IA was not widely known topic for the business. Even the CMS team who decide how to categorise and organise content, including selecting taxonomy were not very familiar with the breath of IA and the impact it could have. When it came to the recommendations specifically about reducing duplication, combining sections and renaming labels we initially received a lot of push back.
Storytelling and the way in which we demonstrated the problems we identified was key to persuading the teams we worked with to make the changes.
One problem we identified was a tab had over 100 matches. We used a very simple but clever way to demonstrate why this was a problem by showing how physically long this page is with a measuring tape beside it.. This grabbed everyones attention. We then reviewed the data and were able reduce the page length by 50% without impacting revenue.
MAINTAINING MOMENTUM
The project gained significant traction immediately after the recommendations were announced. However, as we began tackling more ambiguous and technically challenging problems from the backlog, and as team members were pulled into multiple projects, our resources and availability dwindled. Consequently, many of the recommendations were left pending in the backlog, and everyone's focus shifted.
In retrospect, I should have taken the lead in maintaining momentum and the commitment to implementing the recommendations. I could have achieved this by providing regular updates on our progress, highlighting the impact of the changes, and incorporating qualitative feedback from user testing.