Generative Research for PYRAMID's Ecommerce Redesign

In the discovery phase of a website redesign, I lead user and stakeholder research to understand the current product state and future needs.


Project dates: January-May 2018 homepage



PYRAMID is Lehighton, PA’s largest sports and fitness club serving Carbon, Schuylkill, and Lehigh Counties. As my first project with PYRAMID, I was hired to conduct research and to implement a new website for the club. Stakeholders wanted to increase online sales, and to have their website serve as the primary point of contact with their customers.


PYRAMID had never conducted research before, meaning that generative research would both offer insights about members and introduce stakeholders to UX strategy. My role involed other design and development tasks, so research composed about 3/5 of my working time blocks.


Stakeholder Interviews

To kick off the project, I recruited the four leading roles at the company for semi-structred, conversational interviews. My goal was to understand the workflows of internal systems; where stakeholders' pain points were, and what might help them in the future. I also wanted to gauge their perceptions of members.

They offered valuable feedback about ZenPlanner, their member-management system and SpeedMobi, their WYSIWIG website manager. They clearly maintained close relationships with their members, which could provide additional insights. homepage
Sample stakeholder interview questions

Persona Workshop

To leverage stakeholders' knowledge of their members, I led a participatory workshop where we constructed proto-personas. These were not expected to be representative clusters, but instead a jumping-off point for future refinement. My hope was that these would help me understand who I might encounter in interviews.

Participants had 15 minutes per persona to build a total of five that they felt captured their membership base. These included personal attributes and attributes dealing with their relationship to PYRAMID. homepage
Proto-persona of John, a senior member

Member Interviews

To inform customer journey maps, stakeholders connected me with members who I recruited for 20 minute interviews. These interviews followed a walk-and-talk approach on treadmills or around the club.

I wanted to understand how members' interacted with PYRAMID online, where their pain points were, and where else they shopped online. homepage
Questions for members

True Intent Study

Using a Drift chatbot, I conducted an unmoderated true intent study to explore the customer journey in real-time. The chatbot asked visitors why they were visiting PYRAMID's website, and later about their outcomes & experience.

I wanted to understand how members' interacted with PYRAMID online, where their pain points were, and where else they shopped online. homepage
Chatbot UI elements

Card Sorting Workshop

A noticable area for improvement on PYRAMID's website was its disjointed information architecture. Ignoring duplicate pages, PYRAMID had 15 pages on its website that needed organization. To gather testable approaches to organizing those pages, I built a Trello board with cards representing each page and product. Participants were asked to create as many columns as needed, name them and place the product/page cards underneath them. homepage
Chatbot UI elements

Synthesis and Deliverables


From the data collected in member interviews and true intent studies, we organized our personas into three major archetypes, inspired by Nielsen Norman Group's research on ecommerce personas. homepage
Summaries of each persona

Customer Journey Maps

To understand the differences in puchasing behaviors between these personas, we mapped their journeys from discovery to purchase. homepage
Journey map of Jake, the persona interested in a quick purchase

Affinity Diagram

Using insights from both member and stakeholder I constructed an affinity diagram of needs, organized by role. This was important when evaluating new ecommerce platforms, in order to quickly evaluate whether they would support our needs. homepage
Affinity diagram

Future Research Questions

Since I would be conducting evaluative research as the website was being updated, it was important for me to test hypotheses that arose during generative research. homepage
Research Questions


Next Steps

With generative research complete, I would begin the process of evaluative research in tandem with the implementing the new website. I detail that process in my Evaluative Research and Implementation of Ecommerce Website case study.


This research revealed the incredible room for impact at PYRAMID, so much that it's enabled me to work with them for two years longer than we initally expected. Stakeholders now look to research as a tool for reducing through bias when evaluating situations, and we continue to consult and refine these deliverables. This shows the effects of getting quick wins and involving stakeholders throughout the research process.

In retrospect, I would opt for using a tree test instead of a card sorting study to evaulate navigation hierarchy. At the very least, I would have raised the number of participants to 25-30 per cluster in order to eclipse 95 percent confidence. However, 90 percent confidence certainly isn't bad, given the research context.