Summary
A story about how to quantify and use survey data to create audience clusters that are statistically representative of the people that use your online product or service. Edgar joins the dots between multiple research/pattern finding/storytelling techniques, such as proto-personas, affinity mapping, survey creation, quantification of qualitative data, organic clustering of audiences, user interviews, journey maps, and business modeling canvas. It makes research tangible and actionable: instead of being an end-point that provides stakeholders with insights, it takes stakeholders through an inclusive and structured process they can be part of.
Key Insights
-
•
Research often fails when stakeholders are not involved or don't trust the data sources and methodologies.
-
•
Treating research as a product, with stakeholders as users, drastically increases engagement and usability.
-
•
Using proto-persona workshops with cross-department stakeholders helps generate unbiased, relevant survey questions.
-
•
Clustering survey data via Python and correlation matrices can reveal meaningful user segments with statistical significance.
-
•
Statistical validity (e.g., 95% confidence at 5% margin) is critical for stakeholder buy-in in large user bases.
-
•
Behavioral and demographic clusters need to be augmented with emotional attributes to create actionable personas.
-
•
Affinity mapping large qualitative datasets reduces noise and distills coherent narratives and pain points.
-
•
Mapping pain points to measurable hypotheses enables prioritization based on user impact and implementation effort.
-
•
Divergent and convergent workshops using tools like the business model canvas help explore and refine value proposition delivery.
-
•
Iterative, lean research with frequent stakeholder feedback helps identify and minimize bias effectively.
Notable Quotes
"Research should be done the same way products are built, with stakeholders as your users."
"If you frame the research in terms your stakeholders understand, they will become research-hungry."
"We don’t make big deliverables because they create friction; instead, we create shared understanding through conversations."
"Proto-persona sessions helped us avoid departmental bias and uncover the questions stakeholders really cared about."
"We used a 0.8 correlation threshold to tightly cluster similar survey responses, balancing cluster count and inclusion."
"Clusters alone aren’t personas; we turned clusters into interview screeners and then emotional personas."
"We wrote hypotheses as we believe this pain point matters, here’s how we test it, and how success looks."
"Different stakeholders tackle the delivery problem differently, which gave us necessary divergent thinking."
"Biases surface as you iterate, so lean approaches help you find and fix them quicker."
"Marital status mattered because some people buy cars for their spouses or kids, not just themselves."
Or choose a question:
More Videos
"It’s how you think about the relationships between people and these artifacts that makes a design system."
Dan DonaldDesign Systems as a Vehicle for Systemic Change
June 1, 2023
"Consistency between control labels and prototype instructions matters a lot, because mismatches can cause big problems for alternative navigation."
Sam ProulxPrototype Reviews, People With Disabilities, and You
October 1, 2021
"Talking about outcomes first is a really powerful technique for people who have had bad experiences."
Adrian HowardSturgeon’s Biases
September 25, 2024
"Everyone has a plan until they get punched in the mouth, said Mike Tyson."
Adam ThomasSurvival Metrics – Making Change in a Fast, Data-Informed, and Politically Safe Way
December 6, 2022
"Design impact doesn’t exist separate from the work with others – it’s part of the initiative's success."
Peter MerholzThe Trials and Tribulations of Directors of UX
July 13, 2023
"Being trauma-informed means caring for yourself, forgiving your mistakes, and moving forward with humility."
Matt Bernius Sarah Fathallah Hera Hussain Jessica Zéroual-KaraTrauma-informed Research: A Panel Discussion
October 7, 2021
"Stories don’t work by extraction, people tell them and you listen."
Kwabena Opoku Leonie Annor-Owiredu Sam LadnerMethodological toolkit for unique research impact
March 11, 2026
"I was not able to come up for them to be involved so I had to learn how to structure it in a way that we could share how this was important for the team."
Benjamin RealMaturity Models: A Core Tool for Creating a DesignOps Strategy
October 1, 2021
"In Coda, you can create different views for the same data, like one for designers and another for engineers, and all stay synchronized."
Scott StephensThe Next Generation in DesignOps Toolsets
July 28, 2022