%% DATAVIEW_PUBLISHER: start
```dataviewjs
const table = dv.markdownTable(
["Metadata Field", "Value"],
Object.entries(dv.current())
.filter(([key]) => {
const allowedKeys = ["tags", "post_status", "date_modified"];
return allowedKeys.includes(key);
})
.map(([key, value]) => {
try {
// --- Key Renaming ---
if (key === "post_status") {
key = "post status";
} else if (key === "date_modified") {
key = "last modified";
}
// --- Value Reformatting ---
if (key === "last modified") {
try {
const parsedDate = moment(new Date(value));
if (parsedDate.isValid()) {
value = parsedDate.format("YYYY-MM-DD");
} else {
console.log("Invalid date format:", value);
}
} catch (error) {
console.error("Error parsing date:", error);
}
} else if (Array.isArray(value)) {
value = value.map((item) => "#" + item);
}
return [
key, // Removed icon prepending
key === "tags" || key === "post status" ? value.join(" ") : value,
];
} catch (error) {
console.error("Error processing metadata:", error);
return [key, "Error"];
}
}),
);
dv.paragraph(table);
```
%%
%% | Metadata Field | Value |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --- |
| tags | [#user-research](app://obsidian.md/index.html#user-research) [#prototyping](app://obsidian.md/index.html#prototyping) [#product-design](app://obsidian.md/index.html#product-design) [#product-strategy](app://obsidian.md/index.html#product-strategy) | |
| post status | [#article](app://obsidian.md/index.html#article) | |
| last modified | 2025-02-04 | %%
%% DATAVIEW_PUBLISHER: end %%
![[mmm-0_0-cover.png | Mixed-methods Modeling — cover image]]
>[!summary|no-t] Summary
>>[!summary|problem txt-ss] Problem
>>To unlock market share in the $25B promotional products industry, the business needed a targetable, foundational understanding of its customers' needs, discovery processes, and constraints across key industry verticals.
>
>>[!summary|role txt-ss ] My Role
>>As **Lead UX Researcher**, I led a cross-functional tiger team composed of one UX Content Strategist, two UX Architects, and two Associate UX Designers. I was responsible for the end-to-end research strategy, including mixed-methods study design, execution, and synthesis. I also acted as the liaison with the business intelligence team, with whom I collaborated on the survey component and the development of the quantitative lookalike model.
>
>>[!summary|outcome txt-ss ] Outcome
>>Developed and socialized six foundational, data-driven personas. These artifacts created a shared, empathetic understanding of the customer across the organization and became instrumental tools for UX briefs, workshop facilitation, and strategic decision-making.
# the premise
Custom Ink had seen a significant shift in its customer base toward large organizations, but lacked a deep, actionable understanding of this segment. As the lead researcher on a cross-functional tiger team, I designed and executed a large-scale, mixed-methods foundational study to build this understanding from the ground up.
# uncover
>[!column|flex 2 no-t]
>>[!logs-point|no-i ttl-c txt-c ttl-b txt-s] 28
>>in-depth interviews
>
>>[!logs-point|no-i ttl-c txt-c ttl-b txt-s] 3,300
>>people surveyed
>
>>[!logs-point|no-i ttl-c txt-c ttl-b txt-s] 1,184
>>atomic observations capture
>
>>[!logs-point|no-i ttl-c txt-c ttl-b txt-s] 18
>>behavioral pattern variables identified
Our goal was to understand the promotional product needs of customers in large organizations (CILOs) across three key verticals: Tech/Finance, Healthcare, and Education. The methodology was a robust mix of qualitative and quantitative research. We conducted **28 hour-long, in-depth interviews** and fielded two large-scale surveys, gathering **~3,300 total responses**. All qualitative data—over 1,700 minutes of video—was meticulously coded and organized into an atomic research repository in Notion, creating a searchable database of 652 evidence-enriched insights, or "nuggets."
| mmm-1_1 | |
|:------- |:---:|
| ![[mmm-1_1-quant_meth.png]] | **figure mmm.1:** overview of quantitative data gathering |
| mmm-1_2 | |
|:------- |:---:|
| ![[mmm-1_2-qual_meth.png]] | **figure mmm.2:** overview of qualitative data gathering |
# converge
The path from raw data to actionable personas was a systematic process of synthesis. The cornerstone of our analysis was **Behavioral Pattern Mapping**, a method adapted from Kim Goodwin. We identified 18 key behavioral variables—from "Propensity Towards a Shared Platform" to "Account Management Style"—and mapped each of the 28 interview participants against these spectra.
| mmm-2_1 | |
|:------- |:---:|
| ![[mmm-2_1-bpm_vars.png]] | **figure mmm.3:** 18 behavioral pattern variables used to map and cluster interview participants |
This process revealed six distinct clusters of behaviors. Instead of grouping users by their job title, we grouped them by _how they actually worked_.
| mmm-2_2 | |
|:------- |:---:|
| ![[mmm-2_2-persona_clustering.png]] | **figure mmm.4:** the behavioral pattern mapping process |
# generate
The six personas were brought to life by enriching the qualitative archetypes with quantitative data. In collaboration with the business intelligence team, I developed a **quantitative lookalike model**, matching survey respondents to our persona clusters based on their responses. This allowed me to build out detailed profiles with data-backed visualizations—which I created myself—for things like top use cases, discovery channels, and desired product attributes.
| mmm-3_1 | |
|:------- |:---:|
| ![[mmm-3_1-card.png]] | **figure mmm.5:** a mixed-method persona card |
A key deliverable was the **Quality of Experience (QoE) Map**, which charted the emotional journey of each persona across the key milestones of ordering custom gear, powerfully highlighting critical pain points and moments of delight for each segment.
| mmm-2_2 | |
|:------- |:---:|
| ![[mmm-3_2-qoe.png]] | **figure mmm.6:** cross-journey average Quality of Experience (QoE) for customers that look like our personas|
# reflect
## Strategic Impact
- **Behavior beats demographics.** Grouping users by their behaviors, not just their industry or title, unlocked a much deeper and more actionable understanding of their needs.
- **Mixed methods are a superpower.** The qualitative interviews gave us the "why" behind the behaviors, while the quantitative survey data gave us the scale and confidence to know these weren't just anecdotes.
- **A research repository is a force multiplier.** Building our insights as atomic "nuggets" in Notion allowed for near-infinite composability and made the data accessible to stakeholders long after the initial study was complete.
## business outcome
- The six personas provided the entire organization with a shared, data-driven language to talk about the customer.
- They became foundational tools used in UX briefs, design workshops, marketing campaign planning, and sales strategy.
- The research identified clear, low-hanging opportunities for unlocking market share by better serving the specific needs of these high-value customer segments.
![[contact#^83635d]]