DISCOVERY RESEARCH · GENERATIVE + CONCEPT TESTING
Would agents give up their workaround to enter data our way?
The business wanted agents to enter their data into a new system first, adding a step to a workflow they already found painful. This study tested whether that trade could ever be worth it to the people being asked to change.
METHOD
Interviews · Survey · Concept test
PARTICIPANTS
Separate group per phase
CONTEXT
B2B enterprise platform
PHASES
3
01
THE PROBLEM
Listing data was being pulled from MLS (Multiple Listing Service) systems after listings were already posted. The business wanted to flip that order and capture the data directly from agents first.
The upside for the business was clear — owning the data at the source rather than renting access to it downstream.
But there was a catch that made this a research problem, not a build problem: entering data into a new system meant adding a step to a workflow agents already found painful. Several internal concepts existed for how the tool might work, but none were grounded in how agents actually captured and moved data today.
A tool that serves the business but gets ignored by agents solves nothing. The question was whether it could create real value for the people being asked to change their habits.
02
MY ROLE
Sole researcher across all three phases. I designed each phase, recruited a distinct participant group for each, wrote the interview and concept-test guides and the survey instrument, moderated the qualitative sessions, ran the analysis, and delivered findings and direction to the product team.
03
RESEARCH QUESTIONS
What does the current listing-data entry process actually look like for agents?
Where does that process break down?
What would make entering data meaningfully easier — enough to change behavior?
04
APPROACH, AND WHY THREE PHASES
Each phase answered a different kind of question, and each was designed to de-risk the next.
PHASE 1 · GENERATIVE
Interviews
One-on-one interviews with agents across different states, market types, and transaction volumes, to map the current workflow and where it broke down.
PHASE 2 · QUANTITATIVE
Prioritization survey
A survey to identify which listing-data fields agents considered most important, and in what order — evidence that directly shaped the entry-form concept.
PHASE 3 · EVALUATIVE
Concept test
Two competing concepts tested with a fresh group: a single universal standardized form versus dynamic entry pages tailored to each agent's state and MLS.
Why interviews first?
A survey could have told me how many agents found entry frustrating — but I needed to know why, and what a solution would have to look like to actually get adopted. Workflow context and mental models require a conversation.
Why a varied sample?
Listing volume was spread deliberately, from occasional sellers to high-frequency agents. A uniform sample would have produced findings that fit only one kind of agent — and volume turned out to change everything about how much pain a person felt.
Why concept-test two options?
Rather than validate a single favored idea, I put a standardized form against tailored dynamic pages, so the choice was grounded in agent response rather than internal preference.
05
HOW I ANALYZED IT
Each phase used a different analysis suited to its data. Interviews were coded thematically to surface recurring breakdowns in the workflow. The survey was analyzed as response distributions to rank data fields by importance. The concept test combined preference counts with the reasoning behind each choice, so the recommendation rested on why agents preferred one approach, not just that they did.
Illustrative reconstruction. Agents re-entered the same listing data — an address alone might be typed three or four times — across disconnected systems, with no single source of truth.
06
WHAT I FOUND
The workflow was built on repetition
Agents were entering the same listing data into four or more systems, one after another. A single field like the property address might be typed three or four times across the MLS, marketing tools, and transaction platforms.
Every agent described the process as time-consuming and error-prone, with no place the data lived reliably.
And the frustration wasn't only the time. It was re-entering the same information into disconnected systems at every stage, with no continuity between them.
Familiarity beat standardization
In the concept test, agents responded far more positively to dynamic entry pages tailored to their state and MLS than to a single universal form. The dynamic approach felt familiar enough to adopt without forcing them to change how they already worked — and adoption, not elegance, was the whole game for a tool asking agents to add a step.
07
RECOMMENDATION
The research pointed the product in a specific direction. Build the capture experience around dynamic, familiar entry rather than a one-size form. Make the core value the elimination of duplicate entry, with data flowing bi-directionally across an agent's systems rather than trapping it in one more silo. And design the data-capture model to preserve agent-controlled submission, the only terms on which agents would participate.
08
IMPACT
The research gave the team a defensible direction and, more fundamentally, a reason for the product to exist.
The central risk had been whether agents would see enough value to enter data directly rather than defaulting to their MLS. The study answered it: agents resonated strongly with cutting duplicate entry, and wanted bi-directional communication across their systems.
There was overwhelming support for something that would make data capture more efficient — and in some cases automatic — which is exactly the value the business needed to justify building it
09
WHAT I'D DO DIFFERENTLY
Brand and market coverage
The interview sample spanned states and volumes but represented a limited slice of brands. Broader coverage would have tested whether the workflow patterns held across different agent communities.
Larger concept-test group
A bigger sample in the concept phase would have added confidence to the preference signal and let me segment response by agent type.
Earlier secondary research
Some of the constraints agents raised — credential ownership, liability — could have been anticipated with stronger desk research up front, sharpening the discussion guide before the first session.