A full redesign of the internal tools one of Germany's fastest-growing real-estate companies relied on every day, plus a fourth built from scratch. Grounded in hands-on research, and along the way I helped set up a stronger design and research culture inside the company.
McMakler is a German real estate company built around in-house agents backed by its own software. When I joined, the tools those agents used every day were fragmented. Data sat inside each system on its own, and there was no reliable place to see what was happening across a whole transaction.
My work covered Immoforce (central property management), Alla (broker scheduling), and Brokerforce (broker task management, built from scratch). None of them talked to each other. Agents kept their own spreadsheets to track what the software should have shown them.
The PM's brief was simple: a basic UI refresh, a few flow improvements, reuse the existing MUI components, tight timeline, no research phase.
But the product showed me something the brief had assumed away. The app's logic wasn't written down anywhere. The business rules, the required field order, which states set off which workflows, none of it lived outside the heads of the people who'd built or used the system. The brief was asking me to improve a UI without knowing what the UI actually did.
I proposed a one-week research sprint, not an open-ended discovery phase. One week, one deliverable: the logic written down, findings from users, and one direction to test. Three days in person with customer care, SWAT team leads, and ops, plus FullStory session recordings to check what users said against what they actually did.
When someone pushed back in a review, I had a session recording to show them. 75% of designs got approved on the first round, because they were grounded in how the system actually worked.
Not 8 seconds, 80. Agents used Alla to manage their calendars and coordinate property viewings. The load-time problem was a backend issue that engineering handled separately. My job was making the product worth using once it finally loaded.
Three design iterations:
Alla was shut down about three months after the redesigned version launched. Adoption was low. Some agents stuck with tools they already knew, and part of the scheduling work happened over the phone, not in any software.
What I took from Alla: we hadn't asked enough agents, early enough, whether they'd use any tool for this at all, or what it would take to replace what they had. That question should have come before the third iteration, not after. Low adoption after a redesign is usually both an adoption problem and a research problem.
Broker workflow was running on paper and in spreadsheets. Tasks got assigned, tracked, and closed by hand. The goal was software that could replace all that while still fitting how brokers already thought about their work.
With brand-new enterprise tools, people won't adopt software that thinks about the work differently than they do. The design had to match the mental models already in place, then improve the mechanics underneath.
This ran alongside the Immoforce redesign, so I was switching between reworking an existing product and building a new one from scratch. The modular patterns and component decisions from Immoforce fed straight into Brokerforce, so agents using both tools weren't learning two different visual languages.
McMakler was the first job where in-person research was a regular habit for me, not a one-off phase. Sitting with the customer care team while they used Immoforce showed me things a 30-minute interview would have missed.
Watching people work reveals the shortcuts, workarounds, and little hesitations they never think to mention, because they don't see them as problems. To them it's just how the thing works. That changed how I run research. The most useful things I learned at McMakler, I learned by being in the room while someone did their job.