People Behind Products: Meet Ayse Delikaya, Junior UX Researcher

Ayse, you have been working at ImmoScout24 for more than 2 years now. What is your current job and why did you decide to pursue a career in UX Research?

I have a Psychology background and feel in my “natural habitat” at ImmoScout24 as a UX Researcher. I found my passion for research very early on at university. That’s why I always worked in academic research during my studies. In fact, I worked in various faculties of biopsychology, medical psychology and integrative life sciences. Initially, my goal was becoming a professor of biological psychology because I was very much in love with Psychoneuroendocrinology. However, I quickly realized that the academic research conditions weren’t good for me. I was missing the aspects of practical application; everything was so regimented and fixed and there was little room for creativity. In addition, I got bored after a while, because it was dreadful to work on one study for 5 years.

I started exploring my options since graduation was nearing. I was dead set on wanting to do research but not in the academic field, so therefore UX Research crystallized as the perfect intersection. I tried it out as a working student at ImmoScout24 and loved it. The reason? I am able to study my favorite subject: humans! And I still have the privilege to provide benefits to people in their day to day experiences, instead of over the process of 15 years.

The total rent search is one of the features that you helped developing. What is ImmoScout24 trying to solve by developing this feature?

The total rent filter enables you to get results that are in your real budget. Previously you could search using the net rent, which doesn’t include heating costs, other ancillary costs etc. As a result, an apartment could be advertised as 600€, but through other costs you would realistically end up paying around 850€. We knew our users were not choosing their homes based on net rent, but on the total rent they ended up paying each month. Furthermore, we had feedback from price sensitive users, who saw many objects in their result lists that were exceeding their budget when they took a second look. Because the proportion of ancillary cost in listings varies, estimating the correct net rent that would reliably show the users objects within their actual total rent budget will always result in showing them some objects that are above their budget. We had additional costs for most, but not for all listings, which was keeping us from building a simple filter. Therefore we had to get creative.

How did you and your team develop this feature?

The total rent topic was a major pain point for our apartment seekers and a tough nut for us. My other colleagues had tried to crack it a few times before. When I joined the company as a student I had the opportunity to pick my own challenge, so I decided to focus on our users’ filtering needs and pains in an approach with mixed methods. We already knew of the total rent issue and tried to validate if it still held true in the current heated market. It did. So we explored the actual problem behind the wish for a filter: Who are our price sensitive users, who demands this filter? Why do they want to know the total rent? Why is net rent not sufficient? What other problems do they have? Would this or that solution work? Why yes, why not?

To give you an example, we knew from our price sensitive users about the heated market, which caused FOMO (fear of missing out). Users were adamant about not wanting to miss any objects. That implies that simply filtering out objects without ancillary cost was not going to cut it. We needed to show all available listings that rent seekers want to see. So we came up with the idea to build an estimated total rent model for listings without additional cost info. Together with our Product owner Markus Meixner and our Designer Chris Lüders we built and tested prototypes with rising intricacy.

As soon as we had found an interface that could prove its usability and value in user tests the data science team started doing their magic and building the estimation model. This way we were serving two segments at once: We were solving the pain of our users, we bumped up quality of leads to our landlords and landladies because contact requests were coming more from fitting candidates and it gave landlords and landladies without data a rough estimation on what ancillary cost would amount to. It was a win-win for everyone.

As soon as the model was ready and developer capacity was at hand we tested the feature with a small share of our users. Our goal measures looked very promising, we were pumped! With full roll-out we were still hitting our target measures, adoption was increasing and reviews looked good. In fact, it was so good that it is now in the default setting of our App.

What's in store for the future? Will there be improvements or adjustments for this feature?

The total rent feature has proven to be very successful in its first implementation and effectively solved the problem of our users. It is not an MVP (Minimum Viable Product), but a finished product. That is why there are no adjustments planned for the filter itself. Our data shows that users with a total rent search see less objects, but contact more owners of those objects. This means less time and energy is spent to find your new home or office. Our awesome team is working at all times on making the user experience better and there are exciting new features in the making so stay tuned!

What are other projects that are in your focus at the moment?

I am working with a large team on a multitude of projects that mainly focus on all sorts of communications our platform has with its users. That includes pushes, mails, store notifications, etc. You will hear more from us very soon.