🤙🏼 👩🏻‍💼 👋 How we at Cyan improve a product through user experience research 🔱 🆑 👶🏼

Hello, Habr! Cyan pays attention to interesting technical solutions. They are also interested in how users will live with the results of these decisions, how much these solutions are needed and how they will help people who are looking for or selling real estate. In this article, I’ll tell you how we built the user experience research process and what it gave us.

What we study

The purpose of the research is to determine what needs to be done so that the product meets the needs of the audience and helps to successfully cope with the tasks.

You can, of course, go this way

Formally speaking, User Experience Research is the study of human behavior and reactions when using a product, including a non-digital one. But this is the term, but what are we really studying?

Motives and human needs

All people who come to our service want to solve some problem, formed by life circumstances. Someone moves to another city and wants to rent an apartment, someone is waiting for a replenishment and is looking for a new home for purchase. We study such situations in order to offer the best options for real estate and help with the process of buying or renting.

Application Use Context

We pay a lot of attention to the environment and conditions in which our products are used. Therefore, we study what happens around our users when using Cyan, and how the surrounding reality affects usage scenarios. Whether they are looking for real estate on the road, with children in their arms, in the office or at home - all this is important to know in order to adapt sites and applications to the needs of users.

Analysis of familiar services, products and typical tasks

Every day, our users open many sites and applications, look at terminals and ATMs and form habits or patterns of access to digital products. When designing an application, we need to understand how users see the digital world and what metaphors they understand so that our products are easy for them to learn, behave predictably and not make them think hard how to do this or that action.

Usability

There are attributes by which sites and applications are considered “good” or “not very”. Digital products should be easy to manage, pleasing to the eye, so that the user does not have to think long about their actions. Therefore, we pay much attention to the study of the usability of Cyan itself. In order to evaluate and then improve the usability of our products, we invite users to our house and observe how they work with our sites applications, what emotions they experience, what problems they encounter and how they solve them.

Why do we need research

We have a large development team and 14 million active users. Each of the team members has their own view on the development of the product, and we can try to independently determine what the product should be and decide for users what is important and what is not.

The bitter truth is that we are limited by our own life experience. And there are many users, and they have other patterns of using digital products, other ideas about the search, transactions and the work of realtors.

Research helps to understand how the world works outside our comfort zone, and combine the experience of a large number of people to create a convenient product that helps to solve both simple problems and cope with "narrow cases".

Place of research in product development

Cyan has always understood that user behavior needs to be studied. Such initiatives came from our director, Maxim Melnikov, so we did not have to prove the value of research.

At first, however, our research department was “cloudy” - all the researchers sat together, and the products or designers came to the consultation and ordered the necessary research when it was needed. But with this scheme, the research process was delayed.

There are several product lines in Cyan, and each of them is engaged in improving the user experience and developing new features for its audience segment. Here are the largest areas:

The audience. Responsible for search experience in renting and secondary real estate.
Resale. Responsible for working with owners and realtors.
New buildings. He is responsible for working with developers and search experience in new buildings.

Since a certain researcher was not assigned to the direction before, when new projects appeared, each time it was necessary to delve into the context for a long time and study what was happening in the product.

Taking a look at this process, we decided to find a researcher for each of the large areas. We emphasized making the researcher a full-fledged member of the team and increasing his immersion in his part of the product. In this structure, the researcher understands the business goals of the direction, so studies become not just observations made when communicating with users and customers, but discoveries that help to achieve the goals of the direction and build a product development strategy.

The role of the researcher is to convey to the team how people work with the product, show what can be improved in the user experience, and explain what this improvement will bring to the business. In our world, a researcher is the initiator of obtaining and using knowledge about users, as well as the keeper of research methodology.

We will show the process from the inside: how exactly research is organized in Cyan and what results are obtained.

Formulation of the problem

All research begins with the statement of the problem. It all starts with quarterly planning: before the quarter starts, the products decide what they will do with their team in the next three months, checking against the annual goals. Research must be included in the plans. Ideas for this research can come from products, designers, and the researchers themselves.

We do not do inapplicable research, so each idea should be formulated as follows:

We want to research A to make B.

For example, the hypothesis is that our object card is overloaded with information. We want to know what is important for users to see on the card in order to remove unnecessary and redistribute emphasis. So the card will become easier to perceive and the cognitive load on the user will decrease.

Method selection

We teach research customers not to speak the language of methods, but to designate only a question that needs an answer. Thus, the researcher flexibly selects the optimal method: one that does not take too much time, does not require large investments and gives a result that you can rely on when making product decisions.

Speaking directly about research methods, there are two groups: qualitative and quantitative.

Quality methods

This is the most common and easy to organize group of methods. The goal of quality research is to get as many answers as possible to the question “What and how do users do?” and discover the maximum number of patterns, life situations, queries and other components of the user experience. Qualitative research includes usability tests and interviews, as well as focus groups and co-creation sessions.

To speed up the decision verification process, we came up with the practice of regular respondents. Once a week, users come to the office of Cyan for quality research, and we show them a beta version of the features coming up for release. Some might argue that you can work even faster and cheaper by testing layouts. But testing layouts differs from testing the final product, because in the layout it is impossible and disadvantageous to reproduce the level of interactivity that a live assembly provides. This does not depreciate the testing of layouts: if such a method is suitable for finding the answer to a asked question, we will choose it.

With the help of regular respondents, we quickly catch usability problems before the updates are released on the prod and constantly keep our finger on the pulse, receiving a continuous stream of observations of real users. In the course of such studies, we not only test new functions, but also find old bugs.

Quantitative methods

The second group of research methods is quantitative. Now they are less common in the IT industry, because such research requires a broader theoretical base: you need to understand how many and what kind of people you need to interview, where to find them and how to correctly interpret the results. But the small spread of quantitative research is a matter of time and technology.

Quantitative studies are used to answer how widespread a particular phenomenon is and how often situations that we observed in the process of qualitative research are encountered. Quantitative research methods include AB tests, questionnaires, questionnaires and event analysis in web analytics tools (Google Analytics, Yandex.Metrica, etc.).

Additional data sources

In the course of our work, we realized that the totality of knowledge from different sources helps to objectively look at the problem posed and possible solutions.

In addition to the classical research methods listed above, we also study additional sources: contacting customer service, open information on the Internet, reviews and mentions on social networks.

To quickly monitor user opinions about the site, we use the UX Feedback tool, which consistently gives us 600 live comments per week, and this is only the desktop! We do not force users to write reviews, but unobtrusively give them the opportunity to share thoughts in any part of the site. In this way, people write that they don’t like it, or praise us for our work. We have built a process of continuous analysis of reviews from different sources and now each direction can track the dynamics of opinions about its part of the product.

Study progress

We have seen from our own experience that the more forces a person puts into obtaining knowledge, the more willingly he uses it. Therefore, not only the researcher is involved in the research process, but also people who make product and design decisions.

Developers can also follow the progress of the study. They have the opportunity to listen to people who come to our office, and we encourage their interest in respondents - tell us why people act in one way or another, and answer all questions that arise.

When is the time to stop?

A phenomenon called Analytical Paralysis is common in research and analysis. Usually it comes when you are very much afraid to make a mistake, making a difficult decision, and dig in an endless analysis and study of a topic instead of moving on.

As we know, research is aimed at eliminating the risks that might arise when making product decisions based on our own limited experience. I would like to believe that research has a result and a fixed point at which it will become clear that study is enough. But there is no such thing as “enough research.” Usually sufficiency is determined by the moment when you have something “clicked” and the puzzle has developed. Of course, this does not cancel the adherence to the methodology - you must follow all the rules by which the selected method works.

In order not to become paralyzed, we try to identify the factors that cause the greatest risks and to identify high-priority issues. After finding the answer to them, you can move on. You can’t stand still - you need to find the main reference points and continue moving, constantly expanding your field of knowledge and eliminating the unknown.

If you are interested in learning more about the adequacy of research, write in the comments, there is enough material for the whole article.

Results: how they look and how they are applied

Before starting a study, we always agree with stakeholders on what to consider the result of the study. In some cases, verbal retelling and a few tips are enough to fix, while others require the creation of long-reads and large presentations.

The general rules are such that we must get a reliable answer to the asked question and understand that a negative result is also a result. You must always remember that testing can fail, and the hypothesis is not confirmed. Of course, this is sad, but this is the purpose of research - to dispel mirages and show reality, because the sooner we find that we were somewhere wrong, the more time we will have for the right actions.

Here are some examples of artifacts.

Last year, we decided to study the interaction with users in order to understand how we communicate with them: is the communication too obsessive and relevant. As a result, we put together a board in Miro that shows the points of contact with the user:

This visualization helped us to systematically look at our communications and see that the initial hypothesis that we remind ourselves too often of the user is not true. But looking at the feedback from the site, we also saw growth points: users left comments, which is unclear how to unsubscribe from some mailings, and that they periodically receive irrelevant recommended ads.

Another example of an artifact will be shown by examining the main page. We set out to facilitate our main desktop and rid it of useless information and buttons. We heard that the main page really looks too complicated, so we decided to study the events in Google Analytics. We found that there are few transitions in the extension above the menu, and decided to ask our users how they understand it. The result was unexpected - users did not even pay attention to this part of the main part, it simply created a visual noise. Specially drawing the attention of users to this part of the page, we received comments on each of the items and designed them into a presentation. Here is an example slide:

Based on this study, we finally made sure that we need to redo the main page, and to this day we continue to move in this direction.

PS: In this article, we briefly covered what the researchers at Cyan do and how we built the research process. We hope that we have done it simply and clearly. If you are interested in participating in our research, take a short survey .

How we at Cyan improve a product through user experience research