Does a research repository make us better designers?

From Lucille Ritter von Marx (Lead Product Designer at leboncoin) & Mathias Frey (Design Manager Thiga - Product Management. Redefined.)

We believe that any scaling product organization faces challenges in structuring its research practices, especially in centralizing its user learnings. Whether you are a Product Designer or a Product Owner, you may be wondering how you can :

• Easily find the learnings collected in a previous study
• Capitalize on the learnings of other teams in your organization to avoid doing the same work over and over again
• Prioritize user problems for your next discovery phase

We would like to share with you how we implemented tools and best practices at leboncoin that allowed us to better structure and prioritize our user research.

Our context at leboncoin: launching a new marketplace for professional sellers

Leboncoin, the French leader in classified ads, has traditionally been a marketplace where consumer goods are sold between private sellers and buyers. Facing the rising power of Amazon, leboncoin decided in 2021 to open its marketplace to professional sellers as well. However, professional sellers’ sales experience was not adapted to their needs. Therefore, a crew called Marketplace Pro was born with the mission of becoming the most efficient and economical transactional marketplace for professional sellers who sell second-hand and refurbished items.

The Marketplace Pro crew is composed of :

• 3 feature teams: 3 product owners, 3 engineering managers, 23 engineers, 4 QA analysts
• 2 Product Designers, 1 UI Designer
• 1 Product manager, 1 Lead Tech and 1 Design Manager

A lot of user research was conducted and difficulties began to emerge…

Discovery and delivery activities are carried out in parallel, on a quarterly basis at leboncoin. Roadmaps therefore allow several weeks for the product and design teams to plan and conduct their user research activities.

After 6 months and 20 user studies conducted, we started to see some problems emerging…

We organized a workshop between the designers of the crew to discuss the pains we were encountering with user research.

We identified three major problems:

😅 A need to gather all of our questions on the remaining user needs

😓 Difficulty in finding insights collected from past research

😬 Since we collected lots of insights, we had trouble prioritizing insights to be addressed

We set 3 goals:

🎯 Better anticipate our next discovery phases

🎯 Allow anyone on the Marketplace Pro crew to easily find all of our research findings

🎯 Easily suggest optimisations to Product Managers, through structured, weighted and therefore more credible insights.

​​To meet these goals, we decided to create a research repository for our Marketplace Pro crew. We wanted to test this as an MVP to ensure the value it brings and learn from it on a smaller scale before potentially rolling it out to other crews or to the entire UX team.

To do this, we turned to Airtable because this tool offers tagging elements and thus allows an advanced and quick search. Airtable is a tool that people at leboncoin already knew about, so there was no need for training.

Prioritize user research through the creation of a ‘discovery backlog’

Building a Marketplace on leboncoin’s website by connecting professional sellers and buyers is quite a challenge: a professional seller does not manage his sales the same way a private seller does. We decided to start working on the ad depository and the order management experiences.

By adopting a user-centric posture, we have to ask ourselves a lot of questions: How do professional sellers manage their orders? How do they use the tracking number? What is their desired payment frequency? What are the real needs of professional sellers in terms of restocking?

Faced with the multitude of questions that we were asking ourselves, we decided to create a backlog allowing us to list all these questions related to the uses and needs of our users.

To help us see more clearly, we have categorized these questions by :

• User journey on which it intervenes (e.g.: order management, stock management, ad deposit, etc.)
• The way of obtaining the answer (e.g.: user interview, data analysis, quantitative study, etc.)

Once all our questions were filled in, we set up a criteria allowing us to prioritize them: each question has a “Risk” score from 1 to 5 based on both the priorities of the roadmap and the need we have to get learnings to that question. A question at 5 is said to be risky and deserves to be answered quickly and 1 being much less of a priority.

A status is associated with each question: To be processed — In progress — Answer obtained

Benefits for our team:

👉 Having a common base of questions on which to reflect between Product and Design

👉 This allowed us to prioritize the questions we were asking ourselves based on the risk and status in order to decide what we would address for future user research.

👉 Capitalizing on the questioning of each team to conduct common user research and thus avoid duplication

What we could have done in a better way…

🤔 Be more rigorous in prioritizing each question: the risk score could be more accurate if it contained the notion of reach (how many users are affected by the question we are asking) and impact (if we don’t get the answer to this question, does it have an impact for our future solutions?)

Sharing the results of our research with as many people as possible by centralizing all our studies

Between the number of studies we conducted over the months and those conducted by other teams than ours, it became difficult to find our way around. We took the initiative to centralize all our studies in one place.

To facilitate access to the information, we have decided to categorize our studies as follows:

• The name of the research (e.g.: Behavior of buyers towards professional sellers on leboncoin)
• The link to the study report
• The type of research (User test, quantitative study, user interview, surveys, etc.)
• The journey concerned (e.g.: posting an ad, managing orders, searching for an ad, etc)
• The persona impacted (sellers, buyers, etc)
• Date of the study (e.g.: Q3 2022, Q4 2002,…)
• Status (Past study, In progress, To come,…)
• The designer(s) responsible for the study

Each study contains the link to the presentation of the study on Google Slides.

Benefits for our team:

👉 Easily find previous studies related to our users

👉 Avoid losing user knowledge as time goes by

👉 Allow anyone in the organization to have access to all previous research conducted (it is very useful for newcomers or evangelizing outside the team)

What we could have done in a better way…

🤔 Make sure to establish a reflex for all team members to ensure that all studies are available in this space.

Make it easier to prioritize insights by centralizing and weighting them

Before having this amazing research system, we had to access each of our studies presentations to find the key insights, and weighing them was difficult…

1. Centralizing all the facts observed during a user research

First, let’s define what a fact is: a fact is the “what”, the direct information, the element observed during a study (user test, user interview, survey).

The facts observed during a study are stored in this space. Each fact is sorted by its title, the element tested, the user quote, the user journey, the persona impacted, the recurrence of the fact, a screenshot and is linked to the concerned study.

👉 We noticed that this allowed us to save time during our studies reports and convert them into insights afterwards.

2. Building our insights repository

We based our insights repository on the Atomic Research Principles:

1. Atomic: the system should contain insights tagged with personas, observations/user quotes, recommendations.
2. Visible: insights must be visible and/or audible to people who go on the repository
3. Accessible: must be accessible to everyone in the company
4. Verifiable: the system must carry the voice of the users and must contain 0 bias
5. Iterative: it must be fed because it is linked to the life of the product
6. Connected: personas, observations/user quotes, recommendations must be connected to each other.

An insight is the “why”, the contextualized explanation of the fact. An insight can answer several facts. It can be a problem as well as a positive finding. → “This fact(s) make us say that X” As an example: for an hypothetical research, a collected fact could be: ”70% of users are between the ages of 25 and 35.” The related insight could be: “Users in their late 20s to early 30s are drawn to the product because it fits into their busy, on-the-go lifestyle.”

Insights are formulated from one or more facts observed in one or several studies conducted. An insight cannot be created without facts.

The more facts associated with an insight and the more experiments conducted, the more confidence we have in it.

If an insight is associated with not enough facts, we have two possibilities:

• If we consider that the insight has little impact on the product, we do not hold it against it.
• If we consider that the insight has an impact on the product, we can conduct additional research to increase our level of confidence and thus better take it into account.

To facilitate access to the information, we have decided to categorize our studies as follows:

We have categorized each insight as follows:

• its title,
• its type (negative / positive / neutral / opportunity),
• the associated fact(s),
• its status (to do / in progress / done),
• the concerned user journey,
• the concerned study(s)

Benefits for our team:

👉 Researching insights is now really easy thanks to Airtable’s advanced filters! ✌️We usually need to find insights when preparing for an ideation workshop or start working on a new project to make sure we stay user-centric.

👉 It reinforces the fact that we can no longer imagine product optimization without insights.

👉 This has reinforced our work and posture as user researcher in the eyes of product managers.

What we could have done in a better way…

🤔 Better highlight the number of associated facts and the different experiments. It is a bit difficult to quickly notice the number of facts and experiments associated with an insight…

🤔 Connect our insights to recommendations: we have not included recommendations related to negative insights because we work on them outside of this repository. We can ask ourselves if it would be valuable to include them in our insights repository.

What we have learned…

Let’s go back to the objectives we set by highlighting the impact and benefits we have achieved and also some lessons we have learned:

1. Better anticipate our next discovery phases

We saw a better collaboration between Design & Product to agree on priorities. It also strengthened the collaboration between the designers of the same crew: the fact of mutualizing our questioning allowed us to think and lead together these phases of research.

2. Allow anyone on the Marketplace Pro crew to easily find all of our research findings

All crew members now have easy access to our studies through a single entry point.

3. Easily suggest optimisations to Product Managers, through structured, weighted and therefore more credible insights

The insights are easily accessible, more structured and are weighted. They have helped us to optimize our product. However, we would like to drive Product teams to properly use this tool because we believe that a better adoption at their level could allow us to make sure that our repository learnings will fully feed the roadmaps. To this end, we constantly need to feed and optimize this repository to keep insights fresh. This seems difficult for the moment because we rely for now on the product designers’ discipline. However, we think that scaling this type of approach requires to assign this task to a dedicated user researcher. Unfortunately, there is no user researcher yet at leboncoin, so our product designers conduct and own their research.

To conclude, the studies centralization seems to be more valuable because it requires less investment than an insights repository and allows us to quickly find all the studies conducted. This makes us want to deploy it to the entire design team at leboncoin soon!

We would be curious to know your opinion on this subject, do not hesitate to express yourself in the comments in order to further enrich thinking on the subject! ✨

From Lucille Ritter von Marx (Lead Product Designer at leboncoin) & Mathias Frey (Design Manager Thiga — Product Management. Redefined.)