Preventing revenue loss for online stores

How we leveraged User Research + Lightning Decision Jam to enhance a B2B SaaS experience.

Role

Platforms

Collaborated with

TL;DR

Problem

My work

Solution

Impact

Overview

In 2019, marketplaces like Amazon and Mercado Livre were transforming e-commerce in Brazil, driving a 13% growth in marketplace sales—outpacing overall e-commerce—with 95% of online buyers using these platforms.

Our client, an early-stage startup supporting 100+ sellers, launched a B2B SaaS solution to automate payment transfers from these marketplaces.

Their product replaced manual profit-sharing processes and spreadsheets, integrating the marketplace data and generating reconciliation reports from it. Revenue came from each approved transaction.

As one of Brazil's first payment conciliators for marketplaces, the client tapped into a high-demand, low-competition market and positioned itself for rapid growth.

The Challenges

1. High churn

The company had a 21% churn, a high number considering the market's growth.

3. Lengthy onboarding

The software setup was complex due to varying marketplace rules, and the support team spent excessive time training new clients.

2. Reconciliations drop

The company had a 21% churn, a high number considering the market's growth.

4. Lacking user focus

The company had few interactions with their users/clients in their process.

The Design Process

I was inspired by the Double Diamond framework to structure the design process, as it is adaptable and allows for flexibility in tackling complex problems. This process was divided into four phases:

Discovery

To ensure a deep understanding of both the product and its users during the discovery process, we employed a range of research methods:

CSD Matrix

We conducted a 1-hour CSD Matrix workshop with the team to define the Certainties, Suppositions, and Doubts, which helped clarify what we wanted to know and guided our research approach in the next steps.

Internal Data

Analyzing Freshdesk support tickets helped identify usability issues, including users problems with adding marketplace setups and order details.

Competitor Analysis

Although direct competitors were limited to two (TrackCash and Equals), we also analyzed similar products, their value propositions, and how marketplaces presented their software.

Contextual Inquiry

We conducted on-site visits to three sellers, engaging with employees from sales and financial departments, including managers, coordinators, and analysts. I was accompanied by our Customer Success Analyst, which played a crucial role in fostering deeper empathy and understanding from her perspective as well.

Remote User Interviews

Given that most of our clients were located in different cities, we opted for remote interviews to ensure broader reach across Brazil. We spoke with approximately 20 individuals across 11 client organizations, covering a range of roles. Each session included a member of the Technical Support Team, a Sales Team representative

Research Findings

After our research was done, we could summarize our findings into main 3 topics:

#1 Users feel overwhelmed and micromanaged at work

We gained valuable insights into the dynamics of different roles and their work conditions.

Many operated in high-stress office environments with little autonomy, juggling numerous tasks beyond reconciliation processes. Micromanagement was common, with supervisors closely monitoring their work, adding to the pressure.

#2 Manual work was still a thing

We gained valuable insights into the dynamics of different roles and their work conditions.

Many operated in high-stress office environments with little autonomy, juggling numerous tasks beyond reconciliation processes. Micromanagement was common, with supervisors closely monitoring their work, adding to the pressure.

#3 Need for reliable support

We gained valuable insights into the dynamics of different roles and their work conditions.

Many operated in high-stress office environments with little autonomy, juggling numerous tasks beyond reconciliation processes. Micromanagement was common, with supervisors closely monitoring their work, adding to the pressure.

Definition

We used research data to create impactful deliverables that helped the entire team better understand customer pain points—both during the buying process and while using the product.

1. Personas

We identified 5 types of users distributed into 2 groups of personas (management and operational). We wanted to focus on their problems instead of adding too much information.

2. Customer Journeys

After that, we mapped their actions and emotions during this journey, dividing it into five steps:

Ideation

For the Ideation phase, we wanted to generate ideas with the team, however, we had considerable schedule restrictions with different members, and we couldn't wait days to bring everyone together.

Since we wanted to generate ideas to validate later, we decided to run an ideation workshop called Lightning Decision Jam, created by AJ&Smart based on the Design Sprint, adjusted to fit into a 2-3 hour window.

Step 1: Sharing research findings and prioritizing problems

For the Ideation phase, we wanted to generate ideas with the team, however, we had considerable schedule restrictions with different members, and we couldn't wait days to bring everyone together.

Since we wanted to generate ideas to validate later, we decided to run an ideation workshop called Lightning Decision Jam, created by AJ&Smart based on the Design Sprint, adjusted to fit into a 2-3 hour window.

Step 2: Reframing the problems

After that, it was time to reframe them as "How Might We" questions. We've decided to pick the top 3 problems voted by the teamand transformed them into the following questions:

Step 3: Producing and Prioritizing Solutions

Next step, we had the "ideate without discussion" section for each of these questions. After a considerable amount of ideas were generated, we chose the best ones using voting dots and prioritized them using an "Effort x Impact" matrix.

Design

Step 1: Sharing research findings and prioritizing problems

After the workshop, we started with the high-impact, low-effort ideas, most of which were related to customer service and could be implemented without requiring technology or design development. To start delivering value to our customers as quickly as possible, we selected the two most popular initiatives:

• Create a FAQ page featuring answers to the most common customer questions regarding their onboarding.
• Follow up with inactive customers every two weeks to assess their needs and explore ways to reactivate their usage.
  
  

While these tasks were being executed, the tech and product teams concentrated on more complex ideas that required greater effort and strategic planning.

Step 2: Reframing the problems

After the workshop, we selected Ideas with high value and high effort to be discussed, simplified, and prototyped later.

Some of them were transformed into less complex ideas so it could be implemented by the development team to deliver them quicker to our users. Others like the new onboarding process required more planning

Our focus was to reuse some UI components to dedicate the effort to how these solutions would be presented, validating with the development team if it was viable or not:

Notification system: emails and desktop push notifications

Improve reconciliation screen

Self-onboarding process

Conclusion

What have we learned?

• The best solutions aren't always technological. Sometimes, the simplest approach can have the greatest impact.
• This project reminded me of a phrase I deeply resonate with: "The role of the designer is to be a facilitator." Collaborating with people from diverse teams and roles is always a rewarding challenge. It's gratifying to see individuals realize that generating great ideas is not the exclusive domain of the "genius designer" or the "skilled developer," but a shared responsibility that everyone can contribute to—breaking down stereotypes and fostering collaboration.

What could have been better?

• While the Lightning Decision Jam (LDJ) is highly flexible in accommodating varying group sizes, we noticed that having too many participants made it challenging to stay focused during certain stages.
• Some participants were inclined to delve deeper into discussions, even when the approach emphasized avoiding debates and focusing on "selling" ideas.
• Allocating more time for usability testing on proposed solutions before moving into development would have helped validate ideas further.

Thanks for reading!

Looking to discuss a project or view some of my work? Let’s connect!

carlosbcapel@gmail.com

//

Linkedin

//

Dribbble