Interface Optimizer

Challenge

The system's routines are very broad, as they aim to serve different business sizes, becoming complex and tardy in some contexts, which generates low user satisfaction.

This complexity sometimes ends up affecting users' cognitive load and generating more friction in use, burdening the completion of tasks that really matter, in addition to impacting the TCO (Total Cost of Ownership) by affecting the user experience with delays in loading the interface.

All of this also ends up affecting system administrators, who manage ERP in companies and are compelled to modify these interfaces in a sustainable way to maintain the correct functioning of the system as a whole.

Approach

Together with two other researchers, we started the project with a well-grounded research phase, seeking to raise doubts (CSD Matrix) to be explored with key users and extract all expertise from the core team, which consisted of six product specialists, most of whom had nearly three decades of experience in the field.

We then conducted a total of 12 in-depth interviews to understand the needs and challenges that users and system administrators had in their day-to-day lives when carrying out this type of operation, from the emergence of this demand to the implementation of the changes, and we identified that it was a slow and risky process to carry out.

Thereafter, we synthesized all the research with the core team using the following sequence of dynamics: Affinity grouping, Insight generation and HMW. In this way, we found that the most aggravating need was to analyze the impact of the change in metadata and a non-standardized process. The dots began to be connected.

Then, we carried out the Brainwriting dynamic to generate ideas based on the synthesized material, which served as a basis for prototyping a solution that would assist users throughout the customization process, creating a mental model with clear steps and information that would contribute to decision-making.

The key point here was the inclusion of an intelligent functionality within these steps, which would perform a check for unused fields in the customer base. This needed to be very clear.

With that, we built a low-end prototype (Miro) and validated the concept with another group of recruited users:

From this moment on, I took over the management of the project, starting with the analysis and prioritization of the lessons learned in the first validation. This made it possible to iterate the prototype with functionalities and flows not previously mapped, such as:

• Inclusion of an editing workflow – as I emphasized the need for freedom and control for users;
• Elimination of redundancy in steps when creating customization – to simplify and make the process more intuitive;
• Need for additional information about routines – to facilitate decision-making;
• Application of improvements to texts and labels in general – to provide more clarity on the necessary steps and actions.

With the approval of stakeholders, I consolidated this stage by iterating the prototype into a medium-fidelity version (Axure RP) and invited around 8 designers from other product areas to carry out a Design Critique dynamic, in order to bring a perspective without the bias of context for evaluating technical usability issues.

After this brief refinement, I conducted usability tests with 13 key product users, now emphasizing on the tool's usability. This new validation round provided more insights for improving writing, as some actions were still not clear enough on first access. Additionally, I also added new information in the data visualization to help users analyze and customize their routines.

During the usability test, I also applied A/B tests to understand user preferences regarding two structures designed to display information in the analysis stage of the customization routine.

Although the designed structure worked, users ended up preferring the second option, which would provide an even better understanding of the information due to a preconceived mental model of the system itself.

To complete the usability cycle, I added comments to the prototype containing all the highlights from the transcriptions made during the usability tests, and later facilitated a dynamic for analysis and prioritization with the main team on each of these points.

With this, we were able to obtain some more valuable insights to improve the tool's operation and make it more intuitive and functional, such as:

• Removed the search for customizations when they have not yet been created – to avoid confusion and make the tool more intuitive;
• Changed the location of the customization deletion – to make it easier to access the action;
• Added new information to the analysis table – to help assess the importance of the element;
• Improved the title of the last step of the process and in the completion message – to make the final customization step more cohesive.

Another 7 improvement opportunities were also identified that did not fit into this tool's MVP, but were documented for technical feasibility study by the development team.

---

Delivery

Finally, I consolidated all the learnings gathered during the approach into the project, crafting the pixel perfect of the interface. The end result was a wizard that divides the customization task into 3 guided and intuitive steps, with a dashboard that allows you to manage each of them.

The tool's workflow begins on a page for managing the system's routine customizations, where the user can later manipulate each of them with actions like editing, deleting, or even disabling it:

Following the workflow, the user can create a new customization or edit an existing one, following the same step by step with the same mental model:

• Description and routine: the routine to be customized is selected and its personal identification is defined to identify it in management;
• Analysis and removal of fields: allows you to optionally perform intelligent analysis of the database, and then define the customization;
• Template review: Finally, the user does a general review of their definitions to make sure they are correct.

All of this resulted in a significant reduction in the time taken to customize interfaces due to a simplified flow, also promoting greater confidence through the availability of intelligent technologies that enable prescriptive analysis and assist users in decision-making.

Key Results

In the end of the project, I analyzed the data obtained throughout the usability cycle, noting the following metrics:

• Success Rate: resulted in a total of 94.62% in task effectiveness.
• NPS: Reached 76.92%, with 11 of 13 users being promoters and an average rating of 9.4.

I also defined metrics using the HEART framework, but I will not address them here to maintain confidentiality regarding strategic aspects of the product.

It was found that in some cases there can be up to more than 40% reduction in unused fields with the use of the tool, which also affects page loading time, resulting in a better experience and a more user-friendly interface.

Customer Testimonials

See some of the feedback received during prototype validation:

In total, around 90 people were involved in the Design Thinking approach, including companies like Mitsubishi Electric, John Deere and AMG Motors.

Are you interested in my job? Contact me here.