Workbook exploration
Designed the entry point, configuration, and result for removing duplicate rows in a workbook table—turning Sigma’s most-requested support issue into a clear, deterministic workflow.
The number-one support request
Removing duplicate rows was the most-requested workbook feature from support engineers, affecting more than thirty-two customer organizations, including some of Sigma’s largest accounts.
The only workaround required users to understand grouping, drag columns manually, then add sorting and filtering steps. A task that felt like one action in Excel became thirteen clicks across several menus—and became impractical for tables with hundreds of columns.
Why the Excel pattern did not translate
We could not simply copy Excel’s two-click feature. Spreadsheets operate on individual cells, while Sigma operates on rows backed by cloud data that may continue to update.
The result therefore had to be deterministic. Users needed to choose which columns defined a duplicate and which row should be kept, with the sorting logic visible enough for governance and auditing.
Finding the expected entry point
I interviewed partners across product, support, engineering, and design, then tested six potential entry points with beginner and intermediate users. People consistently looked for the action in the column menu—even though removing duplicates was technically an element-level operation.
I followed that mental model. The action lived where users expected to find it, while the product handled the mismatch between their language and Sigma’s underlying architecture.
Making configuration explicit but lightweight
The configuration modal auto-selected comparison columns from the user’s context and made the keep rule explicit: preserve the first or last row according to a visible sort order.
I worked with a content designer on concise instructional copy and used smart defaults instead of asking users to interpret a vague “default” checkbox. The workflow remained quick for common cases without hiding the decisions that changed the result.
Hiding implementation complexity, not user control
Sigma already had grouping logic that could power the feature, but exposing that machinery was the source of the original confusion. I used the existing capability behind the scenes and kept grouping out of the user-facing flow.
The system preserved the original data by creating a child table, performed the intermediate transformation invisibly, and presented a clean result table as the object users continued working with.
A familiar action became genuinely accessible
The new workflow reduced the number of clicks by 92 percent and cut completion time by 84 percent. What previously took more than a minute and required specialized grouping knowledge could now be completed in seconds.
The project reinforced that an enterprise workflow can be both simple and rigorous. The interface should remove product-specific complexity while keeping the choices that make a data operation trustworthy and repeatable.