Empower everyone to explore data

A self-service business intelligence and analytics solution built for your cloud data warehouse.

Sigma Computing

Time: June 2016 - May 2020  |  Role: Head of Design

Sigma is the only self-service BI and data analytics tool that is made specifically for business users. Build dashboards, charts, tables, and reports in a familiar spreadsheet interface - all without any programming involved.  

For the first 2 years of this product, I was the sole designer. I designed the product from the ground up – everything from the initial research, UX/UI, user testing, establishing the design system / GUI of the product, all the way to working closely with engineers to getting everything implemented. 

The Problem 

A New Era for data

Until relatively recently, data was structured and small in size allowing Excel to be one of the primary tools for analysis. However, with the advancement of technology and the internet, large amounts of data are accumulating every day. Every click, purchase, and customer interaction generates data. This is often semi-structured, or completely unstructured. All this "big data" is a potential treasure trove for companies to gain insights and create real-world competitive advantages, but the process to attain this is no easy task. 

The Constraints

1 ) Limited to Data Teams / IT

Today data scientists are the professional experts who know how to tease actionable insights out of gigabytes of data. They spend a lot of time in the process of collecting, cleaning, and munging data, because data is never clean. This process requires persistence, statistics, and software engineering skills.

2) Requires SQL knowledge

Structured Query Language (SQL) is the standard programming language used to communicate with a database. It is primarily used to find and pull information from large databases. SQL has become a very important tool in a data scientist’s toolbox since it is critical in accessing and modifying data. 

3) Waiting and reliant on others for answers

With data informing just about every business decision, business users require the help of data teams / IT to help them make data-driven decisions. However, most companies require users to submit request tickets to the data team, which could take anywhere from 1 week to 3 months for a response. 

Frustration goes hand-in-hand with data teams as they are professionals who are qualified to create complex models and write scripts are instead being tasked to create basic reports.

4) No easy tool for business users

  • Requires SQL knowledge or a proprietary language

  • Steep learning curve

  • Not for business users

  • Not flexible

  • Only does certain things well (not end-to-end)

  • Requires multiple products

  • Very expensive

The Hypothesis

A Self Service Tool

This necessitates a self-service model that empowers the domain experts, aka business users, to be able to explore and answer their own data questions. 

The Process

How it all started

In 2016, I was recruited by Sutter Hill Ventures to join any one of their portfolio companies. I decided to join Sigma as the 8th employee because it was in the data space. Six months after joining, everything got scratched and I became the only designer. I started a journey with the co-founders and engineers to design a product from scratch.

Research Interviews

First, we needed to understand the problem and what users were currently using today. 

We didn't have UX researchers or product managers, so I started by interviewing users in our own network from small, medium, and large sized companies to understand how users were accessing data, what they were doing with it, and what the final output was. 

Research Findings

After meeting with a number of users from data teams, analysts, to those who worked with data, we discovered the following things:​​

  • There are 2 sides: the person making data available for others and the person consuming the data

  • Business users are used to a tabular spreadsheet UI (like Excel, Google Sheets)

  • Business users (non-technical) wouldn’t want to view dashboards in Excel, they wanted something simple and accessible on the web

  • They didn’t want to download anything, they wanted it accessible via web/cloud

  • The #1 most commonly used tool amongst all business users was Excel (spreadsheet)

The Ultimate Goal: Empower the business user

We decided to place our bet on making a tool that empowers the business user without having to write a line of SQL. At this time, we didn't know what the product would look like or what it would entail. 

All we knew were:

1) There's a huge market (BI & Analytics market is expected to reach $29 billion by 2022)

2) More and more companies were moving to the cloud

3) There's a big problem no one has solved yet – no tool for business users to clean & prep data for analysis.

My Goal: Simplifying the complex

How do I create an interface to something users usually create writing SQL?

How do I address what users want?

BI Architecture

I researched everything from the big data architecture, data pipelines, BI and data analytics workflows, and everything in between.

Competitive analysis

I identified all the major competitors and researched their products to analyze their user journeys, task flows, target audience, strengths/weaknesses, market positioning, and to identify any gaps in the product.  

Market Analysis

I researched the BI and data analytics market and there was a movement towards finding cloud-based solutions to their data warehouse. 

Brainstorming & Sketches

Working in a startup allowed for quick brainstorming and iterations via whiteboarding. This was the quickest way to gather high-level alignment with the cofounders and any stakeholders before moving forward to the next phase of the design process.

Initial drawings to figure out how to insert visualizations into the Sigma spreadsheet. I looked at other competitors and came up with different options of how we could add them in Sigma. 

Initial drawings for figuring out what the workflow would look like to Join different tables together using primary and foreign keys.

One of the earlier versions of the Sigma spreadsheet.

Final Designs

Sigma = the self-service tool for business users

  • A familiar spreadsheet UI that empowers everyone to explore data

  • Allows complex queries without writing a single line of code

  • Get answers to your data question via an explorable visual interface

  • Analyze your full data set, not just a slice

  • Connects directly to the cloud data warehouse (where data is stored)

  • Live connection, and data is always up-to-date and secure

Step 1: Create worksheet

Create a worksheet and add your data source, which is directly connected to your data warehouse (live data). 

Step 2: Group data for analysis

First, look for any columns to group this dataset by in the right-hand panel (in the "level" section). This will cause the entire dataset to update or pivot accordingly. Continue to select other columns to group and mold this dataset into something insightful.

Step 3: Create visualizations

Once we're happy with the insights we've gotten, we can now create visualizations to summarize the complex information into an easily digestible visual format.

Step 4: Create a dashboard

Once we've created a number of data visualizations/charts, we can now add them to a dashboard to allow users to see at-a-glance the insights and performance of an organization or dataset.

Other features: SQL Editor 

Sigma also has a SQL editor for data teams or engineers to quickly be able to edit and execute SQL statements if needed.

Improvements

In order to keep improving the product, we needed to further understand our user needs, BI goals, and any pain points. 

Personas

We didn't have a UX researcher so I lead the research initiative to better understand our users. I came up with the research structure from scheduling participants, sending surveys, creating questionnaires, aggregation of raw data into insights, to finalizing everything down to the final personas.

 

We interviewed over 30+ users at different organizations to understand user workflows, goals, pain points, etc. Every company had different user types, but we found the primary four users at almost every company.

User Testing

We would continually meet with users to test features, gather feedback, and provide any guidance whenever needed.

Demos and workshops

I would join the sales engineers who would demo the product to new users to understand what questions they would ask and see if we could somehow improve the UX/UI in the product. 

Conclusion

  • Sigma is the only no-code business intelligence and analytics solution built for your cloud data warehouse

  • Intuitive, spreadsheet UI gives teams the familiarity of Excel without the limitations

  • Empowers non-SQL, domain experts to analyze their own data

  • Powered a 90% reduction in engineering time required to report insights to users

  • Main adopters/biggest fans are the data scientists/data teams at organizations wanting to empower their business users