• Table.ai’s mission is to utilize AI to empower brand managers in managing their brands more efficiently.

• Through user interviews and research, we learned that brand managers selling on Amazon and Shopify currently need to access at least 4 different platforms in order to piece together the full picture of how their brand or product is doing, which takes up a good chunk of their day.

• Even with the help of external teams putting together reports, brand managers will still put together a homebrew excel spreadsheet to track progress and usually have different spreadsheets to track keyword performance, ad campaign performance and marketing spend, P&L, etc.

• Many brand managers oversee multiple brands. Due to the significant time they dedicate to data aggregation, they often find themselves spending the remaining time addressing urgent issues rather than strategizing on improving their brands and products.

• Through user interviews, shadowing, and prototype testing, I designed the site architecture and page UI/UX to allow brand managers an intuitive way to triage their brands at the start of the day, and be able to dive into the specifics to figure out what is the root cause of a brand or product doing poorly and how to remedy it.

• Table.ai assists brands in consolidating data from multiple platforms, enabling brand managers to efficiently triage and prioritize their tasks.

• Using machine learning and AI, Table.ai empowers brand managers with an understanding of the causal drivers for their brand or product performance.

Interviews

• Conducted 5 thinking aloud shadowing user research sessions paired with an interview segment to understand existing workflows and pain points.

• Brand managers had to access at least 4 different platforms to determine brand and product health.

• Some companies hired external teams to compile the data from these different platforms.

• Brand managers would create their own excel spreadsheets to manually log information they cared about, but found it difficult to maintain a comprehensive list, so sometimes missed key opportunities to grow the brand/product.

• Brand managers typically think about the metrics in 3 main categories: core health metrics (like revenue and market share), levers (metrics they can modify like ad spend, unit price, discounts), and other metrics they use to validate results or to triage the root cause of a problem.

• Some users manage multiple brands across 1-8 sales channels, with each channel having 20-40 campaigns, meaning they could be managing around 180 campaigns and monitoring the performance.

Competitive Analysis

• We also looked at existing platforms most brand managers currently use: Helium 10, Jungle Scout, Sellerboard, Data Hawk, Sellerize, etc.

• While these platforms provided brand managers with detailed information about various aspects of their business, the holistic picture is scattered across several different platforms. Inundated with data, brand managers found it hard to identify how to prioritize their time.

• The data within the platforms were also presented in a way that was difficult to scan through. The default view was also not aligned with brand manager’s mental models.

Information Overload

As the product designer, one of the main challenges was identifying what information to display? Since we’re aggregating data from multiple platforms, it was crucial not to overwhelm the user with all the numbers while providing enough information for brand managers to be able to make decisions.

Through user interviews, I learned that there were key indicators of a brand or product’s health: Revenue and Market share. I also learned that it was important to distinguish the “levers” from the KPI. Levers are things that brand managers have control over in order to try to improve their brand or product health.

To each their own

After more user interviews, I realized that each brand manager has their own workflow and stats that they look at in order to gauge their brand health. A key decision point had to be made as to whether we set the precedent for how to view the information or allow brand managers to customize their view. There were so many ways you could slice the information: Key drivers/Key Indicators, Own products vs Competitors, Levers vs Other metrics, Ads vs Keywords vs Competitors, etc.

Simple and Clean

When you think of data analytics, you think of a dashboard filled with tables, charts, and a lot of numbers. We wanted to revolutionize the way data was presented. Our main distinguishing factor from other data analytics platforms at the time was our AI insights. Table.ai needed to look clean and minimalistic. We achieved this by only surfacing the top 3-4 insights that were important and/or time sensitive.

Trust in AI

At the same time, we still needed to give the users a way to establish trust in our AI insights. One way to do that was to hide details in nested content. That way, we could allow the users to dig into the actual data to validate our insights.

Initial design

This design was pivoted from a previous implementation of Table.ai’s analytics for a power grid. The spider chart is meant to show users what has the biggest influence over the product’s revenue. However, this chart took up a lot of space and could be consolidated with the Key Drivers section. An exploration into an alternative way to convey impact is shown in the purple bar with “20%” next to the key indicators. The explanation text also took up a lot of room and was later moved to a hover explanation text

Brand managers quickly triage their brand’s health by looking at Revenue and Market share. Different types of charts were explored. We also looked at different ways of slicing the data as you delve deeper. At the brand level, you’d see your own products’ performance vs competitors’ performance. A layer deeper, you would be able to see the health of an individual product, levers, and other metrics.

Metrics could be toggled on and off in the chart by clicking on the metric. Metrics were displayed in order of highest impact to lowest impact.

I also explored bucketing Metrics so you could choose to only see Keywords if you wanted.

A huge request was to increase the size of the chart, so the Core Health Metrics were moved down and the Levers and Metrics were moved into different tabs you could toggle between

I also explored bucketing metrics so that you could choose to only see Keywords if you wanted. Different information was also added or removed based on user interview and user testing sessions.

This was also the first exploration into using waterfall charts instead of line or bar charts. We learned that many brand managers are accustomed to using this type of chart to view their data so we adopted it for easier transitioning. Because you could see how much impact a metric had within the chart, the impact % was removed and exchanged for a chart toggle as clicking on the metric bar to change the chart display was not easily discoverable for users previously.