Prism - Market Intelligence


Business problem

In April 2022, I was working at Sellics when it was acquired by the Ascential Group. The team merged with their existing advertising product called Perpetua, and the Sellics tool was discontinued.

Prior to the merger, the Perpetua business team had noticed that large clients were migrating to other tools that provided competitive intelligence using estimated sales of products on Amazon. As a result, part of the Sellics team was tasked with developing a solution to address this market need.

Market research

As the designer responsible for this new project, I started working with the product manager and the general manager to research the current tools in the market and map their value proposition, offers, strengths, and weaknesses. 

We identified three main tools:


Market research main takeaways:

  • To start competing with the main existing tools, we had to offer at least good data accuracy and flexibility to segment markets
  • All brands offered the market intelligence tool without a direct connection with their advertising module. The differential should be to plan the tool having in mind how it could be seamlessly used with our existing advertising solution. Our assumption was that the market intelligence data could leverage the ad strategy

User research

In parallel to the market research, we interviewed 15 people from different backgrounds, such as sellers, agencies, and aggregators who were using or have used similar tools. The goal was to understand their current workflow, needs, and pain points. At this step of the process, we wanted to collect as much information as possible so we had open conversations where the main goal was to discover:

  • Which market intelligence tools they were using or used in the past?
  • What did they like and didn't like about those tools?
  • Why do they need a market intelligence tool?
  • How is their current workflow when analyzing competitive information?
  • What do they do with this information?

Main discovered use cases:


To define which exact problems to be tackled, we decided to focus on agencies since they were already the biggest user base from Perpetua. Inside the agencies, our main actors were the data strategists and account managers.

User interview main takeaways:

  • Data accuracy should stay between a 10% to 15% error margin to be better than the current tools
  • Market share is the primary metric used to identify problems with their performance in a segment. For example, if the total sales of a brand shrunk and the market share remained stable, it means that the whole segment shrunk. This shows that it's not the lack of visibility caused by little investment in ads but that the segment has a problem or is in a low season
  • Agencies care more about competitor brands than specific products. This helps them to scale operations when analyzing data because they can aggregate it under the same money source
  • Data strategists analyze market intelligence by aggregating data weekly or monthly

User problem statement

Agencies had to rely only on their client's historical data to come up with a strategy to grow their revenue on Amazon. With many transformations and the marketplace becoming more aggressive over the years, players need market data to project more realistic scenarios and understand external factors that are impacting their performance.

The solution

Prism is a market intelligence tool that uses estimated sales data from products on Amazon to model consumer trends in the marketplace. It allows advertisers to define market segments and track their share so they see a more realistic scenario to find opportunities to grow, discover who they are directly competing against, and the seasonality factors that impact advertising performance. This tool allows agencies to have discussions with their clients using data as a reference point.

Main user flow


Main use cases

  • As an agency, I want to see my market share so if it’s going down I can quickly identify that I have a problem with the products in that market and correct it
  • As an agency, I want to see the overall trend of the markets to identify seasonalities, prices, ratings, and review changes
  • As an agency, I want to compare my client's brand trends to their competitors so I can quickly visualize who is attacking me or who is easier to attack. Once they identify the brands they go to the product level
  • As an agency, daily data is not important for me because I use weekly and monthly aggregations to plan long-term strategies. (This also helps to show more accurate estimates)


Based on the collected information, we prototyped the version of the ideal product with some of our main ideas. Instead of wireframes, I used the existing Perpetua design system in the prototype because it contained solid components. Nonetheless, I didn’t spend too much time refining screens and using the exact same patterns.



We went back to some of the agencies we interviewed and pitched them three versions to discover which ideas resonated more with them. We also presented the project to the tech team to review feasibility. Based on the feedback we kept the most valuable features, adjusted what was necessary, and prioritized how we could release the product in phases.



Based on the user feedback and technical feasibility we split the releases based on features:

MVP - Closed Beta (October 2022)

The first thing we wanted to validate was the quality of the data. As the first release for the closed beta, we launched Explorer, which gave access to users to compare their brands with the estimates and then analyze competitor brands.


Version 2 - Closed Beta

We allowed users to use the main filters to create markets and save the reporting.


Version 3 - Open Beta (February 2023)

Improved the filters to create markets and reporting by adding more types of data visualization.


Version 4 - General Availability (March 2023)

We allowed users to dig deeper on the product level so they could see how specific products were performing and which keywords they were ranking for. We combined this data with the advertising data of the user to show which keywords they were already bidding on and which were not. They could then bid more aggressively if they want to outrank a certain competitor.


Project challenges 

  • We had a lot of ideas and problems we wanted to work on. The main challenge was deciding which ones to solve and in what order. To solve them we were constantly evaluating the value we could deliver to users against the technical feasibility of implementing them
  • We worked with an existing component library and sometimes ran into constraints that required customization to solve specific Prism use cases. Good communication between the design team was essential to maintain the visual identity and ensure that all existing use cases were covered

Success metrics

  • Reduce churn of bigger clients
  • Validate the completion rate of the flow starting on Prism and finalizing adjusting ad settings on the Ad Manager
  • Validate the number of Markets created, filters used, and inputted queries on the Explorer
  • 100 clients by the end of the year

After release

After release, we could gather real usage data to validate how the tool was being used and talk to users.

User feedback

  • "It's quick and simple, a five-year-old kid could use it" - Retail Business Manager
  • "It's super flexible to model markets, I can get the information I want in 15 to 20 minutes" -  Media and Retail Senior Manager
  • Fast loading time (we started to make product decisions to not jeopardize the loading time)
  • The search term table showing what is already advertised was the AHA moment (This was the first connection with the Ad Manager)
  • Feature requests to connect more with the advertising features and improve operations in bulk