FactSet EMS
The Problem
Portware is a stand alone execution management system that was acquired by FactSet in 2015. The strategic plan was to expand FactSet’s offering to support the entire investment process from idea generation, screening, portfolio modeling, to trade execution and confirmation. Integrating two systems raises problems, just some of them are below:
Portware’s UI was built piecemeal at the direction of client requests. The result of this was a large offering that didn’t mirror user workflows, but was a hodgepodge of features.
Portware’s solution to their UI problems was to allow for customization of everything instead of following common workflows & use patterns.
Customization and settings were spread out across the application and referred to with non-standard terminology. For example, changing a toolbar was accessed in a section called “algorithms.”
Portware’s visual style was convoluted and didn’t follow any standard or design system.
Settings were hard to find, used unfamiliar terminology, had no description, and zero context.
Execution tickets had high cognitive load, little to no data validation, and didn’t adapt to where they were in the user’s workflow.
Color choices created low contrast data and CTAs. Opening up 100% color customization only compounded this problem.
Multiple blotters were created for different trade types and different steps in the execution life cycle. This created multiple unused blotters that took up space for traders.
The breakdown
We knew ahead of time who are user was: institutional traders. They are very busy and we found it difficult to get meaningful discovery research done. Our UX research team was able to use former traders as a proxy to get started along with FactSet’s previous experience with our trader user type.
I worked with the product team to simplify the amount of UI while keeping the proven UI elements in the new version. I continually made wires frames and prototypes to advocate for design solutions in the face of a constant stream of features users “might” use.
Priorities and scope creep were a constant problem on this project due to the high visibility within the company, sales/management interference, and the lack of easy user testing.
The Solution
After doing several rounds of story mapping and working to understand user workflows, I broke down the entire UI into four main sections covering ~90% of the current offerings for the MVP.
Blotters - where trades lives pre and post trade.
Market Depth - where traders can see the market.
Ticket - where traders enter trade details to execute.
Settings - a consolidated settings accessible from anywhere in the application.
As the blotters were the main interaction point, I mocked up and tested several different versions with different layouts and features. I also updated the configuration UI with drag & drop functionality, a new grouping UI, and methods for users to share their settings and customizations with one another.
Two blotters could mirror the workflow to show parent orders from an OMS or portfolio manager while the associated routes we be displayed in a separate window.
Blotters could be organized via tabs to show outstanding orders, positions, allocations, and other post-trade details.
Using a vertical name with the dual blotters combined standard filters or “views” for the user to adjust the given blotter quickly.
The Results
Early versions were reviewed internally and tested with users. Simplification and intra-app communication was received very well and prioritized in future iterations. Higher fidelity wires with new UIs and controls were praised by users. Additionally several other points were observed when we tested:
Limited customization and custom colors kept the experience focused on trading instead of configuration.
We kept the blotter & ticket UIs consistent with the marketplace for quick adoption. All tests had users jumping right into the deep end and able to trade without getting used to a new and different UI.
Users had more requests around new workflows now that they have access to FactSet’s data and platform. Continuing user research is required to narrow and understand these future requests. For example, visualizing their trades and executing on a chart are solutions we are currently researching.
The first UI for the two blotters used a dark theme, left space for future features, and reduced bloat on the desktop.
An initial new multi-security ticket allowed users to route multiple orders to one broker with one algorithmic execution strategy.
In an effort to make certain pieces of data more consumable, a side panel with contextual data was added. Execution history, an audit trail, and market data benchmarks are all being added to this container.
We created a drag & drop UI to allow a user to save trade shortcuts quickly route future trades via one click. Their order and location is handed via simple drag & drop.
EMS is currently be sold to FactSet clients and we are continuing to build out more advanced workflow support features every two weeks. Additional features are detailed on other sections within my portfolio.