Replacing an expensive and complex third party solution

The retail brokerage business is one where every second counts – stocks can move in milliseconds. Customers need to be kept aware of changing market conditions at all times.

Customers use the platform to execute individual trades, often using market data feeds that provide a traditional on-screen ticker of stock prices. Users typically set up alerts that inform them about specific stock price movements. 

Until 2022, tastytrade contracted with a third-party alert service that enabled traders to establish custom alerts. For example, a user might receive an alert if the Apple stock price went above $500 or if Microsoft fell by more than 5%.  

The problem with this approach was the lack of fine-grain alert scenarios that were more complex than merely stock prices moving up and down. tastytrade wanted to build an alert service that allowed for more complex alerts to be easily established. 

The second challenge with the third-party tool was that customizing the current alert creation logic was too complicated, making it hard to update information received in response to market changes. As a subscription service, tastytrade had pushed it to its limits. 

Seeing the opportunity to create a more robust alerts tool, tastytrade sought to build a service that allowed greater complexity for alert queries. For example, a customer might want to know how one stock price changed in relation to a wider sector, or how long a stock remained at a certain price point before issuing an alert.

In 2022, tastytrade approached Equal Experts with an idea: was it possible to create an in-house alternative to the current alerts platform? The new platform would be entirely owned by tastytrade and needed to be easily maintained and adapted over time.

Delivering low latency in a cloud-based environment

The new platform didn’t just need to be flexible and scalable, it also had to run as a managed service in a cloud environment. This is a significant challenge for a trading alerts service, where vast quantities of data and associated queries must be processed and results delivered in a matter of milliseconds. 

Equal Experts assigned a team of engineers that set to work creating a tool that could process more complex queries.  The initial version of the tool went live over a two-month period and was performant with single conditional alerting as its proof of concept. The team then spent three additional months refining the learning algorithm and building metrics that would measure the response time for queries ensuring that the low latency was delivered.