About the Client
The client is a US-based technology startup that provides aggregation and consolidation of widespread, fundamental information to help assess, analyze, and track investments.
As the client was in the final stages of securing their IPO on NASDAQ Capital Markets, they were seeking a strong partner to overcome their existing challenges, as well as help them achieve their long-term goals.
Business Challenge
The client’s objective was to expand on their current offering, by building two services, one tailored to B2C interactions, and the other for B2B.
Business to Customer (B2C)
The client planned to launch an analytical terminal that would aggregate a large volume of data from exchanges. The data would be made accessible for investors who research and trade US equities, allowing them to efficiently analyze data, create watchlists, and engage with other users. The terminal would be available on a subscription basis.
Business to Business (B2B)
For the B2B component, they intended to create a self-service advertising platform for companies that are publicly listed. This platform would allow the end-users to autonomously manage and execute their advertising strategies.
Companies would have the ability to disseminate their corporate news and press releases to specific user groups, based on their behaviors. The software would turn this content into personalized, actionable material, tailored to the individual target audiences, attracting them to the companies’ assets.
This service would also operate on a pay-per-use model.
The client’s current website already offered corporate news, price charts, real-time market data, etc.
However, they faced several significant challenges and project risks:
- Time to market & quality of the software
- Market data
- Quality of machine learning technologies.
Time to Market & Quality of the Software
Initially, the client tried to implement the product in-house, but encountered challenges related to the narrow specialization of the project. It was difficult to hire skilled software engineers with experience in capital markets. This was jeopardizing the client’s ability to adhere to the timeline of their roadmap. Besides, the complexity of the domain, along with the demanding requirements for systems processing real-time market data, also highlighted the limitations of the existing prototype.
Market Data
While the existing website was already operational, the firm faced issues with its current market data vendor. In addition to the periodic quality incidents, the client was having difficulties integrating the provider’s widgets into the existing project’s ecosystem.
Quality of Machine Learning
To create tailored recommendations by analyzing user behavior, it’s essential to first gather extensive datasets, which are crucial for training machine learning models.
Existing open-source recommendation system implementations did not align well with the client’s needs. Inadequate targeting would discourage advertisers from using the platform, and potentially defeat the business model.
“This part of the project was the most difficult I had faced in a long time.”
— Chairman & President, the client
This led the client to explore whether other market data vendors and software powerhouses could solve these issues and deliver quality web and mobile apps, fully wired with robust market data.
Solution
B2C
Devexperts team swiftly assembled a solution to overcome the client’s challenges and upgrade their front-end user experience.
The solution is a white-label product based on DXfina, our feature-rich, time-tested analytical terminal for retail investors in Turkey. DXfina consists of web and mobile apps and has most of the front-end functionality, required for the client’s objectives.
By licensing DXfina, Devexperts effectively narrowed down the project scope to a mere localization of the interfaces.
B2B
The portal for advertisers was created from scratch.
Machine Learning
Devexperts upgraded client’s portal front-end so that virtually every action of users was instrumented and registered into Devexa’s data lake. This collected data is used to train and constantly improve machine learning models.
“There were a couple of challenges: short life-cycle of items, that users interact with, and the specifics of the data. The data set contained encoded user activity data rather than interaction with recommended items. Because of that, the client couldn’t use an open-source solution. But despite the tight timeframe for the project, our data science team managed to develop a customized solution that works.”
— Ivan Kunyankin, Senior Data Scientist at Devexperts
Fig. Workflow of the user data collection and recommendation process
Market Data
This aspect was a straight-forward deliverable for Devexperts, as dxFeed has coverage of the necessary instruments:
- NYSE CTA (Tape A, Tape B) + Nasdaq UTP (Type C);
- Nasdaq Basic L1;
- Nasdaq Total View L1 + Market Depth;
- OTC Markets L1, L2;
- OPRA Options Composite;
- OPRA Options;
- Composite + Regionals;
- CBOT Futures L1;
- NYMEX Futures L1;
- COMEX Futures L1;
- CME Futures L1;
- BUNDLE: CME Group L1;
- CBOT Futures L1 + Market Depth;
- NYMEX Futures L1 + Market Depth;
- COMEX Futures L1 + Market Depth;
- CME Futures L1 + Market Depth;
- BUNDLE: CME Group L1 + Market Depth
The client was connected to the market data endpoints under the Vendor Of Record (VoR) model.
Results
Within eight weeks, the Devexperts team implemented and customized the analytical terminal in close cooperation with the client. The client received the analytical terminal, branded to match their style, with integrated flow of US market data from dxFeed. The terminal was available through a web interface and as a mobile app.
“Overall, it looked and felt like a pit stop teamwork. Different teams at the firm, Devexperts, and dxFeed stitched together licensable blocks, custom software, data storages, machine learning microservices, fueled it with market data and launched it within the weeks.”
— Chairman & President, the client
“This was far from a linear project; it has been an intricate bundle of deliverables designed to align with customer vision, bolster their capabilities, and push the boundaries of what’s possible in such a short timeframe. A sincere thank you to everyone involved.”
— Ivan Razuvaev, Project Director, Devexperts
This project encompassed:
- Licensing of DXfina, our advanced analytical terminal.
- The customization of DXfina (initially offered for the retail investors in Türkiye) to work with the US stock market.
- Our dedicated Devexperts software engineering team executed a work-for-hire arrangement, implementing a user-friendly web portal.
- Devexa, our state-of-the-art SaaS for platform intelligence and business messaging.
- Customization of Devexa, turning it into a comprehensive telemetry and behavior data collection hub.
- The creation of a personalized market news engine by our Devexperts’ ML team
- Finally, incorporating market data from dxFeed, a Devexperts subsidiary