Engagement and Retention Tool for a Funded Trader Firm

About the Client

The client is a young and fast-growing funded trading firm in Europe.

Business Challenge

The client faced a number of challenges related to high customer churn rates and satisfaction. Their marketing resources were solely focused on acquiring customers, without adequate retention efforts. The client was seeking to

enhance user engagement
reduce the percentage of users switching to competitors
scale their operations without proportional cost increases.

The client’s hypothesis was that generative AI could make the platform stickier and that AI integration could be a recipe for increased user engagement, tailored entertainment, and ultimately, business growth.

Finally, as the client was running on DXtrade, a white-label of Devexperts’ CFD platform, the ideal solution would require native integration with the platform’s web terminal.

Solution

Devexperts proposed to enable Devexa inside DXtrade and use it to tackle these challenges.

About Devexa

Devexa is an omnichannel communication platform tailored for brokers and prop trading firms. It combines live chat, video calls, screen sharing, broadcasting, and chatbots into a single interface. This integration allows real-time engagement and support for traders, while also enabling firms to track users’ interactions with a trading platform, and gain insights to enhance the customer experience.

The platform features AI-driven content personalization, based on user’s interests and smart push notifications. These features ensure traders receive customized content and timely alerts, fostering greater engagement and satisfaction.

With Devexa, traders receive relevant market information, business news, and alerts. Businesses enjoy increased engagement, satisfaction, and trading volumes.

Devexa is fully integrated with the DXtrade platform and can be activated upon clients’ request.

Devexa widget inside the DXtrade platform

Project Charter

After a series of consultations, the client and Devexperts defined the key hypotheses:

A new stock market news feed with personalization:
In order to improve the user activity and engagement metrics, the new stock news feed is to be plugged into Devexa. Devexa can fetch the news upon traders’ request and the embedded AI recommendation engine inside Devexa would reshuffle and filter the news based on the traders’ account.

Stock news push notifications:

To bring engagement to a whole new level, Devexa would push relevant news to online users. As the news and their relevance to the given user are calculated, real-time notifications can act like a squawk box, alerting traders to act. The client and Devexperts can control and gauge the aggressiveness of the notifications, by tuning Devexa’s recommendation algos.

Dutch language support:

Devexa is to enhance its list of supported languages by improving support for Dutch.

This upgrade would enhance the user experience (the client’s users are predominantly Dutch-speaking), increase user engagement, and improve retention.

Implementation

Devexperts enabled Devexa inside DXtrade for this client.

An example of a personalized news feed compiled by Devexa and streamed into DXtrade

Integration of StockNewsAPI.com

Devexperts integrated StockNewsAPI.com into Devexa, extending the list of supported news feeds.

Devexa Backoffice: Operators have the ability to select multiple news feeds from a given list of news sources

Personalization

Devexa already had AI in service: it uses neural networks and machine learning algorithms for natural language understanding, intent classification, retrieval-augmented generation, and news feed ranking.

In the scope of this project, Devexperts added account trading history and users’ portfolios into the data used by the recommender system.

User’s traded symbols will be utilized in the recommendation engine

It turned out there are differences in the usage patterns between the clients of prop firms and the clients of brokerage firms. It was quite insightful to analyze these differences in the datasets on the way they use the trading web app, the AI assistant widget.
— Ivan Kunyankin, Senior Data Scientist, Devexperts

Devexa Backoffice: An example of ML classification of a user (on the right)

Devexa Widget: Personalized news feed with quick trade buttons, light mode vs. dark mode

Push Notifications

Devexperts developed a custom news feed module within the platform, ensuring real-time delivery of relevant news updates to traders. They also implemented a push notification system to alert traders of important news events, enhancing user engagement and providing valuable market insights.

Devexa Widget: Users receive notifications when new items of interest are published

Push Notifications Tuning

The first rollouts of the feature received mixed feedback. While Devexa ensured that traders received only relevant and timely news notifications, some users complained that the news buzz was quite noisy and annoying. The client and Devexperts team tried several versions of the recommendation engine, considering intermittence, and various heights of the relevance threshold. Eventually, the feature was refined to strike a balance between providing timely, relevant updates and reducing information overload.

Widget designs seek balance: Engagement vs. retention

Widget Redesign

The testing has shown that the push notifications – especially when their frequency is intense – break the default conversational UI of Devexa: the news would interfere with the chat design. That triggered a rethinking and prompted a comprehensive redesign of the Devexa widget to accommodate for different user scenarios: live chat, AI assistant, company announcements, and market news.

The new UI features separate tabs, for Support and News Feed, providing a more intuitive and streamlined interface for traders.

Dutch Language Support

Devexa employs proprietary in-house neural networks to process natural language for prioritized languages and fallbacks to cloud AI for less commonly used languages. In the scope of this project, Devexperts has enhanced support for Dutch, the main language of the client’s user base.

Devexa Backoffice: Operators can edit responses used in Dutch conversations

Devexa Widget: An example of a user dialog in Dutch

Results

The client was provided with a package of DXtrade and Devexa. Devexa has been enhanced to specifically address engagement issues, in addition to its existing communication features.

Following the delivery, the client analyzed the user’s feedback and suggested further refinement. After a few iterations, Devexa precisely met the client’s expectations.

The client achieved enhancements in their user engagement.

This was the first launch of Devexa with a funded trading firm. The fact that most of the features requested could be delivered in weeks, suggests that our positioning, which initially targeted brokers, works for the funded trader companies, too. We’ll continue to strive for excellence as we further refine and enhance Devexa.
— Togay Aytemiz, Product Manager, Devexperts.

We have enhanced Devexa’s communication capabilities and further tailored it to trading domain scenarios. We have expanded our personalization engine, creating possibilities to personalize the user experience beyond the news feed. Now, all our clients, both prop trading firms and traditional brokers, can enjoy this ML upgrade.”
— Ivan Kunyankin, Senior Data Scientist, Devexperts.

We are always looking for new ways to increase LTV and improve user engagement. With DXtrade being end-user-facing, we wanted the DX platform to be more stickier.

And with the GenAI hype, we were keen to integrate some AI features and let our users have fun with it, if not get a real value out of it. The Devexa team worked swiftly and we enjoyed their friendly, responsive and agile approach. I quite liked your responsiveness and willingness to accommodate our feedback and suggestions for further improvement.
— Director, the client