Agile for Business

Agile for Business

Business agility for Global FX distribution firm

A leading electronic foreign exchange aggregation and distribution firm had gained a large share of the global institutional client market.  They wanted to grow their market share but  the lengthy on-boarding time caused by the need to install the Java desktop application was creating a barrier to entry. 

A natural next step was to create a web version that is immediately available.  Many more potential customers could be targeted if an HTML5 version were available on the web,

However, replicating all the features and workflows offered by the existing Java UI would be expensive and time consuming. There was a suspicion that whilst porting over everything would (theoretically) satisfy all types of potential user, there was a subset of key features that would be more valuable to release first.

How we helped.

On the face of it, this was a classic agile project. A longish roadmap drawn from the functionality of the existing user interface, standard non-functional requirements relating to deploying any web application and a good set of stakeholders at the client who understood the customer needs. The natural thing was to prioritise the user stories in a backlog, proceed with iterations that would be released to a pilot group of customers to gain feedback, and then repeat.

Due to the nature of foreign exchange trading, there are many integration points for any system that executes trades. For legacy architectural reasons, it would take several months to include trade execution. Whilst the timescale was acceptable, the time gap presented an opportunity to usefully engage the potential customers and test the general business hypothesis.

Validated Learning

Our suggestion was to release an experimental version that showed the competitive FX market prices in real time. There would be no trading. Instead, a link against each product was provided for the customer to click if they liked the price they saw. This action was used in multiple ways, from registering an interest in the full trading version through to fast tracking the installation of the existing Java application.

Since the product prices had to be made available anyway it was agreed to go with this simple development first and see what could be learnt. Publishing indicative market prices without the ability to execute a trade is a benign way to market a trading service to financial users. For this reason and the technical simplicity it was possible to make the first release very quickly.

It soon became clear that potential customers were keen on the pricing. Feedback was generally positive. Interestingly, by utilising the web-app frameworks and components recommended on all Esparter projects, it was possible to analyse and measure specific user behaviours such as which products were most viewed by customers in particular locations.

The output from this analysis changed the priority and focus for the subsequent project. Instead of just adding a broad capability, specific forward pricing products were released to a small but significant customer segments at identified locations. A much greater value product emerged.