CRU is a business intelligence company, and they have an online platform which subscribers use to access articles and data about commodities based on research and expert knowledge from analysts.
Outcomes
45% increase in platform use
New product features to help increase platform subscriptions
Improved internal workflows to cut production costs
GenAI pilot Product Strategy and Design
How can we bring subscribers back to the platform?
- Many users were relying on information sent them in email communications
- Editorial teams needed improvements to their workflows
- Analysts needed a better search experience for finding key intelligence
User Journeys and IA
Users had a hard time searching and navigating the platform to find the information that they needed (articles or the latest relevant data) and there was a lot of extra ‘noise’ in the form of updates and notifications.
Redesigning the Content Model
The platform was suffering from Information Architecture that was a manifestation of the way customers were subscribing to services – not how they expected to find and use content and data. Content was being siloed in a strict hierarchy in the site, with a lot of cross publishing and duplication.
Therefore we proposed to ‘flip’ the model upside down to allow more flexible ways to find and organise content.
Prototyping user journeys
I worked with stakeholders and users to validate a wireframe prototype to communicate what the experience of navigating the site would be like.
Business Goals and Metrics
To formulate a good analytics strategy, I set out a brief guide and rationale for what was needed from a UX perspective. This included revisiting the major business goals and thinking about:
- People – who will manage, interpret and most importantly take actions with data
- Process – the process of report and alert distribution
- Tools – choosing the right analytics tools and integrating the data
- Measures – information overload is the challenge – finding the right measures to use in dashboards, alerts and for optimisation is needed to get the most from analytics
I then took the HEART analytics framework from Google Ventures and linked each aspect with the business goals, what signals to use and the metrics that we would then be able to measure desired business outcomes.
Content Strategy
I engaged with our client develop a formalised content strategy. This can be a daunting task, but by explaining what content strategy means and then breaking it down into a stepped process I helped the content team to start to build out a clear strategy for what content to publish, how, who it was for and what need each type of content was serving.
My main learning from working on this is that it takes time, but you have to start somewhere.
Usability Testing
We were also looking to bake in usability testing into our sprints, so that we would get early feedback from users, and get some early seeds growing for platform champions and changing the client mindset to a user-centred one. We also wanted to have usability testing and research as part of adding new features and services to the platform as the product matures.
UI Design and Agile build
Following all this work I then worked with a visual designer to create a new style and patterns for the platform. I wanted the Development team to feel fully involved, so they joined us through the design process, and we produced Adobe XD handoffs containing all the assets and specifications they would need to implement the new UI.
After the initial design system was signed off, I took on the role of UI design for any iterations and further product development.
Dashboard and data cards design
New product features that I am proud of are the design of commodity dashboards and data cards.
I worked closely with the client to research ideas, we then also ran some user research sessions to uncover what information clients need regularly at a glance. That allowed me to design some scalable, interactive dashboards and data cards that would become a big selling point for new subscriptions.
Introducing GenAI into the platform
CRU knew from customer feedback that a key priority for many users was the ability to get summaries of analyst research, as customers often have limited time to get information to support their needs.
So they wanted to launch a GenAI pilot where users could ask questions around commodities research. The model would be trained on a limited dataset, so we could gauge the quality of responses returned.
A second feature was to supply summary and translation of any article on the platform by using a simple added feature in the article UI. Again we would use this to benchmark the quality of translation returned by the model.
Project Requirements Document
To support the initiative I too ownership of the PRD from the start, to make sure that we were clear about the customer needs, aims and scope of the project.
I cannot show any UI for this at the moment. The pilot is still in testing.