Carma (2015-2016)
Carma is a global provider of media intelligence solutions. As a result of recent acquisitions and expansion into European and North American markets they required an entirely new platform for their media monitoring and analysis services.
While our initial design sprint had focused on the customer facing product, further discovery revealed the extent of the work required upstream from this.
In addition to the client facing product, completely new internal workflow systems and interfaces would be required. Several recently acquired companies with differing tools, workflows, and methodologies needed to be transitioned to a common analysis platform. The monitoring operation, with production facilities spanning several countries across Europe, the Middle East, and Asia needed a seamless, uninterrupted transition to the new platform. Automation would have to be developed for custom workflows which had evolved over years of client familiarity.
Global Audience
The main product this was replacing had been primarily used in the Middle East and North African market. The replacement was to serve a worldwide customer base. Aligning customer expectations and familiarity was fraught with difficulty. Services that customers in one region had come to expect would not be possible in another. With the goal to avoid complex customisation for different regions and customers, we had to work with account management and client relations teams in multiple regions to determine how this could be best accomplished on a case by case basis.
A Daily Snapshot
Carma and it’s companies provide both media monitoring and analysis services. It was important to understand the primary motivation a customer would have for visiting the site, and to see how that might break down by service.
- Monitoring
- What's happening today?
- Analysis
- What does it mean?
Eventually the product would serve as a single location for both, but the initial focus was monitoring. These two questions guided the designs, with a dashboard style overview page focused on sumarizing the current day serving to answer the former, and a moduler design allowing expansion into the latter.
Scannability
The article list is the primary interface for the customer to check all their new articles, search for old items, and compile reports for export. The amount of content a customer receives daily can vary significantly though, depending on the nature of their search and the media types subscribed to.
Using guidance from research and interviews I prioritised the article metadata, which I then divided into three categories: significant, expected, and not needed or unusable.
The design had to work for customers who received a handful of items and scrutinised each one, to those who received hundreds and needed to quickly scan through for items of importance. The design needed to be relatively information dense but scannable.
Article Landing Page
Many customers of the existing product had become used to an email only workflow. While the long term goal would be to transition these users to more active use of the website, in the meantime the article page may be the only part of the site they would see outside of the daily email, as the landing page when they clicked a headline in their email.
While a clean presentation of the summary and media was important, there was particular value in proving additional value through a clear and useful presentation of the clip metadata.
The metadata sidebar had to clearly convey key information about the article. The design needed to be flexible enough to accomodate each media type’s distinct attributes, and to present a publicly accessible version with private customer information excluded.
To give additional context to the impressions and AVE values I designed a simple visualization. They show the values on a scale relative to the those previously received, giving the customer a sense of the value of this article in context. The orange dot allowed the user to get a quick feel for the relative value at a glance, before even checking the numbers, useful when working through a large number of articles to identify those of significance.
Tags had been grouped into people, companies, competitors, and other classifications; and color coded as such. This scheme was continued throughout the product.
It was important to ensure any visualizations simply enhance already clearly presented, readable information, and not replace or confuse it.
Visualizations
I initially worked with Accurat, a data visualisation firm, to develop a variety of innovative ways to visually extract meaning and significance from customer content. I gathered feedback and then worked to further develop and integrate the most successful and feasible options into the product.
The variety of content types and sources, and the fact that the data gathering and processing systems upstream were being developed simultaneously, meant it was quite challenging to reduce the possible chartable attributes to a reliable set.
Operational Efficiency
The internal workflow systems allowed print, tv, radio, and online news media from across the world to be gathered and matched to customers’ criteria. Carma had been conducting a relatively manual approach to this workflow, which we needed to introduce some automation to in order to scale Carma expanded and the volume of content increased.
The transition needed to be handled with care as customers could not notice a sudden change in the quality of content they received.
A new workflow was designed, with existing roles and permissions refined. New tools were designed to allow editors to validate and adjust the results of the automated processes. As these activities are repetitive and time sensitive - with articles scanned as the first edition of a newspaper is published and expected to be in the hands of a client by open of business - the interface needed to provide efficiency while maintaining quality.
Spending time onsite with editors in Dubai as they validated and released articles to customers in the early hours of the morning really helped.
We ran several internal beta tests; gathering feedback, tweaking the interface, and monitoring efficiency and quality metrics to validate and instruct further iterations.