2015 | Unicredit Group

Merchant Analytics

Business Intelligence, Visual Analytics, and Location Intelligence

A suite of interactive location intelligence solutions that visualize real-time geolocalized data on daily credit cards transactions, digital points of sale transactions, and branch activity.

Brief

In the spring 2015, UBIS, the IT company within Unicredit Group, contacted Accurat to imagine and conceptualize a suite of interconnected digital products that could leverage the value of data related to daily transactions recorded from thousands of Italian stores, gathered from the bank’s credit and debit cards and its digital points of sale (POS), and integrate it with other streams of data gathered from their network of physical branches, to understand how and when the services they offer are used by customers. The goal was to transform a large amount of raw and unstructured data into useful information to inform the bank’s commercial efforts. Some topics were considered particularly interesting: mapping the evolution of Unicredit’s POS penetration in the Italian market, monitoring commercial performance of different merchant categories, analyzing peaks of churn rates, and allowing potential future data monetization.

Accurat imagined, designed, and developed a scalable and functional web-based software platform that interacts with Unicredit’s data structure and back-end technologies. During several months of collaboration with UBIS, Accurat’s data visualization designers, interaction designers, and developers translated business inputs into an articulated and multi-layered tool developed with the most modern web technologies.

Solution

The applications provide quick visual access to an always-updated stream of data: transaction information that can be explored by geography, time, single merchant, or merchant categories, while branch activities can always be analyzed by client type (branch client, bank client, non-client), by type of service (atm, desk, self service) and organized over time or by geographical aggregates. To allow a wide range of internal users to perform a complete and critical analysis of this data, Accurat designed a simple interface based on interactive maps, customizable timelines, and a list of suggested KPIs based on performance. The geographical exploration of credit and debit cards’ data allows for the visualization of information at different spatial granularities; starting from an overview of the entire country, users can navigate to more specific areas, down to the district level. During the navigation of the different geographical levels, transactions can be viewed as an aggregate for the selected area. With a similar logic, the selected time frame can be modified freely at any zoom level thanks to a standardized calendar, enhancing the analysis around the transactional data and enabling the visualization of day-to-day trends. At any time, the information visualized in the application can be filtered by merchant or client category. The color-coding and the additional information on the merchant’s ATECO code (the Italian ISIC code) help to group the data and allow exploration of the transactions at a specific geographical level (a specific city, for example) by comparing its product categories. To allow users to dive even deeper in the transactions dataset, Accurat imagined and designed a “Merchant Summary” section of the tool; this view collects all the useful data around a specific merchant (type of merchant, type of account, group that owns the merchant) and enables the comparison between its transactions and the ones performed on average at other merchants of the same brand or of the same product category. When exploring physical branch activity, instead, users of the application can easily compare traffic data between locations or over time and break it down by service or by customer segmentation, focusing on either absolute or percentage values to map the interactions between the bank and its customers.
The transactions analysis is by design split between two separate exploration environments: ‘Acquiring’ (what the bank can discover from its points of sales) and ‘Issuing’ (what the bank can discover from its credit card clients); this clear separation enables different types of analyses on the collected data that are translated into clear visualizations for both the number and volume of transactions.
DETAIL - Acquiring amount
DETAIL - Issuing amount
DETAIL - Total amount

Process

When dealing with Big Data, unexpected drawbacks can easily arise if the scope of the project isn’t clearly defined and kept in check during development; these risks increase when working with multiple data sources, continuous streams of information, and different types of users. This is why Accurat’s team approached the design and development of the application with flexibility, based on continuous product iterations that allowed for multiple reviews with all stakeholders involved. This approach led to a constant and controlled redefinition of the requirements of the application during development, to ensure that both the technological and the design efforts followed a clear vision and ensured added value. Each component of the application adopts a custom stack of front-end technologies, selected and combined to optimize performance and maximize the effectiveness of the user experience on every different type of analysis. Accurat’s analysts and dataviz designers started by developing simple, static data visualizations with real data from the cities of Rome, Milan, and Palermo to identify the tangible operations that could be performed, and proceeded based on concrete observations rather than hypotheses. At the same time, Accurat’s interaction designers and front-end engineers developed the first concept into a more complete tool by working with a limited set of simulated transactions and activity data that served the prototype from an in-house temporary back-end, without dealing with complex integration tasks before testing how the user interface worked with real users. Once the actual usefulness of the analyses and the effectiveness of the interface were tested and approved, the two processes converged in the creation of the final tools. As the platform gained its skeletal structure, the “Suggestion” module, a predictive search engine with the capability of creating multiple comparisons with various benchmarks, was also added.

Machine learning techniques are used to create a ‘Suggestion List’, that is now widely used within the organization as an instrument to foster quick discovery of opportunities for direct monetization: this feature automatically highlights performance trends, both positive and negative, of geographical aggregates and single merchants in order to develop possible commercial strategies.

Results

The “Unicredit Location Analytics” suite is the first instrument the bank utilized with embedded data visualization principles that help present large amounts of data in a clear and flexible way without asking its user for any programming or statistical skills. The application is built entirely on widely adopted open source technologies, making it easy to expand and maintain by the bank’s internal IT teams.
Team
Paolo Corti
Giorgia Lupi
Pietro Guinea Montalvo
Alex Piacentini
Simone Quadri
Gabriele Rossi
Marco Vettorello
Tommaso Zennaro
Services
UX/UI Interactive Data
Visualization
Visual Storytelling

Dataviz, Branding, and Art

Interactive, Dataviz, Storytelling, and Experience