Information and Insights

A common challenge identified for a large portion of CPG companies was the ability for end users – from senior leadership to operational staff – to ingest the insights being delivered by current reports. As described previously, many CPG companies indicated that over the last decade they have moved from “information poor” to “information rich” but not necessarily resulting in better insights and decision making.

Further research revealed that a key reason for this shortcoming was that insights were getting “lost in translation”. Information was not being delivered or “served” to end-users in a format that was most conducive for them to efficiently garner the relevant insights to aid in their decision making. This is perhaps mostly due to decades of users being overwhelmed by simple row-and-column reports, bar charts, and pie charts. These methods of conveying information often still require individuals to exert significant energy to cobble together the required insight to make a decision. While these formats have not always been intuitive for “small” or existing internal data, they will be even less effective when introducing the volume and variety challenges posed by Big Data. Many Big Data assets will simply be impractical to view in rows and columns due to their less structured nature (e.g., text, video, digital image, audio).

Many of the more mature analytical CPG companies have recognized that their end-users (e.g., sales, marketing, supply chain), like most people, are more visual learners, and therefore have started to adopt and use new visualization capabilities. These emerging data visualization methods use new formats to “bring data to life” with intuitive, interactive interfaces. The visual experiences can be shared with end users on tablets and other mobile, touch-screen devices allowing for much deeper interaction and exploration. This interactive capability allows decisions to be made in real-time rather than the delayed results typical of traditional ad hoc analysis.

New visualization techniques also align with a rising trend in which employees are expecting corporate tools to be as usable and intuitive as the “apps” they use in their personal lives. This trend is commonly referred to as the “consumerization of IT.”  More analytically mature CPG companies recognize this shift, and as a result, have started to reengineer their information delivery tools to be aligned more around the user – adopting principles that have long been used when building consumer-focused applications. This means moving from a traditional “user-requirements” based approach (commonly associated with ERP or other enterprise technology) to a “user-experience” based approach. This evolution involves new approaches for designing information delivery tools involving personas, journey mapping, and other techniques. Ultimately, the goal is to ensure that any investments in capabilities supporting decision making are tailored to the specific needs of the end user.

For a more detailed description of how and why CPG companies are using new data visualization techniques, please see the video clip “Big Data Visualization – From User-Requirement to User- Experience”

The new user experience: data visualization

The near-term opportunity that data visualization presents is its potential application to many current CPG reporting use cases. Data visualization can also leverage existing investments in the underlying infrastructure (e.g., MDM, data integration, EDW). For example, PepsiCo and Safeway worked with Deloitte to leverage data visualization capabilities to better understand their joint challenges around managing out-of-stocks and optimizing inventory levels.


There are many additional use cases in which data visualization capabilities may be applied when Big Data is involved.

For example, data visualization is being used to understand:

  • “Buzz and Chatter” fueled by unstructured text data from social media

  • Sources and nodes of the supply chain

  • Early warning signals for food-borne illness and other related product quality and safety issues

Paula Davis from Deloitte's Highly Immersive Visual Environment provided a brief perspective on various types of data visualization that can help companies develop analytical maturity in the category of "Information and Insights."

Deloitte's Highly Immersive Visual Environment (HIVE) Highlights Several Forms of Data Visualization


Your data - especially your Big Data - is complex and multi-dimensional. Visualization can help you to interrogate that data to find answers to the business questions you need in order to manage your business. Below are a series of visualization examples and the questions they can help to answer across your enterprise. 


Streamgraph is a technique for displaying large scale time series based dataset and can be used to overlay multiple time series data.  This type of visual could be used illustrate changes in sales volume over time for different products or geographies.  Because of the use of thickness and color of each stream, it is a more compelling way of showing temporal data than a stacked bar chart.


Force-directed graphs

Force-directed graphs are used to display networks with multiple interconnections.  If each distinct node in the network represents an individual, the lines relationships, and colors show brand preference—the visual would allow sales and marketing teams to monitor product adoption among the groups and seek out those who are centrally located to target populations with special offers.


Tree maps

When data volumes are very large and hierarchical, tree maps might offer a good way of summarizing and interacting with the data.  Colors on the graph can indicate categories and size of the rectangles can address the volume of items in each category. 


Radial visualizations

Similar to the tree map, radial visualization techniques allow viewing multiple levels of data while keeping the context of the entire data set.  This is a great replacement or alternative to a pie within a pie chart visual.


Bubble charts

Bubble charts are often used for visualizing data that have more than two components.  Bubble charts can be combined with a scatter plot to add additional information through the uses of size and color.   For example, this could be used to show revenue growth over the years with color representing the type of organization along with size representing profit.


Many Eye Bubble charts

This type of bubble chart uses color and size to categorize and group items based on multiple parameters.  Similar to a tree view size could be used to represent volume while color and placement can be used to represent category.


Word Cloud visualizations

An increasingly common visualization technique used to represent trends in social media is a word cloud. Adding color and even relative position as a dimension, it is possible to further segment the visual.  The size of the word can translate to the frequency of occurrence or relative relevance of the word compared to the rest of the text based data.


Time Series visualizations

Time series visualization is often used to represent data that has a real time component to it.  For example, this type of visualization can be used for monitoring twitter feeds, or stock market price variation. 


Geospatial visualizations

Geospatial visualization is used to find spatial relationships in data sets or to overlay data with other geographic information, such as demographics, weather data, or geologic and resource information.  It can be used to understand the distribution of various indicators spread across different locations.  Geospatial analytics can provide ways to look at the KPI’s of stores across the US to analyze which regions are performing well.


Parallel Chord visualizations

This type of visual, commonly used on data sets with a large number of dimensions, can help to highlight how various attributes from across disparate systems interact with one another.  This visual can help to discover insights in data by spotting trends and cause/effect.  The chords or threads show connections between dimensions while the colors could represent products, days of week, under/over stock, or just about any other type of dimension associated to the data.


Authored By Paula Davis - Deloitte Highly Immersive Visual Environment (HIVE)