Analyst Insight: Corporate data is a $200 billion problem, and a problem that generates more CO₂ than all the cars in South America.  Is data a valuable corporate asset or is it a waste by-product of business processes? 

The estimated global data footprint is now 97 zettabytes, with another 23 zettabytes forecast to be added in 2023. A zettabyte is a trillion gigabytes. Ninety-seven zettabytes equates to 24 smartphones full of data for each person on the planet. Using the 6.84 billion smartphones in the world as a proxy for personal data, assuming smartphone data is half of personal data, and that all personal data is duplicated in cloud back-up, the global personal data footprint is approximately 15 zettabytes. This leaves commercial activities to account for 82 zettabytes of data, or 84.5% of the global footprint. While 82 zettabytes is an extremely rough estimate, the magnitude of this footprint is worth bearing in mind while assessing the environmental and business impacts of the global data footprint attributable to business activity.  

What is the environmental impact of 82 zettabytes of data? The most conservative estimate for the carbon footprint of 1 Gb of data stored to a data center or the cloud (McGovern) is 0.015 kWh per GB, and 0.0042 kg (0.0093 pounds) of CO₂ per GB. For 82 zettabytes of business data, the total carbon footprint is 1.23 trillion kWh and 344 billion kg (758 billion pounds) of CO₂. To bring 344 billion kilograms of CO₂ into perspective, these emissions equate to almost 75 million automobiles (more than the total vehicles in South America).  

One of the challenges with driving environmental sustainability is the business case. For the commercial data footprint, the business case is clear. Using $0.165 per kWh as the global average cost of electricity, the 82-zettabyte data footprint will cost global businesses roughly $203 billion in 2023. This cost includes only the storage, not transmission, validation, access or security.  

Is 82 zettabytes an acceptable “data overhead” for business? Should firms accept this and focus on areas with greater potential for sustainability? The cost of energy ($203 billion) is only 0.18% of the value of world trade (approximately $112 trillion).  Viewing this overhead differently, however, 82 zettabytes supporting $112 trillion of global business equates to 728 megabytes of data per dollar. This is the equivalent of a 2TB laptop supporting a paltry $2,700 of annual business revenue. While these calculations are rough and the assumptions many, equating the global business data footprint to only $2,700 of business value per laptop underlines the need for more sustainable data management practices.  

Read more: 2023 Supply Chain ESG Guide

When developing sustainable data strategies, organizations need to come to terms with a critical conundrum — is data a valuable corporate asset or is data a waste by-product of business processes? The answer is both. The business value of information (where information is defined as organized and contextualized data that provides logical meaning) is undeniable. In the information context, data is a corporate asset. Classifying some data as waste may run contrary to the mantra of “Data Is The Biggest Asset On Your Balance Sheet“ (Forbes, September 2020). Forrester, however, reports that 60% to 73% of all data within an enterprise goes unused for analytics, while Google estimates 30% of data on the Internet is a copy of other Internet data. Veritas Technologies reports that 33% of enterprise data is redundant, obsolete, or trivial (ROT) and 53% of enterprise data is of “unknown value.” 

Once an organization has accepted that not all data has value, the organization can develop and implement data waste management strategies to reduce data footprint (and the associated environmental impacts). Typical data waste management strategies can include: 

Data governance and data ownership: Any strategy to manage an asset or waste is only effective when an owner is identified and responsibility assigned.  Data, whether an asset or waste by-product, is no exception. 

Reduce data waste at source, including:

• When determining data requirements, work backward from user needs, not forward from the available data.  Retention of data “in case we need it” leads to volumes of ROT data.
• Improve data quality at the source so that the data can be consumed by analytics and so that the data can be easily classified for retention policy determination. 

Data classification, data retention policies: The challenge with ROT data is that is does not rot.  Data governance needs to include retention policy classifications in the same manner that information protection classifications are rigorously applied to all documents.   When to delete data or move the data to off-line storage should be defined when the data is created, not applied retroactively when data volumes become problematic. 

One source of truth: Enterprise systems are sometimes a mix of similar data sources with different owners and overlapping purposes.  For example, cost accounting may manage the accounting value of inventory while sales has their own inventory system data, order fulfillment manages a third view of inventory data, and enterprise analytics takes data from the other three.  Multiple sources of truth for the same information exponentially increase the data footprint and dilute enterprise data quality.  

Rationalize systems and adopt composite application strategies: Application System rationalization reduces data duplication, reduces storage occupied by applications, and simplifies enterprise data management.  Composite application strategies using interchangeable components and robust fabric connects reduce resource requirements and data footprint in comparison to siloed systems.  

Solution selection: In addition to striving for one source of the truth, the data footprint should be considered in software and solution evaluation.  Systems that leverage existing data sources should be favored over solutions that require their own separate data.  Extend total cost of ownership calculations to include environmental impacts.

Programming and architecture: Different programming languages, programming patterns, and algorithms consume different amounts of energy and require different volumes of data for the same task. Consider this when making programing choices.  

The benefits of reducing an enterprise data footprint extend beyond cost reduction and more sustainable business practices.  A larger data footprint may increase security risks.  Unnecessary retention of customer and personal data complicates administration of privacy policies such as the right of an individual to request deletion of their personal data.  Data duplication threatens data quality and the usefulness of data for analytics. 

To sum up, the global data footprint of commerce is roughly 82 zettabytes.  This footprint comes with an environmental price tag equal to the emissions from 75 million automobiles, and an annual financial cost of $200 billion for just the electricity.  Once firms come to accept that not all data is valuable, meaningful strategies to approach data waste, in order to reduce environmental and financial impacts, can be executed.  Data waste reduction is much more than an IT strategy; it is business strategy that drives sustainability, marketing, branding, compliance, profitability and risk mitigation.