In our previous post, we showed that since business costs are increasing, it is wiser to reduce data costs and maximize the use of current data and analytics tools than it is to invest in new tools that have yet to prove their worth. Indeed, the data and analytics tools you have now can help you attain increased revenue, profit, operational efficiency, speed to market, and employee retention. Naturally, these benefits have costs. This blog post aims to help you manage those costs effectively.
If your organization is like most, then yours spends a lot on data. You spend time, effort, and money on storing, processing, cataloging, distributing, and securing data. Therefore, the faster you find the data you need, the better. The same goes for improving the way business units utilize and spend on cloud resources. Let’s take a closer look at some of the most impactful ways to reduce data costs.
Activate metadata
While the statement “metadata is data about data” is often said, its implication is not discussed enough. Metadata covers many of aspects of data, such as:
-
Details regarding creation – Who or what created the data? When and how was it created?
-
Context – What business purpose does this data serve?
-
Security and permissions – How sensitive is the data? Who has control over it? Who can access it? What can authorized parties do with the data?
-
Operational characteristics – Who accesses the data, and how often do they do so? What are the changes applied to the data?
-
Data quality – How accurate, consistent, and/or complete is the data?
The problem with metadata isn’t in not having it, but rather in having it, but not activating it — i.e., using it to better manage data. All of the aspects of data listed above helps enterprises do the following:
-
Enjoy a comprehensive, high-level, and up-to-date view of your data. You can have a clearer view of data lineage (i.e., the map of direct and indirect dependencies between data entities in your environment), context (i.e., the origin and history of your data) and compliance, (i.e., the degree to which data is being used appropriately and in accordance with regulatory requirements).
-
Monitor data quality in real time. ML-powered tools can assign values such as data accuracy, completeness, and consistency and alert data managers when quality trends downward to prevent poor data from negatively affecting decision-making or operations.
-
Automate data governance. ML-powered tools can automatically fulfill tasks pertaining to data retention and other actions relating to data regulations compliance.
-
Improve decision-making via analytics. ML-powered tools can enrich metadata by adding more context and insights into data, such as trends, patterns, and correlations. For example, the tools may find that most customers provide negative feedback via social media comments, so decision-makers may find it wise to focus their customer service data gathering and PR campaigns in social media channels.
Now that enterprises can’t help but gather ever-increasing amounts of data, the ability to obtain the most meaningful and impactful data in the shortest amount of time is a competitive advantage. That is, if your organization can focus on what’s truly important, then it can save time, effort, and money from being spent on less worthwhile tasks.
Let’s take the data generated during canned food production, which include canning dates, who were the persons handling production, and the identification numbers of the machines utilized to produce said items. Metadata for such data may include descriptions for what the data is (e.g., ‘canning data contain details of the circumstances surrounding the production of each canned food item’) and why it was created (e.g., ‘canning data helps trace when and where an item is produced so that if and when an item is discovered to be of poor quality and/or poses a danger to consumers, the rest of the items that were in the same batch as the defective item may be easily recalled, and the factors that contributed to the drop in product quality may be corrected’).
Therefore, in the event of a product recall, human teams or AI-powered applications can use metadata to quickly and accurately find the data needed to execute the recall — no time is wasted on aimlessly sifting through virtual piles of data.
Improve data governance
When developing new products, who in the enterprise ensures that proprietary information is kept in the strictest confidence? This and other questions pertaining to ownership of and accountability for data assets are answered by data governance.
As an organizational system, data governance defines who exercises control and authority over data assets and how these assets may be used. While this as a concept may seem simple enough, many things add complexity and cost to data governance:
-
Compliance with changing regulations from different jurisdictions
-
Differing data requirements of different internal stakeholders
-
Differing protocols for different business units
-
Countless data streams to manage and monitor
-
Inconsistent data quality across the organization
All these challenges can be addressed by comprehensive continuous improvement of the existing data governance framework of an enterprise. For example, standardizing security protocols across the entire organization ought to streamline processes and decrease the time and effort (and ultimately, cost) to audit data processes and keep data secure.
Optimize cloud spend with FinOps
One of the primary functions of an organization’s finance department is to utilize financial analysis information to aid in decision-making. To illustrate, projected fuel price increases may result in the prioritization of goods that incur the least transportation cost.
FinOps works in a similar fashion but is more focused on how business units spend on the cloud to derive business value. For instance, the engineering team may be spending the most on the cloud to operate consumer-facing apps and build new features for it, but the returns the apps are getting from ad and subscription revenues dwarf revenue from other sources. FinOps enables other business units to see this and build upon the engineering team’s success. One concrete way those units could do so is by handling shared cloud costs better via a FinOps practice called cost allocation.
In more general terms, FinOps empowers engineering, finance, and other relevant teams to work together so that they can:
-
Implement cloud best practices, such as utilizing cloud resources from a common pool.
-
Build the most efficient cloud architectures for enterprise applications.
-
Manage and govern the company’s cloud in a unified manner via an interdepartmental cross-functional team called a Cloud Cost Center of Excellence.
-
Negotiate better cloud terms with providers.
Integrate more data into decision-making processes
Obviously, collecting, storing, and processing data incurs costs, but no organization would do those things merely for the sake of doing them. Rather, they do these so that they can obtain insights that help them make the best business decisions possible. However, there are many barriers to putting data to work:
-
Data is kept in silos; and teams fight over whose data is correct.
-
There are few to no data strategies that support the achievement of business objectives.
-
Analysis is performed only when a special purpose is identified (as opposed to being consistently performed to inform business decisions).
-
The quality of data provided to data consumers across the organization either varies wildly or is mostly poor.
-
Data literacy among personnel outside of the IT department is poor.
Enterprises that have not yet overcome such barriers are said to exhibit data immaturity. In sharp contrast, enterprises that utilize data and analytics as an integral part of their business strategy are said to exhibit the highest level of data maturity or be simply described as being data-driven. To illustrate, B2B sales managers who used data and analytics to identify and prioritize sales growth opportunities, activate customer touchpoints to maximize sales ROI, empower sales personnel with relevant insights, and iterate for continuous process improvement saw EBITDA increase by 15 to 25%.
Growing into data maturity doesn’t happen overnight. Rather, it is a gradual organization-wide transformation that Lingaro can help facilitate for your enterprise. Technical requirements will vary from business to business and thus require intensive assessment.
Data spend is only half of the Data ROI equation
Optimizing data spend is but one side of the Data ROI coin. In Part Three of our Data ROI series, we’ll look at the coin’s flipside, namely maximizing the business value created with data and analytics. In that post, you’ll discover how to do so via methods like use case prioritization and tools like Lingaro’s Intelligent Insights.
Read more articles on Data ROI:
Data ROI: Challenges and Opportunities
Data ROI: Maximizing Value From Data and Analytics
