From Cost Centre to Value Driver: Quantifying Data ROI

Introduction: The Persistent Perception Problem

Like many business functions, Data and Analytics capabilities are seen as a cost-centre. A necessary cost, rather than a strategic investment. Along with this, many organisations perceive their data as being merely a by-product of their business processes.

In the context of this, many business leaders, especially data leaders, are under increasing pressure to demonstrate tangible business value and return on investment (ROI) from their data assets. Factor in the growing appetite for “doing AI” and organisations really do need to articulate the business value of data and AI initiatives.

In this article, I am going to provide a framework for quantifying the value, and subsequent ROI, of data and data-related initiatives, including how to effectively communicate this value to the business.

Why Quantifying Data ROI is Crucial (and Challenging)

Historically, when asked to describe the business value of technology (and data) investments, the answers were always abstract, along the lines of “we can make better and quicker decisions”.

As data landscapes become more complex, building data solutions becomes more expensive and business benefits that are only anecdotally described are not going to make a good business case. This becomes even more challenging when trying to describe the value added by foundational activities such as data engineering and data governance.

It is therefore imperative that the business insights that are served to business stakeholders as well as the foundational data work be mapped to very specific and measurable business outcomes.

The obvious, positive business outcomes that organisations seek are revenue growth and cost reduction, but we must not lose sight of other benefits such as risk reduction/avoidance or measurable efficiency gains. Once these can be clearly identified and explained, executive buy-in and budget allocation becomes easier. More important than this though, the strategic importance of a data function or capability becomes apparent.

The next section lays out a framework that can be used to help articulate this value.

A Simple Framework for Measuring Data Value

Whilst a little more time consuming, I recommend taking a top-down approach that involves first understanding exactly what the organisational strategy and goals are. Alongside this, it is key that there is an agreed set of objectives that will support any business case or funding request. As mentioned in an earlier section, a good start is:

• Revenue generation or growth. Any data initiative that achieves this outcome by enabling things like new product development, more targeted marketing, improved sales effectiveness, or new market opportunities
• Cost reduction or efficiency gains. This would be any initiative that results in process improvements, introduction of automation, or reduced operational friction
• Risk mitigation or avoidance. A data initiative would have to improve compliance (avoid fines), enhance security, or support better forecasting
• Strategic enablement. Any data initiative that is core to the organisation’s existence, for example, the ability of a financial services organisation to retain a banking licence or a pharmaceutical company to retain its manufacturer’s licence and wholesale distribution authorisation

Next, consider the organisation’s strategic goals or objectives. To bring this to life, we'll use Lloyds Banking Group as a real-world example. Their 2025/26 strategy contains for core objectives (or pillars):

• Deepen and innovate in Consumer, measured by a three percent increase in depth of relationship and an increase of about 50 percent of active customers served per distribution full time equivalents.
• Create a new Mass Affluent offering, measured by a more than 10 percent increase in Mass Affluent total relationship balances, including assets under administration.
• Digitise and diversify our BCB business, measured by maintaining small business deposit market share and by digitising 50 percent of key servicing interactions
• Develop our Corporate and Institutional business, measured by an approximate 45 percent increase in CIB other operating income and a greater than 5.25 percent increase in income over average risk-weighted assets

At this point in time, it is already evident that Lloyds Banking Group will need the relevant data to highlight how the organisation is progressing against achievement of the stated targets.

Unfortunately, this in itself, does not tell us anything about the value of the data needed. Nor does it help us understand how data will help Lloyds achieve its strategic objectives. To do this, we need to delve a little deeper. Lets consider how Lloyds would “Deepen and innovate in Consumer”. To achieve the stated goal of a three percent increase in Depth of Relationship (defined as product holding for customers retained since 2024) , certain activities need to take place, either projects or business as usual processes, usually a combination of the two. This might look something like:

• Determine what success looks like (i.e. is a three percent increase in depth of relationship reflected by a cross-sell ration uplift in absolute product numbers, a three percent increase in interest income from those customers, a three percent increase in the lending book, or something else)
• Understand who the customer is, so that it is known which products they hold
• Identify customers that present a cross-sell opportunity (i.e. credit card, short-term loan, ISA account, etc.)
• Determine and execute the marketing and sales initiatives to achieve this cross-sell
• Measure the progress and outcome
• Adjust actions where necessary

Lets imagine, to achieve the ambition of successfully delivering the above, a requirement is building a single customer view and segmentation model. Using the success metrics (e.g. a three percent increase in interest income) vs. the cost of building and maintaining a single customer view and segmentation, Lloyds can calculate what a likely return is on their data investment. This should, of course, be done using a robust return on investment model, considering all the cost factors and time value of money.

Whilst this is a simplistic example, because a single customer view would have other benefits (both tangible and intangible) supporting other business initiatives, it illustrates how tangible outcomes and return on investment can be assigned to data projects and investments.

Most importantly, it ensures that data investments are made in support of achieving the organisation's strategy, and not because analysts or peers are saying that every organisation needs a single customer view.

It would be remiss of me not to highlight that simply delivering a data programme, project, or initiative will not ensure success. Organisations need to ensure that the quality of the delivered (data) asset is of a high standard and able to support the intended strategic initiative and that the necessary effort is invested in the organisational change efforts required to drive adoption. We’ll cover this in the next section.

Practical Steps and Metrics

Not to be confused with the business metrics discussed in the previous section, you should next seek to define some metrics that can be used to measure how well your data is supporting the organisational objective you have identified.

As with business metrics, you need to ensure that you are selecting the right data metrics, ideally balancing leading and lagging indicators. For example, usage metrics would be lagging indicators in that they describe an event that has already happened. In contrast, something like data quality metrics would be leading indicators inferring that poor data quality would reduce the quality of decisions made or decrease trust in the data.

Some ideas for these types of metrics are:

• Data quality metrics such as data accuracy, completeness, timeliness, consistency, and issue resolution or remediation times
• Data management and operations metrics such as availability, uptime, and performance
• Data governance metrics such as critical data elements curated and governed, data stewardship, and data ownership
• Data cost reduction metrics such as the reduction in spend on legacy applications and infrastructure

It is important that each of these metrics are baselined so that progress can be tracked. The metrics should then be presented in data value scorecards or dashboards and shared with relevant stakeholders.

Communicating Value Effectively to Stakeholders

Instead of just publishing the data value scorecards or dashboards, stakeholders should be actively engaged and encouraged to understand the significance of the value added by the organisation’s data. This messaging should be tailored to different audiences such as Finance, the C-suite, and business unit leaders. Stay away from technical jargon and use this approach to build a narrative around data as a strategic asset and value driver.

Don’t wait until your data programme is complete (not that they ever are!) but identify a quick win that you will focus on before even starting your programme. Once this goal has been achieved, celebrate it and demonstrate it to stakeholders.

Conclusion: Making the Shift

Being able to demonstrate ROI is key to elevating the organisation’s data capability from a cost centre to a strategic investment. Remind stakeholders that linking data efforts to business outcomes needs to remain deliberate and concerted.

I would love to hear about your own challenges or successes in quantifying the value of your data so feel free to leave a comment or message me directly.