We’ve all been in situations where people use quality of data as a smokescreen to avoid the real business conversation at hand. It’s a natural human reflex that causes an insane amount of inefficiency in any given organisation. It's also completely incompatible with business success.
We’re drowning in data
From my vantage point as CFO, I’ve seen a massive wave of tools and activities in the last few years with the goal of measuring everything. Instead of enhancing the way we all work, this abundance of data results in an overwhelming amount of material to sift through. It’s become incredibly difficult for decision makers to assess what matters, especially in finance.
Separating the wheat from the chaff in analytics is a really hard job; this in itself is not new. Organisations have been dealing with data quality management since the beginning of time, but now it becomes a scale problem. When you're dealing with a stream, you can manage it with a bucket and pail. When you have a Niagara Falls of data, you can’t process it fast enough in a manual way. If you don’t use some data management software platform and automation to help cut through it, you’ll just drown.
Looking ahead, the problem of scaling data only grows from here as organisations hunt for new customer experience touchpoints, explore avenues to reach new audiences, follow new trends, and grow more invested in customer success. Those activities require mountains of data, and data only works for you if it’s catalogued, processed, and applied appropriately.
Without the right systems in place, you’re potentially creating damage by feeding your organisation the wrong metrics, or metrics that are just wrong. CFOs would love to rely on IT for clean and compliant data, but too often that relationship is strained. How can we hold IT accountable while strengthening the relationship between our departments?
Corporate data management is failing us
First, let’s be clear: finance professionals aren’t getting what we need from our IT investment. Finance organisations need absolute certainty that can stand the test of investors, public auditors, and regulators. We put a massive amount of energy and work into establishing that level of control for our reported financials and metrics; it’s a critical standard which we test out every year. But what about the leading indicators that guide strategy and capital allocation? The question for any CFO should be “are your operating metrics and analytics audit-ready?”
At most companies, the answer is no. In a recent worldwide, cross-industry survey on corporate data health, only 38% of finance executives reported that they always trust the data they work with. Only 56% reported making even half of their decisions based on data. Across all departments, 64% of executives report making the majority of their decisions with data — 8 points higher than finance, but still distressingly low.
What’s going wrong?
As decision makers struggle to derive business value from data, it’s easy to group all kinds of challenges under the vague umbrella of “data quality.” That’s why it’s important to look a layer deeper than trust of data. Trust of what? What dimension? There are so many different ways to think about data trust: completeness, accuracy, accessibility, traceability, recency, relevancy, the list goes on. Until you unpack which aspects of your data are problematic, it’s hard to facilitate a conversation about why a decision maker feels hesitation or reluctance or indecision about some particular analysis.
Let’s rethink our relationship with IT
It’s hard to quantify ROI on IT. CFOs in general don’t feel like we get enough ROI for our IT spend. We perceive that we’re investing a lot and not seeing tangible results. Part of the problem is probably how high the bar is for quality in our own data.
The threshold for trust in your data is always a function of the decisions that you’re making. In the finance organisation, we spend all our resources and energy on gaining the most accurate view of what’s in the past. Our role requires highly accurate data in order to report the news, as opposed to very rapid data required companywide in order to make the news.
From IT's perspective, looking at organisation-wide data strategy, it’s challenging to apply the right degree of quality for these situation-specific needs. Perfection is often the enemy of good, but at a minimum we need enough insight to calibrate our expectations and make active trade offs.
Trust and transparency turn out to be very helpful as communication mechanisms between IT and Finance organisations. A platform that is able to give you at-a-glance visibility into the reliability of any dataset can help, because it quantifies the integrity of data in multiple dimensions so that decision makers can see how much to trust their data.
What if you could measure the performance of entire data management systems based on the trustworthiness and decision-readiness of corporate data enterprise-wide? Imagine if the IT leadership could prove to the finance leadership, “look, you can see our quality.” Both sides would find it valuable to have a metric to demonstrate the value proposition and prove it out.
What’s measured matters
This still leaves CFOs with two big questions:
1. How are we supposed to determine whether finance data is audit-ready?
2. How are we supposed to determine the ROI of corporate data management?
Today, companies lack systems that really serve as the backbone for reporting and quality assurance in enterprise data. It's not that we don’t believe in the value of data quality standards — it’s just a need that hasn’t been met. We strive to fill that gap by instilling trust across metrics that come from any system. This provides a uniform and consistent gauge of quality that allows decision makers to calibrate the degree of trust or assurance that you need for a particular dataset.
Of course, no one system or vendor can be a cure-all for data culture. Metrics themselves aren’t enough; they can only be successful when coupled with strong process, commitment, and organisational buy-in to ensuring sound maintenance of the data. That’s why enterprises in every industry must embrace the idea of governance and unite under common data standards. In our research 95% of executives agree there’s a need for cross-industry data quality standards.
Success depends on a central effort and combination of the software with the processes and systems that ensure data health. Once you can do that, you can prove the business value of your data. Suddenly the disconnect between investment and return vanishes.
Create a data-centric culture focused on data health
As we’re flooded with more and more data, it's important to remember that the context and usability of data is paramount. If you can systematise context in a way that a larger audience can consume and interpret it, you supercharge the business value of the data. The right data systems and culture grant people the ability to understand the “why” as well as the “what,” which means that the data isn’t just there — it’s decision-ready.
Good data governance is critical to teasing apart the actual challenges of working with data, and solving them at a human level. Democratising data with appropriate governance and universal quality standards ensures that you’re putting safe data in the hands of the people who need it to make decisions, and that those people understand the life history and quality of the data they work with.
That’s why it is important to embrace a data health approach. Corporate data is no longer a collection of assets that people share with particular tools; it’s a living thing that requires care and feeding, a common language, and common understanding. Everyone who works with data needs to have the context to understand the data itself and its value to the business, and shares responsibility for maintaining the visibility and reliability of that data.
Once you achieve data health with transparent and trusted data processes, not only can finance and IT agree on the ROI of data management investment, but no one can fall back on data quality problems as an excuse.
