Poor data quality costs businesses, charities, and governments trillions of pounds each year. Most organisations talk of digital transformation, and as business becomes ever increasingly digitised and complex, this cost is likely to increase. Managing data quality issues is seen as one of the most significant challenges for leaders. Gartner research highlighted 60% of respondents cited data quality as one of their key data governance issues, along with bedfellow’s data literacy and silo-orientated attitudes, which can often result in a lack of investment in sustainable data quality improvement.
But how much should you invest in data and data quality? This question is easy to answer: Don’t bother investing in your data if you’re not willing to invest in its quality.
This may seem like a hard-line position, but unless you recognise that data quality improvement isn’t, unfortunately, a one-time activity but a continuous process that requires ongoing effort and commitment. As attractive as a big bang–silver bullet program may seem, this is doomed to failure and repercussions in the board room.
But deciding how much you should invest in data quality though can sometimes feel more like an art than a science. But what are the costs of getting it wrong?
The UK Governments COVID-19 test and trace system is an interesting use case. Designed to identify and track people who had come into contact with COVID-19-positive cases, it suffered a technical glitch, resulting in 16,000 positive COVID-19 cases being omitted from the official reporting.
What caused this monumental failure? An Excel spreadsheet reached its maximum file size, which caused data truncation and resulted in the loss of thousands of COVID-19 test results. As a result, individuals who had tested positive for the virus were not notified, and their contacts were not traced or advised to self-isolate undermining the effectiveness of the whole test and trace system leading to a huge loss in public confidence.
But poor data quality isn’t just limited to the public sector. In the infamous case of Knight Capital Group, which experienced a loss so large, the company had to be acquired by another financial firm to avoid bankruptcy after a software upgrade no longer in use was reactivated and started sending a flood of erroneous buy and sell orders to the market.
The data quality issue stemmed from the failure to adequately test and validate the software changes before deploying them to production, costing the company approximately $440 million dollars.
In both use cases, we see the potential risks of poor data quality.
But how do we improve data quality?
Make people care!
Before any progress can be made in improving data quality you have to make business leaders care about the issue. You may want to share the sobering outcomes of the 2 use cases above, which may help you illustrate the point!
The trick is to find the areas in which your leaders have a personal interest and show how poor data quality is detrimental to the business.
The first step is to expose the pain caused by poor data by identifying key business outcomes and priorities and showing how trusted high-quality data is critical to business success and can give you a competitive advantage. Find out the impact of poor data quality and connect its impact with your data and analytics initiatives. Look at historic risks and any internal audit reports highlighting the impact that your poor-quality data has or could have on your business. But don’t forget to validate your problem statements with your key stakeholders to make sure you use the right language and priorities. This is about winning hearts and minds without being the zealot in the room spreading misery.
Show & Tell (show the impact your poor quality data is having on the business)
Being terribly British and a nation of armchair activists, we all know when something is going wrong but doing something about it is different. We work in data, so use that data to demonstrate how your business is suffering and what your bad data is costing the business today. Identify those critical business processes and their owners and determine key indicators (KIs) that are the most impactful to those processes.
Connecting the impact your poor-quality data is having on your data and analytics initiatives can be a huge help. We increasingly see at board level senior management increasingly frustrated that their investment in data hasn’t reaped the rewards first promised. Use your data quality profiling to analyse critical data elements and their impacts on business performance. This way, you can provide concrete evidence of the impact of poor data quality on your business. If you can’t get the basics sorted, how can you hope to move on to the ‘sexy’ Artificial Intelligence and Machine Learning?
Identify the root causes of your problems and show how you can fix it
As we’ve highlighted, data quality isn’t a one-hit wonder but a process of continuous improvement requiring you to get to the heart of the problem. Don’t attribute all of your issues to a single area such as technology or data but be objective in scoping out the underlying causes in an objective way. By understanding the root causes and systematically sorting your solution scope based on your business priorities, you can identify the practical scope for improvement.
At Oakland, we’re experts in improving data quality and management and have many war stories on the value of data quality applied in business. For example, good process management in shipping data means less shrinkage of profits through over- or under-shipping.
Great data quality in warehouse management has led to operational efficiencies that reshape balance sheets and business models. Accurate and trusted customer data has enabled delightful experiences across retailers, infrastructure companies, and manufacturers.
Likewise, poor data quality defeats company efforts and delivers slim profit margins and gains in production, poor advertisement targeting, and poor customer service.
In short, quantifying the value of data quality is not an art, it is a part of each decision about how and how much business you will do.
If the data is poor quality, then anything that relies on it will likely be poor quality as well, if data quality is effectively managed, it will add significant value to products, services, and your company.
Please reach out to The Oakland Group for a data quality assessment to kick start your data quality improvements and so add value to your products and bottom line.
