Banking Review

Dirty data

Is bad data – inaccurate, outdated and incomplete information _
killing your customers’ experience and interaction with you?

How often has a customer called your company, only to have the operator ask for their details – again – or worse, to find that your company’s records are out of date, inaccurate or duplicated?

Although certainly not unique to the industry, this scenario is all too common in the world of financial services. Acquiring a single, complete and accurate view of your customers is not only beneficial to your company’s relationships with them – and would certainly improve your ability to market new products and services – but it would also significantly reduce your exposure to risk.

Let me explain. The concept of a ‘single customer view’ is nothing new, certainly not in the world of financial services marketing. Over many generations, companies have attempted to consolidate their various information systems and data sources but, sadly, these efforts have often been relegated to the end of the data lifecycle – that is, in the data warehouse.

It’s all well and good to have a neat and tidy data warehouse, allowing you to generate impressively accurate retrospective reports on the business, but this back office cleanliness does little to improve your company’s day-to-day operations.

So attention needs to shift to the front office, or where your business transacts with your customers. This is where bad data impacts on customer service. It’s also where bad data originates, through inaccurate entry, via differently formatted records across different business systems, or through variable naming conventions, to name a few. It’s a fact that every large organisation in business today grapples with inaccurate or incomplete data at some level.

Complicating matters is another fact, that most large organisations have multiple data and information systems spread across multiple operating divisions. This means the prevalence of bad data is exponentially multiplied across the organisation. Given subtle differences in format and capture, multiple records often exist for any single customer. Bad and duplicate data within one system is a problem – but multiplied across the many systems used to support typical financial services businesses, the capacity for accurate and timely processing becomes hugely impacted.

Forget for a minute the implications this bad and subtly duplicate data has on a company’s ability to properly market back to its customers; what does this say about its ability to meet its government compliance requirements or, worse yet, its exposure to risk?

Using just one example, what is the risk to a company of having a customer blacklisted for credit in one division, but, due to the link between two subtly different records, they are granted credit in another? And can companies be certain the information they produce for compliance reports is anywhere near accurate, or at the very least, up to date?

The challenge here is how to achieve – and maintain – a so-called ‘golden record’ for the business. This is a single, cohesive, clean set of quality data across the organisation, regardless of division or information system. In the golden record, Rob Mills is not Rob R. Mills or Rob Miller, and if the business happens to have two or more customers with the same name, a combination of technology and process allows them to be consolidated if they are indeed the same person.

But even if you redesign your business processes so that data quality is attained at the point of contact with your customers, and then used to create a consistent record across your organisation, how do you keep it from becoming out-dated?

The longevity of a golden record is threatened almost immediately by what I call significant data events: a change in a customer’s surname, a change of address or phone number, a change in a customer’s credit rating. These events occur hundreds, even thousands of times a day at multiple points in any large organisation. So, if your data governance practises and technologies are not fine-tuned to capture significant events, any effort to maintain a golden record is compromised.

Historically, recording significant data events has been difficult if not impossible using traditional approaches. Large, complex, proprietary systems were difficult to interface with, and didn’t integrate well with more modern business systems. Even today, large organisations necessarily maintain a mix of new and older systems, sometimes in the same division.

It’s therefore critical that the technology used to create a golden record of customer data should also be able to interface natively with each of these systems. This can’t be done without a standards-based, uniform integration layer, capable of detecting events in each of the connected data sources.

Event detection is an emerging technology approach to keeping golden records up to date, and its application with large financial organisations – many of which have grown by acquisition over generations – will provide profound benefits to the first movers.

While a golden record is arguably the starting point for establishing a successful financial services marketing campaign, the exercise has potentially even more important benefits for the business, reducing its exposure to risk, and greatly improving the data accuracy for compliance purposes.

Rob Mills is Vice President of the Asia Pacific region for Information Builders.