Customer data management: Pain (and opportunity) for every organisation

Customer data sits at the forefront of many of our clients’ data management concerns. Availability, security and quality of data on customers is not only a regulatory issue for the many financial organisations we work with; it is also a key pillar in most of their data strategies. When a firm sets out its vision for sophisticated analytics capabilities built on well-integrated corporate datasets, customer data is inevitably picked out as one of the most critical success factors.

And yet, achieving well-integrated, high quality, controlled customer data is far from a simple task. It demands an effective operational capability to source and manage such data throughout the customer lifecycle – something many organisations find a persistent source of challenges, including:

• Onboarding: A modern bank needs a lot of data simply to start doing business with a new customer. The totality of data needed to perform KYC, finalise legal agreements, make credit decisions and fulfil relevant products / services for a new customer can easily run to several thousand discreet data points to be sourced and validated; this sits in tension with the need to offer customers a rapid, painless onboarding experience.
• Aggregation: Analytics use cases often depend on being able to combine time series views of customer activities together with the static data on those customers. Both activity data and ‘static’ client data is frequently sharded across many system and organisational silos – by product, division, channel etc. Aggregating these (often extremely large) datasets into a common data warehouse presents a range of technical and organisational challenges
• Updates: Customer data is never truly static: names, addresses and contacts can change, companies can merge or be acquired, and the customers’ use of products and services can evolve over time. Processes to update customer records accordingly need to ensure that (a) changes are auditable, retaining precision over what was believed to be true about a customer at any given point in time; (b) updates are replicated consistently across all relevant datasets; (c) updates trigger relevant business workflows (e.g., additional KYC) where needed.
• Security: With customer data being automatically confidential in nature and subject to restrictions such as GDPR, financial firms need extremely precise methods to identify customer data across all their data assets, provide transparency over the ways in which it is accessed and processed, and demonstrate that there are effective controls in place to prevent breaches or other unauthorised use.

Each of the above is a complex architectural domain in its own right, which for most of our clients drives continuous cycles of technology and business change while consuming significant operational resources. We expect the complexity (and hence cost) of customer data management to continue growing over time, driven by organisation’s appetite to both capture more data on their customers from a range of sources, and to leverage it in increasingly diverse ways. Operational analytics in particular (in which insights from analysis of big customer datasets are fed back to and surfaced within front-office processes) are proving to be a source of many innovative use cases for client data, creating demand for so-called ‘reverse ETL’ technologies.

More broadly, we see the proliferation of data technologies actually creating as many problems as it solves – architects and CTOs now have huge-ranging choices in how they build applications and pipelines to capture and distribute customer data, the range of options adding to design complexity and in some cases hampering productivity in its own right.

Our recommendations to clients facing questions on how to improve their customer data management remain simple:

• Focus on business – not technology – priorities. The biggest cost drivers in customer data management are the huge (and yet still generally over-capacity) customer operations teams whose work amounts to validating and remediating data issues day-to-day. People who oversee these processes are generally best-placed to determine what changes might, especially in the short term, provide optimum return on investment in data management. ‘Automate everything’ is not always the highest priority (or even a valid objective) in many customer data management contexts.
• Always consider the customer’s experience of data management outcomes. If a customer sees an out-of-date name in an online banking portal, needs three different customer numbers and passwords to access poorly-integrated services, or isn’t confident that the bank will protect their data, it can rapidly undermine confidence in the corporate brand. The long term cost of failing to optimise customer experience often outweighs most over factors.
• Adopt a minimalist, rationalist approach to defining customer data requirements. Adding progressive attributes to customer datasets can attract hidden costs from many directions, most obviously the additional effort needed to process the data and the potential burden on clients to provide/validate it. Analytics functions can also bear hidden cost in terms of the effort required to curate attributes for use by business end users, and the ‘noise’ it can generate when trying to understand complex patterns. Customer data product owners would often do well to err on the side of ‘less is more’.

Customer data, more than perhaps any other of a financial firm’s data domains, is complicated by the degree to which data management is tightly integrated into day-to-day business operations. Viewed through a certain lens, capturing and managing customer data can, in fact, be seen as the primary function of any bank, hedge fund, or other service provider. Accordingly, it’s an area where effective data management can deliver real opportunities for differentiation in a competitive market.