TDBank Group, the second largest financial institution in Canada, began re-evaluating the relevance of its own data a few years ago as branch traffic declined -- but without a subsequent increase in digital transactions.
Financial institution data analysis suffers from what Przyklenk referred to as the four S's:
- Siloed, with several systems or applications for different products, business lines or stakeholder groups, and limited integration
- Structured, with several layers of data ownership, governance and access levels, requiring vast institutional knowledge
- Super-expensive, with internal cost structures including chargeback models for processing time and bandwidth
- Strange, as some systems are older than the analysts using them and the information generated is difficult to understand
A fifth S, speed, is actually a benefit. Financial institution mainframes tend to offer better processing speed than many other systems. However, the structured and unstructured data available from an increasing number of sources overwhelms traditional databases and systems, meaning that analysis isn’t possible using traditional tools, Przyklenk said.
“A lot of the data is dirty, and data scientists aren’t happy with that,” he said. “You can include all sources, but there are consequences of doing that. The proliferation of unstructured data actually decreases accuracy without proper ETL or robust queries in place to account for variation.”
There’s also an opportunity cost for investments made to collect and analyze Big Data – the funds may be of more use in other parts of the organization, Przyklenk advised. Marketers should keep this in mind as they present their data-related business cases to management.
"Choose what you need based on your business requirements. There’s no value unless you’re sure how you will make money (with additional data analysis)," he said.
Among the data collection and analysis investments that yield the most marketing value, Przyklenk said, are a centralized data warehouse, a data management platform and customer journey mapping.
Data Warehouse Benefits
A centralized data warehouse enables marketers to consolidate all reports in one place, provides data attribution modeling and can be integrated with CRM and transactional records. It also provides a company with increased context for digital marketing decisions, while enabling analysts to focus more of their time on running their campaigns rather on simply consolidating available data.
“Higher quality insights will result in lower acquisition costs,” Przyklenk said.
Przyklenk suggested including the following data in a data warehouse:
- Daily granularity keyword data from paid and organic search platforms
- Paid search cost-per-click data and campaign metadata
- Email and direct mail campaign metadata and tracking code usage
- Micro and macro conversion data
- Transactional details including products, promotions, discounts and cost of goods sold
Data Management Platform Benefits
Prospects and customers have come to expect personalized, customized offers based on behavioral and relationship data, but marketing channels today aren’t fully integrated to make these offers, a challenge that a data management platform will help solve, Przyklenk said.
The platform will provide organizations with a 10 percent cost reduction in advertising as well as increased conversion rates, Przyklenk said. An additional soft benefit is higher overall customer experience scores and loyalty due to the platform’s personalization capabilities and potential for targeted advice.
Przyklenk recommended including the following data in a data management platform:
- Primary sales channel transactional logs (online, mobile, phone, retail)
- CRM metadata such as age range, demographic, product purchase and return history
- Advertising campaign creative and classification metadata
- Third-party cookie pool or list data
A 360-Degree Customer View
Customer journey mapping is designed to provide a 360-degree view of the customer, which many companies don’t have today because information is often siloed, limiting marketing’s efforts to cross-sell and upsell. Omni-channel integration will help break down these channels, permitting cross-channel optimization and reducing operating costs by up to 5 percent.
Przyklenk advised including the following data in customer journey mapping:
- Raw Web analytics "clickstream" data
- Phone channel or IVR data
- Retail transactional (purchase and returns) data
- (Optional) social media interactions, both direct messages and indirect data, such as sentiment analysis
"Take a balanced approach to Big Data," he concluded. "Consider the resourcing/people impact and the cost-to-benefit impact. Keep costs in check by including only the data that you absolutely require."
He also recommended pursuing marketing efforts with the greatest impact (the low-hanging fruit) and documenting everything to illustrate the benefits to management.
Phillip J. Britt writes for a number of technology, financial services and business websites and publications, including BAI, Telephony, Connected Planet, Savings Institutions, Independent Banker, insideARM.com, Bank Systems & Technology, Mobile Marketing & Technology, Loyalty 360, CRM Magazine, KM World and Information Today.