Performance and Insight Optimization for Banking
Analytical Customer Relationship Management Evolution
Contributor: Gerhard Held
Analytical CRM for Banking
In 2007 Thomas Davenport and Jeanne Harris convincingly described in their book “Competing on Analytics” how business insight based on analytical analysis helped companies around the world to build distinctive capabilities that resulted into a lasting competitive advantages. Some companies even built their whole business model on the usage of analytics for critical business processes, such as, Amazon, Best Buy, Capital One or Barclay’s Bank in the UK for cross-selling to their consumer finance line-of-business. Superior business insight pays back on a large scale if it is backed up by management support and organizational alignment. In the case of Barclay’s it resulted in 25 percent increase in revenue per customer account.[1]
Analytics can be applied to many business processes. One of the most successful and best-known examples is analytical customer relationship management. IT Analyst firm Aberdeen reports results of a survey that extensive usage of “data driven marketing” results for the best-in-class performers in 148% mean-class Return on Marketing Investment (ROMI) and 63% mean-class growth in revenue[2].
Due to the recent recession with particular implications for the financial industry banks are fiercely competing on new and current customers. Current customers are in particular focus as they generate the bulk of revenues and they are on average five times cheaper to keep than to acquire new customers. Competition on financial markets is intensified through new market entrants (on-line banks, retailers) with different price structures and aggressive conditions. In this competitive environment current bank customers are very attractive for the competition. Customer loyalty has decreased and time to anticipate and react on changed customer behavior has decreased as well. Therefore planning and careful targeting of marketing campaigns to increase the loyalty of existing customers is currently one of the best investments for a bank.
While the above mentioned financial rewards are impressive they are evidence of a somewhat longer journey to become an analytic competitor. In this short publication we would like to describe some major steps of this journey for the specific case of analytical Customer Relationship Management in banking in sequence of consideration below.
Data
Data is the lifeblood of all fact-based decision-making. The banking industry is particularly rich in customer generated or market related data (see figure 1). ERP systems assemble accounting and controlling data whereas banking transactions are collected in line-of-business oriented transactional systems. Commercial software solutions, such as, SAP® Banking Services are useful to streamline these data sources and deliver standard reports through the SAP Bank Analyzer. Access protocols inform about the usage of digital channels (web, email or social networks) a new and very rich source of information. Operational CRM systems, such as SAP® Customer Relationship Management collect customer interactions and most banks contrast internal information with external sources about the state and trends of financial markets, external credit ratings and geo-location data with relevance to their customer base.
Figure 1 Data Sources in Banking
Data and particularly customer data involve three key issues which need to be addressed to supply value for analytical CRM: data quality, size and form of the data. The quality of data has been cited as one of the major inhibitors to analytical CRM (Aberdeen Group, 2009). As a prominent example a bank in the UK (company name disguised) wanted to find out how many customer it had. As a first attempt it used data assembled across various banking systems and counted twenty five million customers. When cleaning this data and eliminating duplicates the bank found that this number reduced to seven (!) million customers.
The pure size of transactions in line-of-business systems used to prevent financial companies from analyzing this data for analytical purposes. Other industries, such as, telecommunications and retail were first to successfully apply transactional data to predict customer churn/attrition. Recent advances in in memory-based computing, such as, SAP HANA open up a fascinating new avenue to uncover the value of customer behavior buried down in transactional data.
What is required is to finally assemble this data in a form that is readily consumable for analytic CRM. Traditional source - or product oriented data need to be combined to customer-oriented data marts which serve a particular purpose (analysis of line-of-business or specific goals such as retention). At later stages of the analytical CRM journey these data marts will turn into departmental or corporate warehouses as is typical of the best-in-class analytical competitors described by the Aberdeen Group, Davenport/Harris or Peppers/Rodgers (2010).
Analytical CRM Creates Complexity
“There’s no shortage of customer data, but there is a shortage of insight about how to grow and use customer data for better decisions.” (Mike Linton, Ex CMO Best Buy and eBay).
The final goal of all marketing activities is to present the right product to the right customer at the right time using the right means of communication. If we visualize this goal (see figure 2) then this almost immediately raises a number of questions both for each entity (which customers?) and relationships between entities (which product to sell to which customer segment?).
Figure 2 Complexity of Analytical CRM
Our journey on analytical CRM needs a structured approach to address this complexity. Based on the vast experience of the PIO team both in the banking industry as well as in analytical solutions we suggest the approach visualized in Figure 3 which is discussed in the remaining paragraphs. SAP PIO covers each stage of this analytical CRM maturity cycle with a dedicated solution offering.
Figure 3 Recommended Approaches to Complexity of Analytical CRM
Segmentation
Knowledge about existing customers, their preferences and their behavior stored in the customer base (and transactional systems) is the competitive asset of a bank. The first logical step in analytical CRM would be to systematically analyze differences between customers with a goal to evaluate their current and potential future profitability and the emerging opportunities (and potential risks) for the bank. Detection of customer differences resulting in customer segments would then give rise to treat these segments differently in further steps of the analytical CRM maturity lifecycle.
Segmentation is a common practice in financial industry, but often these segmentations follow some predefined fixed structure based on demographics or a single measure of profitability (A, B, C segment or silver, gold, platinum). These predefined segmentations are easy to handle when preparing campaigns but experience shows that the usage of more accurate data-driven segments based on actual behavior typically generate at least double the conversion rate in campaigns compared to usage of traditional pre-fixed segments[3].
The purpose of customer segmentation is to place customers into groups with common needs and characteristics which best indicate those needs. In a commercial world the common needs are of course the customer preferences to buy which commercial products at what time. The characteristics to indicate those needs fall into several groups. Segmentation based on demographics would work if you believe that buying behavior depends primarily on customer attributes such as age, zip code, or marital status. Attitudinal segmentation could be used if buying would correlate well with attitudes obtained in customer surveys. Contrary to these approaches PIO would argue that the best indicator for buying are the day-to-day transactions across the line-of-business, a rich, always current and most relevant indicator for current and of course future buying preferences.
Retention
When existing customers are the primary source of value for a bank then defending these customers in loyalty programs should be a primary goal of analytical CRM. The goal of retention analysis is to predict which customers are likely to cancel a particular product or service under consideration within a foreseeable future (mostly the current year). Given estimates about typical net contributions this gives a good indication what value would be at stake if the cancellation of the product would materialize.
Retention builds on top of segmentation as profitability-based segments typically show different retention/attrition rates. The interaction of profitability and segmentation is of particular importance for marketing strategies and resulting marketing campaigns (see figure 4). Retention campaigns should concentrate on the low loyalty/high profitability segment. Retention analyses would require historical data on current and past customers to detect what would indicate a high likelihood of churning. Transactional data will give useful hints through decreasing amounts or frequencies of transactions when a risk materializes. Demographic data (change of address, change of profession) or external data (credit ratings) might point into the same direction.
Figure 4 Marketing Strategies for different Customer Segments
Cross-Selling/Up-Selling
Cross-Selling is defined as the action of selling an additional product or service to a current customer (same level, for example selling an insurance policy to a loan customer) whereas up-selling means the action of selling more expensive items or upgrades to an existing customer, for example selling stocks to an existing savings account customer. One third of the Best-in-Class performers already run cross- and up-selling initiatives whereas other survey respondents concentrate more on customer segmentation and customer acquisition (Aberdeen Group, 2009).
In addition to the data sources discussed for retention cross-selling/up-selling initiatives require information about product relationships (potentially differentiated by segment), in what sequence products were acquired, and finally how customers have reacted to different channels to optimize the means of communication. Cases studies indicate that more in-depth customer knowledge has resulted in cross-sell rates increasing by as much as 40%.
Customer Lifetime Value
The Customer Lifetime Value (CLV) is the discounted present value of all future revenues/costs attributed to a single customer relationship. It requires organizations to also calculate the cost per customer and thus gives a comprehensive picture of a customer value using previous knowledge from segmentation, retention and cross-selling initiatives about expected revenues.
In addition to data sources mentioned previously CLV initiatives should take advantage of the fact that long-lasting loyal customers are typically ready to communicate about their positive experience to their peers and relatives, especially when the positive experience was based on good customer service[4]. Likewise 79% of customers that had a negative experience with a company told others about it. With the recent proliferation of social media comments on companies have a huge impact and attempts should be made to measure the (hopefully positive) content as referral value. Long-lasting profitable customer relationships can also serve as a yardstick which new customers to acquire and what potential revenue could be expected. “Marketers can use CLV to determine exactly how much they should be spending to acquire new customers. This can influence marketing mix and channel decisions”.[5]
Campaign Portfolio Optimization
Analytical CRM initiatives have already quite a history and best performers use it routinely to set up their marketing campaigns and “productize” campaign execution. The efficiency of running campaigns has grown enormously in the last few years. Best Buy has progressed from 50 direct mail campaigns per month 4 years ago to now 800 campaigns per month in 2009. A global bank in Europe runs about 250 campaigns every three months.
Productization of campaigns creates its own challenges (compare Figure 2 above). Campaigns draw from the same resources (employees, call centers, electronic channels), customer contact policies need to be respected, individual campaign constraints (minimum or maximum of offers) and overall budget constraints need to be taken care of. Given cross-selling information, such as, how likely customers would react to offers and through what channel Campaign Portfolio Optimization (CPO) maximizes the economic value of a set of inter-related campaigns while covering constraints on customer contact policies, resources, prices, and budgets. Results of the optimization include maximum economic value usage of resources and optimal attribution of customers to offers ready for campaign execution. Early adopters are very positive on this technology. A smaller North American bank reports 6 million new leads in the first year, a European bank calculates an average increase in campaign ROI of 55%.
Summary
The Analytical CRM arsenal has the potential to create unique competitive advantage. Banks need to engage to discover the business value of their transactional data, clean it and set it up in a form ready for analysis. Data is meaningless without analysis and context. For the analytical CRM initiatives we advocate a sequence of transaction-based segmentation, retention, cross- and up-selling, customer lifetime value and finally campaign portfolio optimization. The goals of the initiatives differ, but each step lays the groundwork for the following in terms of additional data required for analysis, analytical maturity required and business benefits/competitive advantage achieved. SAP Performance, Insight and Optimization are ready to support banks at each step of this journey.