Case study
Deutsche Bank finds value in its customer data
Deutsche Bank is one of the largest financial institutions in Europe, with assets under management worth 100 billion euros (£60 billion). It operates in seven different countries under different names, although the company is considering consolidating into a single brand operating as a pan-European bank.
In 1999 its chairman, Dr Walther, said the company had to improve its cost-to-revenue ratio to 70 per cent from 90 per cent and add 10 million customers over the next four to eight years. That would be achieved by increasing revenues through growing customer value by cross and up-selling, reducing costs through more targeted communications, and by getting more new customers based on meaningful data analysis.
Central to this programme has been the introduction of an enterprise-wide database, analysis and campaign management system called DataSmart. This has brought significant changes to its marketing processes and effectiveness. Achieving this new IT infrastructure has been no mean feat Deutsche Bank has 73 million customers, of which 800,000 are on-line and 190,000 use its on-line brokerage service, it has 19,300 employees, 1,250 branches and 250 financial centres, plus three call centres supporting Deutsche Bank 24, its telebanking service. It also has e-commerce alliances with Yahoo!, e-Bay and AOL.
‘DataSmart works on four levels providing a technical infrastructure across the enterprise, consolidating data, allowing effective data analyses and segmentations, and managing multi-channel marketing campaigns', says Jens Fruehling, head of the marketing database automation project, Deutsche Bank 24.
DEFINE AND IMPLEMENT INFORMATION QUALITY POLICY
The new database runs on the largest Sun server in Europe with 20 processors, 10 Gb of RAM and 5 terabytes of data storage. It is also mirrored. The software used comprises Oracle for the database, Prime Response for campaign management, SAS for data mining, Cognos for OLAP reporting, plus a data extraction, transformation, modelling and loading tool. ‘Before DataSmart, we had a problem of how to get data from our operating systems where it was held in a variety of different ways and was designed only for use as transactional data. There are 400 million data sets created every month. We had a data warehouse which was good, but was not right for campaign management or data mining', says Fruehling.
The new data environment was developed to facilitate all of those things. It also brings in external data such as Experian's Mosaic. ‘We have less information on new prospects, so we bought third party data on every household the type of house, the number of householders, status, risk, lifestyle data, financial status, age, plus GIS coding', he says.
For every customer, over 1,000 fields of data are now held. These allow the bank to understand customers' product needs, profile, risk, loyalty, revenue and lifetime value. That required a very sophisticated system. For every customer, there is also a whole bundle of statistical models, such as affinity for a product and channel, profitability overall and by type of product.
Questions
1 What were the reasons for this information management initiative?
2 How have DB changed their approach to information management? How has it changed the quality of their information?
3 Assess the possible risks associated with this approach.
‘These are calculated monthly so we can perform time-series analyses, so if their profitability is falling, we can target a mailing to them', says Fruehling. DataSmart has allowed Deutsche Bank to make some important changes in its marketing process, allowing it to operate more quickly and effectively. ‘We have a sales support system called BTV in our branches to communicate with each bank manager. They can see the customer data and are able to add information, such as lists of customers who should be part of a branch campaign, who to include or exclude, and response analyses,' he says.
Previously, typical marketing support activity involved segmenting and selecting customers, sending these lists through BTV for veto by branch managers, making the final selection, then sending those lists to BTV and the lettershop for production.
‘There were many disconnects in that process we had no campaign history, nothing was automated. Our programmers had to write SAS code for every selection, which is not the best way to work. We had no event-driven campaigns', says Fruehling. An interface has been developed between PrimeVantage, BTV and each system supporting the seven key channels to market. Now the database marketing unit simply selects a template for one of its output channels. This has allowed Deutsche Bank to become more targeted in its marketing activities, and also faster.
‘Regular selections are very important because local branches do our campaigns. We may have up to 20 separate mailings per week for different channels. That is now much more profitable', says Fruehling. Customer surveys are a central part of the bank's measurement culture and these have also become much easier to run.
‘Every month we run a customer opinion poll on a sample of 10,000. Every customer is surveyed twice in a year. That takes half a day to run, whereas previously it took one week and 30 people using SAS. If a customer responds, their name is then suppressed, if they do not, they are called by the call centre', he says. The bank's customer acquisition programme, called AKM, now uses up to 30 mailings per year with as many as 12 different target groups and very complex selection criteria. ‘We flag customers using SAS and PrimeVantage recognises those flags', he says. ‘We are now looking to move to a higher communications frequency so every customer gets a relevant offer.'
Source: European Centre for Customer Strategies www.eccs.uk.com/advice/case_studies/CS_Implementation/case5.asp.