Consider the following scenario: A customer is in the process of buying a shirt and is at a checkout stand paying with her credit card. The clerk swipes the card, and the register displays a message that says this particular customer’s favorite designer jeans are on sale in aisle 10 – and her size is in stock. Because she is an excellent customer, she is entitled to another 10 percent discount. The clerk informs her of the pant sale, and given the extra-good deal, the customer opts to buy the jeans as well.
Scenario number two: An irate wireless service customer phones a call center and complains that he was dropped “for the third time this week” at a particular location. The call center rep confirms from the customer’s call records that this was indeed the case. The rep also notices the customer is using an outdated phone and determines that such a phone does not allow the user to take advantage of a new digital system. The rep then offers to replace the old phone for an upgrade fee, waive the new phone activation charges and offers to mail the customer a preconfigured new phone that same day. The customer gladly accepts.
In the first scenario, technology helps turn an already happy customer into a happier one; and in the second, an irate customer’s complaint becomes an opportunity not only to re-build customer satisfaction, but also to up-sell a new product.
Both situations are good examples of effective customer relationship management (CRM). And neither would be possible without an active or, as it is also called, real-time data warehousing system.
Promises to Keep
According to an August report released by the Conference Board, a survey of 96 multinational companies showed that 52 percent of the surveyed companies have implemented a CRM product or package. CRM solutions are clearly on the rise, but only one in four respondents characterized their implementation as “very successful.”
Scores of other studies also support the disappointing results of CRM in real-world installations. It’s hardly a secret why. Although CRM-related technology has existed for some time, integrating this technology into legacy systems and company culture has been far from easy. Data – no matter how critical – can do very little good for customers, or anyone else, as long as it stays snugly inside the warehouse.
But with the advent of real-time data warehousing (RTDW), for the first time, CRM solutions that are truly customer centric are not only possible, but also practical. With new advancements in technology, RTDW systems can now make operation and transaction data easily accessible to customers, personnel, suppliers and other targets, without hindering operation and transaction database performance.
New Market, New Paradigm
Given the growing need for real-time data warehousing and its resulting host of benefits, what should an ideal RTDW system look like? In an effort to approach real time, many data warehouse experts have favored an increasingly frequent extract, transfer and load (ETL) approach. Updating data warehouses in such a manner is used to create data marts and OLAP cubes.
Emerging technology today, however, allows for a completely new – and I believe better – approach. Replacing the ETL approach with a method based on replication, minimal transform, and continuous and smaller load (Rtl), this new type of RTDW can become the sole source of data and information not only for CRM purposes, but also for a number of applications, such as data mining, business intelligence and reporting.
The ideal RTDW should:
- Be structured so that data schemas, fields and the data itself would be familiar to most of the enterprise. Data schemas (structures) should be similar to existing operation and transaction systems. Certain fields (columns) could be excluded, and some records could be aggregated or retained “as is” with aggregations stored in separate tables.
- Accept high rates of record (row) inserts and updates, and simultaneously allow complex queries by a large number of users.
- Allow for access of the data by and at all levels in a company, instead of being limited to high-level executive or campaign summaries.
- Minimize or avoid the need for specialized data marts.
- Consist of archived or completely static, normally static, and live data and information.
- Be capable of fast location and time-series queries.
- Return result sets that could be used “as is” in the operational and transaction systems.
- Incorporate unstructured (text, e-mail, comments) and semi-structured (XML, spreadsheet) data and information.
- Include binary large object (BLOB) files, such as images, video and audio.
The advantage of such a paradigm for real-time data warehousing is the system’s ability to enable utilities and applications typically used with data warehouses, data marts, OLAP cubes and other traditional approaches to be created without crippling system response. This is no small feat because, in the past, the driving force behind the creation of high-level data warehouses, data marts and OLAP cubes was the need to overcome database performance limitations and features.
Such limitations are no longer the problem they were. There are new breeds of database technologies that are becoming increasingly capable of supporting the demands of the most complex utilities and applications. The ability to implement successful CRM programs is perhaps the best possible illustration of why these new database solutions will become increasingly critical in the future.
With the growing emphasis on satisfying the individualcustomer, the rise of real-time data warehousing is inevitable. To better serve, cross-sell and up-sell the customer, pertinent data must travel across multiple channels – and it must do so at speeds that approach real time.
