Data processing: financial – business practice – management – or co – Automated electrical financial or business practice or... – Discount or incentive
Reexamination Certificate
1999-03-10
2001-12-25
Gravini, Stephen (Department: 2162)
Data processing: financial, business practice, management, or co
Automated electrical financial or business practice or...
Discount or incentive
Reexamination Certificate
active
06334110
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to temporally defining virtual communities, and specifically to a system and method for analyzing customer transactions, such as purchases, in a commercial setting, as well as analyzing customer interactions, such as via browsing on the Internet
2. Description of the Prior Art
Presently, there exist systems that capture purchase transactions about customers. This information may be loaded into a scaleable data warehouse (SDW) for analysis by marketing personnel to determine product correlations, trends in sales, brand and color preferences, and trends in buying behavior based on demographics, psychographics, or geography. These systems may perform market basket analysis (products bought together at the same time), propensity analysis (buying product X predicts a subsequent purchase of product Y), or customer segmentation (defining a set of customers that buy or may buy a certain product or products).
These systems include the Knowledge Discovery Workbench available from NCR, the Management Discovery Tool available from NCR and the software available from Sterling Douglas which is used to create predictions of sales volumes.
SUMMARY OF THE INVENTION
The present invention addresses the need to temporally analyze customer transactions (i.e., purchases) and customer interactions (i.e., browsing on retail sites of the World Wide Web of the Internet). In other words, this invention analyzes customer behaviors based on the time when those behaviors occur. The invention captures information about customer interactions and transactions over time, classifies customers into one or more clusters based on their time-based interactions and transactions, or both, and uses this classification to perform selected target marketing or cross-selling.
Customer point of sale purchase information may be obtained by capturing point of sale scanner information. With the emergence of electronic commerce, it is also possible to collect customer purchase information and customer interaction behavior. In addition, the marketing analyst can examine differences in transaction behavior across in-store purchases and purchases on the Web (e.g., in-store and Web store “channels”). For example, some segments of customers may browse in the physical store and make a purchase through the Web store. Others may browse the Web store to gather product information, then only buy the product in the physical store. Some customer behavior may be product specific. For example, a customer may buy books through the Web, but purchase food items at a physical store.
The time when an item was purchased or browsed is important. Each purchase and browsed item may be tagged with the time it was bought or accessed. Temporal information on purchases or interactions may be captured when customers use scanners, Web browsers, kiosks, make calls to an inbound customer care center, or react to a call from an outbound customer care center. For Web browsers, by capturing keyclick information, it is possible to know what is browsed but not bought on the Web. This temporal browsing behavior, or interactions, resulting in no transaction is also useful to the marketing analyst because it reveals what customers considered but did not buy. Gathering the time an interaction took place may be useful in segmenting customers, because browsing time may be quite different from transaction time.
In-store temporal transaction information can be merged with Web temporal transaction information to provide a more complete report on customer behavior. Subsequent advertising or “segment-of-one” marketing campaigns can be built to target specific customers. With temporal profiling and temporal campaign management, marketing analysts can identify when people tend to be amenable to advertising. By building a profile of customer roles linked to time-of-day, day-of-week, and/or week-of-year, the marketing analyst can better predict when an advertising message might be most effective.
In summary, the following steps are performed. In step one, sources of information are temporally tagged by the customer “touchpoint”, which in the preferred embodiment may be a scanner, a kiosk, a customer care center, or a web browser. However, this invention may use other sources as well and includes any point at which a customer interfaces with a business. This information is kept in a SDW database such as the Teradata database available from NCR.
In step two, temporally tagged transaction or browsing information is analyzed to create temporal profiles. These profiles are created one per customer, and represent a complex time series object that captures the sequence of browsing and buying activity as well as the timestamp of each activity and the product being browsed or bought.
Individual temporal profiles can then be clustered using traditional data mining methods to identify groups, or segments of customers with similar browsing or buying behaviors at particular points in time. These clusters are called “virtual communities of interest”.
In step three, the marketing analyst then uses these segments to develop segment-specific advertising campaigns to appeal to these virtual communities of interest. Each temporal campaign involves an offer that will be made through one or more of the channels whenever the customer is identified through those channels and the profile indicates an opportunity for dynamic advertising. In this step, the marketing analyst decides which offers will be made to which communities of interest, through which channels, and when. This information is also kept in a database such as the Teradata database available from NCR.
Step four occurs when a customer who is a member of a particular virtual community of interest browses or buys at a specified time through a channel that can connect to the database. A check is made whether unique targeted advertising or offers are to be made to the customer. If so, the advertising information content is sent to the end user.
Step five involves the gathering of statistics over a period of time to determine the effectiveness of the advertising by channel. Temporal profiles that are successful are highlighted for reuse or further decomposition. Profiles which were not good predictors of subsequent buying behavior are discarded.
REFERENCES:
patent: 4649481 (1987-03-01), Takahashi
patent: 5305196 (1994-04-01), Deaton et al.
patent: 5426282 (1995-06-01), Humble
patent: 5481094 (1996-01-01), Suda
patent: 5537314 (1996-07-01), Kanter
patent: 5541835 (1996-07-01), Dextraze et al.
patent: 5634101 (1997-05-01), Blau
patent: 5642485 (1997-06-01), Deaton et al.
patent: 5687322 (1997-11-01), Deaton et al.
patent: 5933811 (1999-08-01), Angles et al.
patent: 5974396 (1999-10-01), Anderson et al.
patent: 6129274 (2000-10-01), Suzuki
patent: WO 98/56154 (1998-12-01), None
Smart Marketing by Jennifer deJong, Computerworld, Feb. 7, 1994.*
Database Marketing by Mike Carr, Marketing Intelligence & Planning, 1994.
Schrader David K.
Walter Joanne S.
Gravini Stephen
NCR Corporation
Needle & Rosenberg P.C.
LandOfFree
System and method for analyzing customer transactions and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for analyzing customer transactions and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for analyzing customer transactions and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2586666