Multi-source consumer behavior tracking system

• US 20120054213A1
• Filed: 09/01/2010
• Published: 03/01/2012
• Est. Priority Date: 09/01/2010
• Status: Active Grant

• Current Assignee: Acoustic, L.P.
• Original Assignee: International Business Machines Corporation
• Inventors: Puttaswamy, Hemanth, Woodruff, Travis, Ogireddy, Srinivasa, Batra, Nipun, Howard, Steve
• Granted Patent: US 8,768,943 B2
• Time in Patent Office: Days
• US Class Current: 707/758
• CPC Class Codes: G06F 16/903 Querying for retrieval from the web G06F16/953
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data

Abstract

A process for tracking a consumer'"'"'s behavior based on his or her entries into a number of cybernetic device sources such as personal computers, automated goods and services dispensing kiosks, automated teller machines and cell phones does not depend exclusively upon the presence of cookies implanted into such devices by vendors or upon a consumer'"'"'s having provided a positive identification form. The process uses a pairing algorithm to comb elements of information from the source device visitation records in search of one or more common characteristics that can be presumptively attributed to a single consumer. Characteristics such as source device identifiers, phone numbers, street addresses or email addresses are each weighted by a weighting factor that is used to identify the consumer and attribute the visitation record and its associated activity to that consumer at a calculated confidence level. The attributed information can analyzed to track that consumer'"'"'s behavior.

Claims

1. A process for associating unattributed elements of information collected from a plurality of cybernetic sources with a plurality of consumers, said process comprising:

creating a master registration containing at least one characteristic attributed to a first consumer;

collecting a plurality of visitation records containing elements of information non-exclusively attributable to said first consumer;

detecting in said elements a value matching a portion said at least one characteristic;

said value originating in a first of said records; and

,

attributing said first of said records to said first consumer.

2. The process of claim 1, wherein said records are collected through a plurality of disparate marketing channels.

3. The process of claim 1, which further comprises assigning a record confidence level to said first of said records.

4. The process of claim 3, which further comprises:

each of said records comprising activity data related to a cybernetic visit;

a plurality of said records being attributed to a plurality of master registrations; and

,

running an analysis of said activity data across said plurality of registrations and said plurality of records.

5. The process of claim 4, wherein said running comprises selecting a minimum confidence level;

and screening said plurality of said record confidence levels against said minimum confidence level.

6. The process of claim 4, which further comprises assigning a master registration confidence level to said first master registration.

7. The process of claim 6, which further comprises adjusting said master registration confidence level according to a record confidence level said first of said records attributed to said first master registration.

8. The process of claim 6, which further comprises updating said record confidence level according to at least one weighting factor determined by the context in which said first of said records was created.

9. The process of claim 6, which further comprises updating said record confidence level according to a number of matching elements found within said first of said records.

10. The process of claim 9, wherein said updating comprises assigning an intrinsic confidence level to each of a plurality of element types common to said records.

11. The process of claim 10, which further comprises changing said intrinsic confidence level according a context adjustment weighting factor.

12. The process of claim 1, which further comprises linking said first master registration with a source identifier code associated with said first of said records.

13. The process of claim 12, which further comprises subsequent detecting of a second value in a second of said records matching said source identifier code.

14. The process of claim 1, which further comprises said portion of said characteristic being an entirety of said characteristic.

15. The process of claim 1, wherein said value is of a first type and said at least one characteristic is of a second type, wherein said first and second types are different.

16. The process of claim 1, which further comprises:

flagging a number of elements associated with said value;

determining whether said number pf elements has an associated confidence level which allows said value to be used as a future characteristic in said master registration; and

,

updating said master registration to have a new characteristic containing said value.

17. The process of claim 1, wherein said first master registration remain un-associated with an identified one of said consumers.

18. The process of claim 1, which further comprises assigning a weighting factor to said characteristic.

19. A process for associating anonymous strings of information collected from a plurality of cybernetic sources with identified consumers, said process comprising:

detecting in said strings characteristic values non-exclusively attributable to a consumer;

flagging strings containing a plurality of said characteristics;

combining a plurality of said flagged strings into batches as a function of similarities in said characteristics; and

attributing each of said batches and related strings of information to a single one of said consumers as a function of information in said batches.

20. The process of claim 19, wherein said attributing comprises using a pairing routine.

21. A process for associating strings of information collected from a plurality of cybernetic sources with a subject, said process comprising:

extracting from each of said strings at least one characteristic non-exclusively attributable to a subject;

combining a plurality of said strings into batches as a function of similarities in their respective characteristics;

attributing each of said batches to a subject; and

identifying said subject through information contained in said attributed batches.

22. The process of claim 21, wherein said combining comprises scanning said strings to extract at least one common characteristic.

23. The process of claim 22, wherein said scanning comprises using a pairing routine.

24. The process of claim 22, wherein said extracting comprises assigning weighting factors to a plurality of said characteristics.