Browse Prior Art Database

Inactive sales opportunity detection and sales manipulation behavior and clique discovery

IP.com Disclosure Number: IPCOM000234132D
Publication Date: 2014-Jan-14
Document File: 4 page(s) / 93K

Publishing Venue

The IP.com Prior Art Database

Abstract

Sales activity would generate massive sales opportunity in the sales pipeline, though some of them are promising, while still are inactive and zombie opportunities that would otherwise waste sales team's effort if they are not cleaned from the pipeline. This disclosure proposes a method and system for detecting the inactive low quality sales opportunities from the pipeline database using a semi-supervised learning approach. In addition, detecting sales mainpulation , which intentionally underestimate or overestimate the quality rating of a given opportunity is also presented.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 4

Inactive sales opportunity detection and sales manipulation behavior and clique discovery

Effective managing sales and sales opportunity is key to business performance. Each day, many new sales opportunities are identified and recorded in the pipeline by sales, and many old opportunities become closed (win/lose). It is common that some sales are pressured to input some garbage/false opportunities (we name it inactive) to fake: they are on good track. Some opportunities are recorded and they become pending for long time. Many (100+) well structured information associated with sales opportunity is recorded in the pipeline database. This can be good indicator to evaluate the quality and reliability of the recorded samples. But such patterns are unknown and may change continuously. Thus it is an uergent demand to develop an automatic garbage sales opportunity detection system that can also help identify manipulation by sales.

Key differentiation

Key differentiation:

:: How to generate the sample list for sales to label

                How to generate the sample list for sales to label?
1 Sales are reluctant to label his own pending opportunity as garbage, so avoid letting sales to label his own cases


2 It is still possible that sales may form some interest cliques, in the garbage labeling process, they may mutually support each other when they are given the cases from opposite owned. Sales A may rate good quality to sales B's and vice versa. But in real, the opps. are no good.


3 In the above both cases, the labeling quality is poor and misleading. While these two cases are prone to happen in real business environment.


4 Our solution:


4.1 in each iteration, leveraging the trained model to detect: a) labeling manipulation; b) sales cliques (cross-over-labeling, cross-under-labeling etc)


4.2 and take actions: a) adjust the weight of the sales labeling who are suspect for manipulation; b) adjust the samples for sales to label who are suspect for manipulation

System Diagram:

1



Page 02 of 4

Key procedure flow

2



Page 03 of 4

Details

Details: :

3



Page 04 of 4

1 Initialization: The samples for sales to label are from their background-relevant but not self-owning opportunities


2 A detector is trained and is used to give the garbage confidence score [0,1] about the samples that have been labeled by sales.


3 Compare the predictiv...