Using big data to predict organizational commitment
Download as PPT, PDF1 like377 views
This document describes a predictive model for organizational commitment using big data. It models the organization as a network of commitment agents and uses a Markov chain to represent transitions between states. Statistical techniques like design of experiments and probability distributions are used. Live data is captured and leading indicators from the model can help ensure higher commitment levels across the organization. There is potential to enhance it to evaluate dynamic business scenarios and power a real-time management dashboard.
Using big data to predict organizational commitment
- 1. Using Big Data to Predict Organizational Commitment Rajiv B. Deo B.Tech. M.Tech. C.I.S.A.
- 2. Š Rajiv B Deo 20152 References ďŹ W. Tantisiriroj, S. Patil, G. Gibson. âData-intensive file systems for Internet services: A rose by any other name ...â Technical Report CMUPDL-08-114, Parallel Data Laboratory, Carnegie Mellon University, Pittsburgh, PA, October 2008 ďŹ Ramesh R. Sarukkai, LinkPrediction and Path Analysis Using Markov Chains, Yahoo Inc, 2000 ďŹ Stefan Wegenkittl, Modeling with Markov chains http://crypto.mat.sbg.ac.at/~ste/diss/node1.html, May 1998 ďŹ M.C. Paulk et al., eds., The Capability Maturity Model: Guidelines forImproving the Software Process, Addison Wesley Longman, Reading, Mass., 1995 ďŹ Billy E. Gillett, Introduction to Operations Research, TMH Edition 1979 ďŹ Frederick S. Hillier and Gerald J. Lieberman, Introduction to Operations Research, Holden-Day Inc 1973 ďŹ W.Feller, Introduction to Probability Theory & Itâs applications, Vol 1 & 2, Wiley, 1971
- 3. Š Rajiv B Deo 20153 ďŹ The success of any business depends heavily on meeting the customer requirements in terms of quality, cost and functionality on or before the agreed dead line. ďŹ This is achieved by ensuring a very high level of commitment among all the sub groups activities ďŹ âcommitmentâ is nothing but keeping the promises made with each interface represented by a sub group Organizational Commitment - I
- 4. Š Rajiv B Deo 20154 Organizational Commitment - II ďŹ Quality model based on Capability Maturity concept developed by SEI is built on the management commitment and involvement at each stage for meeting goals of every KPA ďŹ Thus, there is a need to continuously monitor and predict organization commitment level in every organization aiming at âOptimizingâ level of SEI CMM. Over a period this level needs to improve. ďŹ In this presentation, we shall take a brief look at a quantitative model conceptualized and developed by the author to predict level of organization wide commitment.
- 5. Š Rajiv B Deo 20155 Organization commitment model Request for commitment is made by the service user agent Scope of the request is frozen & Risk + Impact analysis is done by the service provider agent Commitment given to the user agent - the date of expected fulfillment Commitments Commitments tracked to closure Organization commitment model described here is best represented by a network diagram (PERT chart) where each arm in the critical path is represented by a two party commitment transaction involving a user agent and a service agent. The actual process between user and service agents is described below:-
- 6. Š Rajiv B Deo 20156 Commitment Agents & theirInterfaces Senior Management Business Development Software Delivery Infra-structure Quality & Software Engineering SQ Audit performance Team Performance - Schedule - Costs Duration of team meetings # Proposals # Orders # Customers # New Customers Audit Performance SLA performance - Installation - Problem solving - H/W S/W Purchase # Processes introduced / modified / improved - Process Compliance Index Customers Project Management Training Management Resource Management Human Resource Management
- 7. Š Rajiv B Deo 20157 NetworkDiagramforpredicting Organizational Commitment Level C U S T O M E R Business Development Group Software Delivery Group Project Management Group Quality Group Human Resources Group Resource Management Group Training Group Infra-structure Group
- 8. Š Rajiv B Deo 20158 Statistical Techniques - 1 ďŹ Design of experiments technique is used to identify unique independent factors which influence the predictability of the commitment transaction. ďŹ Delivery, Quality, and Cost of each project depends on commitment from - ďŹ Senior Management ďŹ Quality Management ďŹ Project Management ďŹ Training Management ďŹ Resource Management ďŹ Help Desk & Infrastructure Management ďŹ Hardware & Technology Procurement Management ďŹ Human Resource Management ďŹ Business Development
- 9. Š Rajiv B Deo 20159 Statistical Techniques - 2 ďŹ The predictive model for organizational commitment level is dependent on the current state and is completely independent of the previous states of the system. ďŹ The Organizational Commitment Model as seen by the customer is represented as a first order, finite state Markov chain consisting of two channels viz. ďŹ Main channel ďŹMarketing - Pre-sales â Project Management â Implementation ďŹ Supporting channel ďŹResources, Training, HR, Quality, Senior Management, Infrastructure
- 10. Š Rajiv B Deo 201510 Statistical Techniques - 3 Transition probability from stage I to stage J is worked out using pij(s) = P(X(t+s) = j | X(t) = i) where, X is the Markov property derived from the performance of respective commitment agents on the critical path of the organizational network diagram.
- 11. Š Rajiv B Deo 201511 Statistical Techniques - 4 Organization Commitment level is predicted with a certain level of confidence from a stochastic process consisting of collection of OCi{i = 1,2, âŚ. n} where in, each OCi has a specific probability distribution function.
- 12. 12 Degrees of Freedomâ Commitment Agents SM QM PM TM RM IT HP HR BD â 7 7 7 6 6 5 4 6 2 SM 5 0 1 1 0 0 0 1 1 1 QM 5 1 0 1 1 1 0 0 1 0 PM 7 1 1 0 1 1 1 1 1 0 TM 5 1 1 1 0 1 1 0 0 0 RM 7 1 1 1 1 0 1 1 1 0 IT 6 0 1 1 1 1 0 1 1 0 HP 4 1 0 1 0 1 1 0 0 0 HR 7 1 1 1 1 1 1 0 0 1 BD 4 1 1 0 1 0 0 0 1 0
- 13. 13 Commitment Agents Table part I Agent Description Mechanism Interfaces Degrees of Freedom Weight age (Wagent) SM Senior Management Senior Management decision making and reviews Customers, Software Delivery, Quality, Infrastructure, Business Development, Human Resources 12 0.2857 QM Quality Management Audit Performance Software Delivery, Hardware & Technology procurement, Help Desk Support, Business Development, Project Management, Resource Management, Training Management, Human Resource Management 12 0.0714 PM Project Management Project Planning, review, and Tracking Software Delivery, Resource Management, Training Management 14 0.1429 TM Training Management Inter Group Coordination Project Management, Resource Management, Infrastructure, Quality, Human Resource Management 11 0.0714
- 14. 14 Commitment Agents Table part II Agent Description Mechanism Interfaces Degrees of Freedom Weight age (Wagent) RM Resource Management Inter Group Coordination Human Resources, Project Management 13 0.0714 IT Help Desk & Infrastructure Management SLA Performance Project Management 11 0.0714 HP Hardware & Technology Procurement Management SLA Performance Project Management 8 0.0714 HR Human Resource Management SLA Performance Software Delivery, Quality, Infrastructure 13 0.1429 BD Business Development Business Targets Software Delivery 6 0.0714
- 15. Š Rajiv B Deo 201515 Using Prediction Model in practice - I ďŹ To find out what would be the organizational commitment level in the month of August, you would look at the predicted value of OC8 of 9 service providers mentioned in the network diagram. ďŹ Organization commitment level for August 2016 would be 1. OC8 = WSM*OC8SM + WQM*OC8QM + WPM*OC8PM 2. OC8 = OC8 + WTM*OC8TM + WRM* OC8RM 3. OC8 = OC8 + WITOC8IT + WRD* OC8BD 4. OC8 = OC8 + WHP*OC8HP + WHR* OC8HR
- 16. Š Rajiv B Deo 201516 Using Prediction Model in practice - II ďŹ After the organization commitment level for a month is predicted using the stochastic process model, we test the hypothesis that the organizational commitment would be at the predicted value with 95% level of confidence using Chi square test. If the test fails we repeat the exercise for a lower level of confidence till the test succeeds. ďŹ The predicted service provider componentâs level from the model is used by the concerned service providers to give realistic commitments, there by ensuring better predictability and greater customer satisfaction.
- 17. Š Rajiv B Deo 201517 Summary ďŹThe predictive model defined here, was implemented using Hadoop with R and beta tested at many global organizations from 2008 to 2014 ďŹLive raw data captured from Unicenter, Remedy, SAP modules. ďŹThe leading indicators from the model have ensured higher levels of organizational commitment
- 18. Š Rajiv B Deo 201518 Scope forfurtherwork ďŹ The predictive data analytical model can evaluate dynamic business scenarios including organization re-structuring as the degrees of freedom between the service agents change resulting in a different levels of organization commitment. ďŹ A real time Management Dashboard driven by business simulation of different organizational strategies can boost the organization wide commitment level as seen by the customer.