IBM s Analytics Transformation

Transcription

1 IBM s Analytics Transformation Value Capture from Big Data, Analytics and Cognitive Technologies Martin Fleming VP, Chief Analytics Officer, and Chief Economist Chief Analytics Office

2 Analytics Aligned to Strategy Enabling Benefit Realization Innovation agenda aligned to strategy Plan Solve Agile Methodologies Quantification of business value Rapid cycle Deployment Measure Transform 2 Chief Analytics Office

3 Statistical Tracking and Assessment of Revenue (STAR) A STAR in the Forecast Advanced analytic model Revenue projection tool with +0.5% accuracy Transforms revenue assessment process Based on patented algorithms IBM Patent Disclosure Dynamic Ensemble Modeling for Revenue Forecasting 3 Chief Analytics Office

4 STAR Forecasts All Revenue Types Transaction Revenue: Logistic regression with factor engineering Stream revenue: Assessed using time series and regression models for seven revenue types Exponential smoothing has ability to give preference to recent data and quickly adapt to changes in the data: {x t } raw data, {s t } output of exponential smoothing S 0 = X 0 St = CxX t + (1 a) S t 1, t > 0 Where Cx is the smoothing factor, and 0 < cx < Prob = exp ( B 0 + B i x i ) / ( 1 + exp ( B 0 + B i x i ) ) S t = cxx t + (1 cx)s t 1 = cxx t + cx(1 cx)x t 1 + (1 cx) 2 S t 2 = cx [X t + (1 cx)x t 1 + (1 cx) 2 X t (1 cx) 3 X t (1 cx) t 1 X 1 ] + (1 cx) t x 0 4 Chief Analytics Office

5 STAR Scales Across the Enterprise Analytic Engine Information Warehouse Delivery Vehicle SPSS SPSS is used to analyze a wealth of data to project outcomes and generate actionable insights Cognos TM1 Cube is the STAR repository and enables drill-down, roll-up, slicing and dicing capabilities Cognos Mapped Worksheets reporting capabilities create a cross-ibm delivery vehicle 5 Chief Analytics Office

6 Workforce Analytics Drive Performance Talent Acquisition Learning Perf. Mgmt. Workforce Planning Engagement Retention HR Operations Analytics and optimization engine to support end to end HR transformation Workforce Planning Talent Acquisition Talent Management HR Operations 6 Chief Analytics Office

7 Analytics is the Link Across Data, Insight, and Value Corporate DBs Collaboration Solutions IBM Workforce Analytics Seller incentive in Geo X are Y% below market average Social Media Give incentive X to employee Y Systems of Record Public Labor Data Big-Data Cloud end-to-end Workforce Solution Hire X number of employees in Geo Y for product Z Move employees A..B to division X and train them in Y 7 Chief Analytics Office

8 Big Data Sets Foundation Clients Which clients sellers need to sell to? Offering Which offerings sellers need to choose? ACCENT* is the foundational Big Data tool build over time enabling a number of client-focused applications * Analytics on the Customers and Coverage for the Enterprise (ACCENT) 8 Chief Analytics Office

9 Sales Coverage Recommendation Engine 1 Multiple data sources 2 Client-level sales effort and profit analysis 3 Reports combining client profit with other data 4 Sales deployment actions Client ID Sales Exp Sales Activity Data Revenue & Cost data Client Opportunity Revenue & Growth 9 Chief Analytics Office

10 Optimize Routes-to-Market Field Sales Business Partners Digital Sales Digital Initiative Marketing Design sales coverage with optimal Improve effectiveness of lead passing to Partners Increase use of Digital Sales resources with territories to reach new clients & markets Accelerate the transformation to new buyers, markets & offerings Recommend optimal Marketing spending by tactic business unit, geography and other segments Business challenge : Increase revenue from new buyers reached through alternate routes Solution: Predict and target growth opportunities Deliver recommendations prioritizing routes and optimizing Publish regular route-to-market revenue and expense performance 10 Chief Analytics Office

11 Routes-To-Market Reporting Drives Sales Transformation Top Down Route-to-Market Model Methodology Inputs Data Inputs Possible Routes to Market Route Revenue Matrix Opportunity Values Revenue Opportunity Alignment Align opportunity with revenue Legend FF Fulfillment Route Selection Select likely Routes to Market OI OO DI DS BP FS Opportunity Identification Opportunity Ownership Digital Initiative Digital Sales Business Partner Field Sales 11 Chief Analytics Office

12 Watson Cognitive Computing Capability Discovery Create new insights and new value Protein P53 Decision Provide bias-free advice semi-autonomously 25,000 test case scenarios and 1,500 real-life cases Engagement Interact and assist through understanding and reasoning of both content and context Transitioning to civilian life 12 Chief Analytics Office

13 Three Fundamental Building Blocks Cognitive computing defines the end user experience of interacting with an intelligent system Data ingestion* Machine learning (ML) / reasoning** Digital - human interface*** Unstructured data (text, images, audio) Natural language processing Image recognition Speech recognition Structured data (can be used as input of ML model) Decision trees / rule-based models Simple & transparent Regression Models Bayesian models Latent factor models Neural nets / Deep learning Complex & intransparent New ways of interaction with computing systems Ask questions in written language to gain insights (Google) Talk to the system (Siri, Amazon Echo) Convert the system output into spoken language 13 Chief Analytics Office

14 Offering and Solution Recommendation Engine Offering Recommendation API Industry Solution Recommendation API Business Rules Event Triggers Digital & F2F Industry Solution Model Ensemble Selected Models Client Feedback Social Media Linked and Integrated Datasets Internal and External Data Sources 14 Chief Analytics Office

15 15 Chief Analytics Office