Corporate Initiative Response IntelliQuote

Digital Transformation of Print Media Sales

Raman Agarwal, Strategy and Pricing

30th January 2019 Acknowledgements

Project Sponsor Concept & Design Tableau Integration IT Implementation

Sivakumar Sundaram Raman Agarwal Rakesh Baboota Paresh Thakur

Devanshu Verma Vikrant Desai

System Architecture Analytics Vivek Kumar Sachin Kale

Raman Agarwal Raman Agarwal Adarsh Saikia Rajesh Rebello

Prashant Dogra Neeraj Chhabra Rahul Jha

Paresh Thakur Yuvraj Relan Nitendra Singh

Vikrant Desai Sunit Chandan Lalit Anand

Siddharth Dubey Babusha Sharma Consultations Nitesh Kumar Rajat Srivastava Subramanian S Kriti Taneja Deepa Bharti Parag Bhangle Amit Kumar

1 Reference and Inspirations

# Topics Source

1 Pricing Guidance in Ad Sale Negotiations Paper by Google

2 Comparison of Accuracy of Statistical Models Paper by Dr. Hirotogu Akaike

3 Revenue Management and Dynamic Pricing Lectures of Prof. Goutam Dutta, IIM Ahmedabad

4 Advertising Industry and Trends KPMG Report 2018

5 Implementing dashboards in Tableau YouTube and online forums

6 Negotiation Theory Wikipedia links on ZOPA and BATNA

2 Foreword / Executive Summary

Bennett Coleman & Company Limited (BCCL) conceptualised, developed and architected a dynamic pricing application – IntelliQuote - which predicts client specific prices in real time for a print inventory.

The algorithm is designed in such a manner that it triangulates prices from two different logics and once a sale is made the machine learns the propensity of the advertiser to pay for the subsequent transaction. The state-of-the-art design is first of its kind in the industry and works with an accuracy of 75%, i.e. majority of its predicted prices have potential to turn into orders.

Apart from predicting the “right price”, the application increases speed of business; monitors negotiation efficiency of the sales team; aids in preventing price-arbitrage; and provides Tableau exportable reports for further R&D and trend analysis.

The application is accessible through a Mobile App or Desktop and is used by over 1500 sales personnel with over 1 Lakh transactions till date, leading to a digital transformation of the sales team and enabling them with quote and floor prices on the go.

The robustness of the analytical engine behind the application is audited and certified by two consulting firms independently.

3 Table of Content

Situation, Complication, Solution 1 The problem at hand?; Plan to address it

Significant Achievements 2 Increase Yield, Revenue; Digital Transformation

IntelliQuote – Under-the-hood 3 Universe of variables; Broad Algorithm; Decision Support System

Appendix 4 Project Timelines; Application Screenshots Table of Content

Situation, Complication, Solution 1 The problem at hand?; Plan to address it

Significant Achievements 2 Increase Yield, Revenue; Digital Transformation

IntelliQuote – Under-the-hood 3 Universe of variables; Broad Algorithm; Decision Support System

Appendix 4 Project Timelines; Application Screenshots Situation: The legacy pricing and revenue management system relied on ad-hoc decisions substituted for business intelligence

Elements of a Dynamic Pricing System

Analytics Decision Support System

Real-Time Trend Transparent Sales Data Cogent Pricing Analysis (MIS) Models Negotiation

Set Pricing Market Prices Strategy Data

Real-Time Analysis Calibration Competitive Market Advantage Intelligence or VOC

Business Strategy + Sales Targets

Existing Partially Existing Missing 6 Complication: The pricing model, negotiation process and route to market had critical drawbacks

Inaccurate Slower Absence Model with Speed of of a Live Issue obnoxiously Issue Business Issue Decision high rates Support 01 03 05 System

Price Inconsistency Issue Lack of Transparency Generic and Issue Archaic MIS for unpalatable 02 business intelligence, rates from an 04 often unavailable for Sales team not outdated system Pricing decisions enabled to take Manual transactions on-the-spot depending on the acumen of Pricing decisions No tracking of a human - Pricing Scientist negotiation efficiency; Risk of team’s churn 7 Solution: A new eco-system is designed with a vision to increase Yield, Drive Productivity and Empower People

Project Intersect A confluence of sales and pricing

Mobile Application – IntelliQuote (iQ)

Pricing Team

Decision Accuracy

Sales Team Transparency

Support Support System

Consistency and

Speed Speed ofBusiness Change ManagementChange

Standard Multipliers Auto Approvals Tracking Discounts Automated & Revised ORs Dashboards Pre-filled Forms Re-visit Triangulate Forecast Regression & Analytics Statistics / Math Rev, Yield, MS Simpler Navigation De-risk Sales Churn Alternate Models Improved UI Integration with CRM Occupancy Third Party Audit Auto Rate Grids Buying Pattern WIP Completed 8 Table of Content

Situation, Complication, Solution 1 The problem at hand?; Plan to address it

Significant Achievements 2 Increase Yield, Revenue; Digital Transformation

IntelliQuote – Under-the-hood 3 Universe of variables; Broad Algorithm; Decision Support System

Appendix 4 Project Timelines; Application Screenshots IntelliQuote is able to increase the Yield by 5%, with an estimated annual revenue benefit of USD 10mn

4.1% Percentage Premium Achieved Over Usual Yield 100% 3.4% 20%

90%

80% G1 G2

70% Area under the curve represents G1 = Average Print Market Growth* incremental revenue 60% (FY 2017 to FY 2018)

50% G2 = Weighted Average Yield Increase (Transactions through IntelliQuote) 40%

% premium charged per transaction transaction per charged premium % 30% Revenue Impact** = ~USD 10 mn

20% (Annual impact if volume is same and full adoption by the sales team) 10%

0% *KPMG India Study 2018 Number of Transactions ** Check appendix for scenario analysis

10 Much more than an App, the IntelliQuote is a digital transformation of the sales team

Request for Quote Machine Rates

1 2 Sales Pricing Team Auto Emails with Quotes Rates curated by SPOCs

4 Near Real-time Model Improvement

5 3

IT

For Regional and National Heads For President’s Office / Directors . No-offline rate to be maintained . Negotiation efficiency of Sales & Pricing Team by Markets, Verticals . De-risk team’s churn 6 . Analytics to roll-out potential Spot Offers (Wish vs Closure) . Triangulation of Forecast Trends Automated Reports 11 In contrast to the legacy system, the IntelliQuote prices are well within the reach for sales to pick business

From

To

12 IntelliQuote shifts the computation load to a machine thereby facilitating richer engagements between the Pricing and the Sales teams

Manual recursive Pricing Supervised Machine Automated Pricing Learning

Archaic way of working Provides opportunity to control rates Idealistic Subject to errors Best of both worlds Need supreme accuracy

1 2 3 13 IntelliQuote fortifies the competitive edge of the Times Group by delivering productivity and sales empowerment

Strategic priorities defined by the leadership to grow Print business in India to be carried till 2020 14 Table of Content

Situation, Complication, Solution 1 The problem at hand?; Plan to address it

Significant Achievements 2 Increase Yield, Revenue; Digital Transformation

IntelliQuote – Under-the-hood 3 Universe of variables; Broad Algorithm; Decision Support System

Appendix 4 Project Timelines; Application Screenshots The legacy system used limited variables for pricing; The rest were considered for ad-hoc decisions

Market Share Share of Wallet Consumption Pattern Opportunity Cost Client Growth Rate Paper Quality Edition Month Industry Growth Rate Innovation Color / B&W Page Position Special Day/Event Premium Circulation / 1 Occupancy 2 3 Ad Size Print Order Deal-Size Launch vs Frequency Affinity to Print Importance for Brand No. of insertions Costs Existing vs New Client Non-Print Bundles

Competition Rates Industry Sector Entry Barrier vs Defend

Traditional Variables Potential Alliance 1 Agency Relationship 2 Variables used in IntelliQuote Government Regulations 3 Contextual Variables

16 IntelliQuote expanded the purview - relevant variables are programmed; Dashboards built for significant ones

Significant variables for which Share of Wallet Market Share dashboards were made Consumption Pattern Opportunity Cost Client Growth Rate Paper Quality Edition Month Industry Growth Rate Innovation Color / B&W Page Position Special Day/Event Premium Circulation / Occupancy 2 3 Ad Size Print Order Deal-Size Launch vs Frequency Affinity to Print Importance for Brand No. of insertions Costs Existing vs New Client Non-Print Bundles

Competition Rates Industry Sector Entry Barrier vs Defend

Traditional Variables Potential Alliance 1 Agency Relationship 2 Variables used in IntelliQuote Government Regulations 3 Contextual Variables

17 IntelliQuote’s analytical engine predicts client level prices with >75% accuracy

Accuracy is defined as predicted prices are about 30% premium on usual yield for the combination Data, Design and Results audited by two consulting firms 18 IntelliQuote exhibits stronger prediction of rates than ever before; Coaxing higher adoption by the sales team

Resolves Issue 1

Source: BW / SAP 19 IntelliQuote automates 66% of the incoming enquiries; increases productivity and speed of business

Resolves Card Rate Issue 2 & 3 10k RFQs Auto Approvals A Price from Legacy System Automation 15k RFQs Approvals by Sales Heads B achieved through the accuracy of IntelliQuote Price price prediction

51% or 28k RFQs Repeat Orders C by IntelliQuote Operating Yield

55k RFQs

70k RFQs 70k 85% of Revenue of85% 49% or 27k RFQs Manually Negotiate D

Fixed + Variable Costs

Variable Costs A + B + C 53k/80k

Note: Estimates are annual and are taken from the MIS on % of business approved below floor 66% Automation Achieved 20 IntelliQuote lends transparency in the system as all stages of ZOPA* are captured to show leakage points, if any

Resolves “Let The Truth Prevail” Issue 4

Confidential Confidential

Confidential

ZOPA*

iQ

Confidential

* ZOPA – Zone of Possible Agreement 21 IntelliQuote, frees-up the Pricing Team to study business context using client-level dashboards

Resolves From Here To Here Issue 5

Wrong Price Right Price Right Value Proposition to Advertisers using Business Intelligence

Note: Under Project Intersect, Real-time Tableau Dashboards are designed for decision making. For confidentiality reasons they are not shown here

22 Table of Content

Situation, Complication, Solution 1 The problem at hand?; Plan to address it

Significant Achievements 2 Increase Yield, Revenue; Digital Transformation

IntelliQuote – Under-the-hood 3 Universe of variables; Broad Algorithm; Decision Support System

Appendix 4 Project Timelines; Application Screenshots Project Charter

Appendix IntelliQuote - Project timelines and charter Several modules were built, working in close quarters with Product, Sales and IT Team

Decision Analytics Change Management Tableau Training Decision Support System

Re-visit regression and data-cleaning Discussions with Leadership Shadow Launch with Train Sales Team for Pilot Draft New Process Develop Models Pricing Team Newsletters and Infuse system for auto approvals Success External Test Fixtures Announcements Benchmarking Create Calibration Cell Pricing KRAs for Adoption Develop Tableau Scratches Training Collaterals Automate Rate Grids

IT Implementation Non-Display Integrate Authorizations Calibration Models Pricing Simulator Build new GUI Execute logics CRM Integration Index Utility Build MRW / App

Design Test Simulator Special Rate Flag Update GUI after VOC Innovations Cost Integration

External Audit - Data Validation and Math

Design Architecture Implementation, Testing, VOC Pilot / Beta Go-Live

Jan 2017 Jul 2017 Jan 2018 Mar 2018 Jun 2018 Dec 2018

25 IntelliQuote – Project plan for architecting the model and processes Assess integration requirements, check analytics models and verify the approach for business measurement

W1 W2 W3 W4 W5 W6

• Data clean up • Data Gap Analysis • Outlier capping Data Standardization • Missing data analysis and development • Data Imputation • Data Standardization

• Understand data distribution • Creating distribution diagrams to understand the Model ready distribution of variables datasets • Creating descriptive statistics to understand the variable relationship • Creating Analytics records datasets

• Multicollinearity check • Model preparation Modeling • Modelling commencement Initial Model • Model preliminary testing Creation Finalized • Iterate to get better Accuracy and precision Model

• Testing the models on sample data • Business validation Model Testing and Iteration • Sharing results with top management

Validation of selected Steering Committee Implement with inventories by business Project Plan from January to July 2017 with President IT 26 Estimated Revenue Impact will be in the range of USD 2mn to USD 15 mn Calculated considering no volume impact with price increase

USD 160 mn (Effective Business on which iQ will be able to impact)

S1: Optimistic S2: Realistic S3: Pessimistic

$ 15.0mn $ 7.0mn $ 2.0mn

Assumptions Effective Volume 80% 60% 40% (Compliance) 10% 6% 4% Price Inc. on Premium Inventory 4% 2% 0% Price Inc. on Inferior Inventory 27 Application Screenshots

Appendix IntelliQuote is a mobile ready application to facilitate requests for quotes on-the-go 1/3

Response IntelliQuote

Copyrights Reserved with BCCL, a Times Group Company 29 IntelliQuote is a mobile ready application to facilitate requests for quotes on-the-go 2/3

Copyrights Reserved with BCCL, a Times Group Company 30 IntelliQuote is a mobile ready application to facilitate requests for quotes on-the-go 3/3

Confidential Confidential

Copyrights Reserved with BCCL, a Times Group Company 31 IntelliQuote offers a user friendly dashboard for the sales team

Quick and easy access of prices by industry / client / publication / position Confidential

Copyrights Reserved with BCCL, a Times Group Company 32 The Pricing Team can search the application and check parallel negotiations and avoid instances of price arbitrage

Download excel for the below search

3 RFQ raised for individual Editions till date

No Packages for the above inventory

One combined RFQ raised for the above inventory

Copyrights Reserved with BCCL, a Times Group Company 33 Live Calibration using Tableau

Appendix IntelliQuote is integrated with Tableau to facilitate continuous calibration by the Pricing Team 1/2

Histogram of errors divided into three parts

Machine vs Average Operating Prices

This part is the Negative This part is the Positive skew skew of the bell curve. of the bell curve. These are all the cases which These are the cases where are below the Average the machine had generated Operating Rate rates above 40% of the average operating rate

Area of concern is to make sure the machine throws rates above the average operating rates The average operating rate

Source: BW. September and October 2018 35 IntelliQuote is integrated with Tableau to facilitate continuous calibration by the Pricing Team 2/2

Machine Prices vs Average Operating Prices

Human Prices vs Average Operating Prices

The data is analysed by 14 individuals, each responsible for an industry vertical

Source: BW. September and October 2018 36 End of Document