The Data-Hub Journey at Erie Insurance Brian Novacek, Erie Insurance

Derek Laufenberg, MarkLogic

© COPYRIGHT13 June 2017MARKLOGIC CORPORATION. ALL RIGHTS RESERVED. Operational Dataflow Challenges Who are my customers? ▪ What services might they need? How do I keep them? ▪ What are the business risks? ▪ How valuable is…?

SLIDE: 2 © COPYRIGHT 2017 MARKLOGIC CORPORATION. ALL RIGHTS RESERVED. Today’s Data Management Challenges

DATA INTEGRATION DATA SECURITY MANAGEABILITY

% OF THE % % OF 60 COST 50 INCREASE 72 BUDGETS Of data warehouse projects is In security budgets, while IT Devoted to just “keep-the- on ETL budgets overall remain flat lights-on” activities

SLIDE: 3 © COPYRIGHT 2017 MARKLOGIC CORPORATION. ALL RIGHTS RESERVED. Erie Insurance

• Fortune 500 Company located in Erie Pennsylvania • A+ (Superior) rated by A.M. Best Company • Over 5,000 employees and 12,000 independent agents • 10th largest homeowner insurer in United States • 12th largest automobile insurer in United States • 15th largest property/casualty insurer in United States What is our mission?

To provide our Policyholders with as near perfect protection, as near perfect service, as is humanly possible, and to do so at the lowest possible cost. Erie’s Challenges

• Application Data Silos • Shadow IT Systems • Complex Workflow & Integration Processes • Data Mastering Challenges • Aging Mainframe Applications & Hardware – Need a 360 view, or golden record Erie’s Business Need

• High Speed and Open/Extensible Delivery Platform • Cross Lines Integration & Data Mastering • Data Integrity with Security Erie’s Technical Goals

• Improve Application Delivery Times • Utilize Agile Processes • One Simple Goal • Turn data into information as fast as possible • Golden record with 360 degree view • Researched Platforms and Options Erie’s Search

SLIDE: 9 © COPYRIGHT 2017 MARKLOGIC CORPORATION. ALL RIGHTS RESERVED. Erie’s Search

• NoSQL & Document üFlexibility üScalability • Needed Security, Transactions, & BI • Support Enterprise Operations

Learned of MarkLogic and Data-Hub Pattern Billing A NEW ENTERPRISE INTEGRATION PATTERN Fixing the Picture Claims Multiple § Traditional integration technologies are CRMs people intensive § Copying the same data to support different needs is the norm because of rigid models Agency 3rd Party Data § A flexible, scalable operational is the foundation for a Data Centric approach Policy Ratings & Risk § Greater convergence between analysis and Quotes operations is critically important § The new Enterprise Pattern is the Operational Data Hub

SLIDE: 11 © COPYRIGHT 2017 MARKLOGIC CORPORATION. ALL RIGHTS RESERVED. Operational Data Hub (ODH) Definition 1.2 An authoritative data repository for cross-functional operations that harmonizes line-of-business data into a canonical forms on an as-needed basis. It serves as a multi-subject, multi-model, contextual, real-time, integrated and data-centric enterprise interchange in support of enterprise operations and analysis/discovery throughout the data lifecycle.

SLIDE: 12 © COPYRIGHT 2017 MARKLOGIC CORPORATION. ALL RIGHTS RESERVED. MarkLogic ODH Pattern

STAGING FINAL (RAW DATA AS IS) (HARMONIZED, INDEXED DATA)

SOURCE 1 ENVELOPED DOCS RDBMS DOCUMENTS OPERATIONAL (ENTITY 1) APPS

f(x)

SOURCE 2 ENVELOPED DOCS SERVE MESSAGE INGEST DOCUMENTS (ENTITY 2) ANALYTICAL BUS APPS HARMONIZE INDEX, SEARCH, & SERVICES

INDEX, SEARCH, SOURCE N DISCOVERY, & ENVELOPED DOCS CONTENT DOCUMENTS HARMONIZATION (ENTITY N) DOWNSTREAM FEED SYSTEMS

SLIDE: 13 © COPYRIGHT 2016 MARKLOGIC CORPORATION. ALL RIGHTS RESERVED.

POC Overview

Objective Scope Prove whether or not document • “System of Record” use case data modeling and storage can • Day 1 ‘Add & Maintain improve flexibility and accelerate Individual Entity Data’ application development. • Compare/contrast application Understand impacts when development with document model changes occur. models POC Summary Relative Applied Time Document Relational

Document POC

Relational DB Why MarkLogic

• Enterprise grade – – ACID Transactions – Backup – point in time restore – High Availability and Disaster Recovery – Security • Flexible data model – easy to work with – fast search • and real-time alerting • Approximate 4:1 improvement overall application delivery Three Pillars for Success

Technology Process People Technology Change…

• Short story – it works! • Longer Story – – Data Hub Pattern – Scalable, Enterprise – … you get the picture Process Change…

• Make your process more agile • ETL becomes ELT • Data model as you go – Load sources iteratively – Harmonize as applications require – No “big bang” • Communicate – repeatedly People Change…

• Socialize the Technology – Data Hub Pattern – Search & discovery – Incremental data modeling • Expanding with the right projects – Match project to team’s skill level – Keep focused • Leverage Your MarkLogic Sales Engineers – Email / WebEx Sessions – Onsite Office Hours Becoming Self-Sufficient

• Select projects that match team’s skill level • Build a Center of Excellence – Open minded architects and developers – Skilled in multiple API, languages – Track record of embracing new technology • Leverage MarkLogic Consulting to augment your domain experts • MarkLogic University Erie Today…

• We are building a Data-Hub using MarkLogic • We are doing it the MarkLogic way • Focused on new & improving processes • Leveraging acceleration while learning • Delivering Faster Q & A