DV 2016: Defeating the Multi-Channel Silo Challenge

© 2016 Tealium Inc. All rights reserved. | 1

Defeating the Multi-Channel Silo Challenge

Matt Wright, Hewlett-Packard Enterprise


Matthew WrightHewlett-Packard Enterprises - Analytics Strategy

• Internal Analytics & Intelligence Consultant with HPe• Work with different teams to establish digital measurement

strategy, including customer intelligence and online behavioral research, measurement strategy, and analytics operations strategy

• Intimately familiar with digital data operations – which includes TMS & Audience identification solutions

Headshot Here


What is a Silo, Anyway?


How do Data Silos Form in Organizations?

Systems Follow Org Structures• Data is typically used by "Users"• Users are defined by

organizational structures with different measures of success

• Use cases reflect user requirements

• The more concentrated the use by certain user groups, the more likely the data is modified / shaped / organized to reflect those use cases

RESULT:• Data that is largely setup for one

use• Difficult to allocate to other uses

Positive ViewCorporate "Scarcity" Model• Every organization is subject to

scarcity and limited resources• Use of data by other teams

constitutes • At best a consumption of a

direct resource, • At worse, a loss of

opportunity• Control the data and you control

the story

RESULT:• Data is 'hoarded' and 'protected'• Data is only shared within

groups that have non-competing interests

Negative View


What Does a Data Silo Look Like?• Different KPIs of an organization

may lead to vastly different approaches to typical execution

• Underlying data that drives the KPIs tends to reflect what makes the KPIs look best

• Data is protected as a resource to tell a story

Examples• Marketing and eCommerce• Finance and Operations• UX Team and Everyone Else• Cats and Dogs

Marketing

eCommerce


Impact of Data SilosCollaboration Doesn't

HappenData is ActionedSub-Optimally

Missed Opportunities


Rise of the "Federated" ArchitectureFederated Topologies• Complex architectures are extremely

hard to manage• Approach is not to create single-systems

with everything, but to join together independent systems into a cohesive meta-system

• Underlying technology shares key information between solutions while maintaining control

RESULT:• Greater agility and flexibility, because

you're not designing to one standard, but many

• Lower risk and cost-containment, because you're not re-designing the world

• Data can flow between disparate silos

Systems


AudienceStreamFederates

Disparate Data


AudienceStream: Federated Customer Intelligence Federated Topologies

• Complex architectures are extremely hard to manage

• Approach is not to create single-systems with everything, but to join together independent systems into a cohesive meta-system

• Underlying technology shares key information between solutions while maintaining control

RESULT:

• Greater agility and flexibility, because you're not designing to one standard, but many

• Lower risk and cost-containment, because you're not re-designing the world

• Data can flow between disparate silos


Case Study: Aligning Offline & Online Data Sales

Engagement

MRM

MRM

Digital Data Offline Account Data


Federation: Our ConceptKey issue with Federated Data Design is Exchange

• Same problem of "common database keys", but amplified by disparate solutions, APIs, and political overhead

• Key design needed to be flexible, but resilient enough to be actionable up-front

• Building from Scratch: very difficult and expensive


Federation: From Concept to Implementation

Connecting Data Became Easier


Once the Key Systems are Setup, Actions on Customer Data are Second Nature

Actioning the Data Became Second Nature


What Doesn't AudienceStream Solve?Non-Customer Centric Approaches• Collaborative Filters based on product

or asset level decisioning• Cross-sell by product (not

customer)• Asset recommendation

• Provided you can map the decisioning back to a customer identifier, you can still make it work

Political Alignment• Focus on win/win outcomes• Less threatening to focus on connected /

collaborative pilots than full-blow projects

• Open framework lowers the risk of losing control of one's data

• Action & Success Sells


Key Takeaways Data that is largely specialized or

protected as a means of controlling outcomes between competing groups

The larger the company, the more likely you'll run into data silos

Data Silos Are Real Federation of data takes disparate

systems and aligns them through a series of consistent technologies and processes

Aligning the data into a 'meta' store of data solutions rather than the 'big-bang' one data store to rule them all

Federation Solves Many Silo Issues

Keying structure between systems has to be correct and customer-centric

Actioning becomes second nature for audience definitions

AudienceStream Breaks Down Data Silos


AUDIENCE Q&A


Thank [email protected]

@mattjwright

DV 2016: Defeating the Multi-Channel Silo Challenge

Technology

Transcript of DV 2016: Defeating the Multi-Channel Silo Challenge