FOCUS ON: CDPs SUBSCRIPTION ACCELERATOR INMA S UMMIT W ORKSHOP , NYC - - PowerPoint PPT Presentation

FOCUS ON: CDPs

SUBSCRIPTION ACCELERATOR

INMA SUMMIT WORKSHOP, NYC FEB 26 2020 STEVE LOK

CO-FOUNDER & CHIEF MARKETING TECHNOLOGIST, M MATTERS FORMER GLOBAL HEAD OF MARTECH, THE ECONOMIST & FRESHFIELDS https://www.linkedin.com/company/mmatt ers steve@mmauthority.com https://www.linkedin.com/company/mmatters

YOUR CUSTOMERS ARE NOT JUST DATA

TOO MANY SIGNALS

NOT ENOUGH MEANING

WHY I’M HERE TODAY

 We started M Matters to help senior executives

bridge the gap between marketing, product, technology, and data into useable strategies at their organizations

24 Years in Tech & Dev

1996 - 2009

Telecom Healthcare IT Education Publishing

10 Years in Digital Media

2009 – 2019

The Economist Scholastic Hachette Freshfields

7 years in MARTECH

2013 – now

M Matters Freshfields The Economist STEVE LOK ALESSANDRO MORETTI

30 mins – Presentation on CDPs as subscription accelerators

 The Problem Today  What’s the CDP and why now?  Exclusive Use Cases: Millennials & Dynamic Paywalls

10 mins – Raising your challenges

 Mitigating revenue loss from the disruption in media businesses?  Changes in the way your customers perceive you due to privacy?  Who is trying to personalize? Who has already done enough to be satisfied?  Do you believe in data sciences?

20 mins – Q&A – I’m an open book!

WHAT WE’LL BE DOING TODAY

CUSTOMER DATA POWERS THE PREDICTIVE ECOSYSTEM: GIVING CONTEXT TO DATA RELATIONSHIPS

Set C Set A Set B Set D Set E Set F Set *

MULTIPLE SETS OF SIGNALS

Set C Set A Set B Set D Set E Set F Set * Aggregated Normalised Owned

CDPs: AGGREGATE & LEARN & DERIVE

DELIVER (EVER MORE) RELEVANT EXPERIENCES

Complexity

• f Intent

EYE KNO YOU

Data Science = Competitiveness

Some CDPs Do Behavioral Scoring

• Using behavioral data from multiple data sources often

create a paradox of choice in terms of field selection

• Behavioral scores provide a multidimensional

description of user behavior

• Scores help to identify behaviorally unique subsets of

users (binge, peruser, etc.) Blog Post: https://www.getlytics.com/blog/post/look_at_lytics_predictive

Content Processing: Affinity

• Automatic: Topic tagging is automatic
• Acts like a Human: NLP and computer vision allow us to

read/view content exactly like a human does, so it can’t be abused like SEO

• Behaviorally Driven: Users are telling you exactly what

they like. (Not “users who like X also like Y”)

NEED: MACHINE LEARNING

• Supervised machine learning

algorithms to derive & predict

• Likelihood to perform an action or

make a change individually or in quorum

• Self-training models use

anonymous data in user and account profiles

Three KPIs we impact with Data Science

Efficiency – suppress people you are wasting money on Prediction – spend more money on people who are ready to buy

1

Discovery – find new behavior patterns which lead people to buy

2 3

THE CDP

• Why now?
• What is it?
• What it isn’t
• Why care?
• Opportunities
• Typical Implementation
• Typical Costs
• Quick wins

WIN

WHY CDP NOW?

Journey Customer Knowledge Location Behavior Patterns Product Relationships Preferences Channel Consistency Processes and UI

MARKETING MOMENTS One to One One to Few One to Many Foundation

RELEVANCY

WHAT IS IT?

A CDP can help us discover and mobilise audiences, across multiple products & businesses. So let’s focus on just the 3 main capabilities, and why they matter.

WHAT IS IT?

#1 A CDP builds 1st party profiles over time using multiple qualified data sources

WHAT IS IT?

#2 A CDP lets you group profiles into meaningful audiences

WHAT IS IT?

#3 A CDP activates these audiences with marcomms

WHAT IT ISN’T

• Not a DMP – information without identity
• Not BI/DW – information without activation
• Not CRM –

information without behavior

• Not Analytics – information without individuality
• Not a ‘file’ –

information without real-time A CDP combines them all: individual behavior + mapped data + RT machine learning = Self-optimizing activated audiences

GREAT – BUT WHY IS THIS IMPORTANT?

YOU GAIN INSIGHTS — AS YOUR CUSTOMERS DEVELOP BEHAVIORS

OPPORTUNITIES

• Content Marketing – More relevancy = more engagement
• Loyalty – Increase relevancy with individual customer journeys
• Value Exchange – Differentiate revenue opportunities across your

existing products

• Data/Insights – Immediately reach the audiences you want, with

limitless connectors to bring data in and ship it out

Decisioning & Machine learning Activation

Advertising Digital Properties Campaigns Analytics/Testing

Cloud Apps Data Warehouse & BI And more... And more... Predictive Models Machine Learning Natural Language Behavioural Scoring Content Affinities

Data Integrations 360° Profile ID Resolution Segmentation Orchestration

Real-time

• r Batch

Real-time

• r Batch

CDP

CDP ECOSYSTEM OVERVIEW

• Pre-built
• Real-time

OOTB

• Data Collection
• Personalisation
• Segmentation

Catalog

• Content

Management

APIs

• Pass data to CDP

via Pixel / URL click

Pixel

• Collect data
• Personalise
• Implement using

Google Tag Manager or other

JS Tag

• Hourly, Daily or

Weekly

• Your SFTP
• Our SFTP
• CSV or JSON

SFTP

JS

CDP INTEGRATION ECOSYSTEM

Cookie: 123456, Email address: abc@123.com Email address: abc@123.com, CRMID: UID123 Email: abc@123.com Loyalty no: ABC12345 Order/Booking: ABCDEF, Loyalty no: ABC12345

Website CRM ESP Loyalty Order/Booking

Userid UID123, Email: abc@123.com, Facebook ID: FB019238

Authentication

DYNAMIC IDENTITY RESOLUTION

DATA BLUEPRINT DATA UNIFICATION AUDIENCE ACTIVATIONS MEASUREMENT

• Scope

implementation phases and create project plan

• Build CDP

implementation to blueprint specifications

• Develop identity

resolution strategy across data sources

• Import and process

data imports to CDP

• Authorize

integrations

• Create centralized

audiences for activation

• Export audiences in

real-time to connected marketing tools

• Measure impact of

improved targeting efforts

• Evaluate process

efficiencies and impact on business

TYPICAL CDP IMPLEMENTATION

TYPICAL COSTS

CDPs: Monthly users (active) x number of events Pricing can vary widely! And is always open to negotiation ☺ Nominally:

USE CASES IN MEDIA & PUBLISHING

• Content – Dynamic comms with personalized content
• Retention & Upsells – Triggered, individual marketing comms on live

behavior – create targeted narratives, tell stories

• VX – Smart CTAs for every unique type of customer
• Audience Mining – Sell high value customer audiences on behavior

CASE STUDY w/CDP: “MILLENNIALS” CAMPAIGN

GOAL: INCREASE AWARENESS AND SUBSCRIPTION ADOPTION WITHIN YOUNGER DEMOGRAPHIC

> $500k salary Single

STEP 1: MODEL CDP AUDIENCES

+ Email: ending with .edu + Data enrichment: IP Addresses of US/UK University + Title: Student, Intern, or Grad Student + Predictive Modeling: High propensity to become premium subscriber

• Exclude: Current subscribers (and low propensity prospects)

(not channel-first, or ad platform-first)

> $500k salary Single

STEP 2: FIND INDIVIDUALS ACROSS PLATFORMS

True Multichannel Campaign

• Facebook
• Snapchat
• Google ads

It works: 35% of ad-driven subscriptions by millennials

> $500k salary Single

STEP 3: ON-SITE PERSONALIZATION

Hold Up!

Free Book!

This worked too:

• 9% Click through rate
• On-site conversion rates

for targeted audiences? 5-10x higher

TL;DR: SMARTER, MODELED AUDIENCES WORK ROAS

[ [

BAU Segments: $2.89 Lytics Segment: $3.95

EFFICIENCY LEVEL UP! 37% Improvement on Return On Ad Spend ROI? YOU BETCHA. < 1 month payback on CDP investment

COST PER ACQUISITION (CPA) WINS: GO LOW

Data-agnostic Structure-agnostic Source-agnostic

• 1st Party
• 2nd Party
• 3rd Party
• MORE EFFICIENT PPC:

43% BETTER ROAS

• CONVERSION RATE

IMPROVEMENTS: NEAR 200%

• SCALE:

13% OF SEARCH CONVERSIONS

[ [

Business As Usual Audiences $35 Lytics-Powered Audiences $20

CPA

• 1. Data Science is a commoditized utility – use it
• 2. Behavioral Orchestration – from Ads and Web all the

way to TV and Outdoor, to paper and post direct mail

• 3. Shifting to audiences of ONE and away from channel-

based measurement, optimization and planning

• 4. Only have a CRM? Good luck!

2020? BEHAVIOR IS THE NEW BLACK

BEHAVIOR IS THE NEW BLACK

CDP

Dynamic Content Paywall:

A Reality in Just One Quarter

A UNIQUE USE CASE FOR PUBLISHING

• Static Paywalls
• Risks & Opportunities
• Leveraging Martech
• Propensity & Paywall
• Quick Wins - Capabilities Now
• Creating Future Opportunities

USE CASE INGREDIENTS:

Reader behaviour is quite different, but because the paywall is static:

• It’s relatively easy to avoid
• Can make people leave earlier than necessary
• Can discourage discovery
• Treats a loyal reader exactly the same as a cold

prospect

Nearly 3/4 of The Economist’s non-paying weekly visitors never saw it

ONE SIZE DOES NOT FIT ALL

Paywall Hits = Content consumed Paywall Hits = Ad impressions Paywall Hits = Subscriptions

= Inability to separate risks from benefits optimisation

RISKS OF A UNIVERSALLY STRICT WALL

“Hard” Paywall

• Ignores the individual experience
• Tweaks the same old metering

levers

• Discourages content discovery
• Blindly blocks our best content

from potential subscribers

• Blocks cross-team opportunity

Dynamic Paywall

• Fully Customer-centric
• Opportunities for different types
• f value exchange
• Encourages content consumption
• Match your best prospects +

their favorite content

• Cross-functional revenue opps

DYNAMIC PAYWALL BENEFITS

Ultra Low Subscription Propensity (ULSP)

05

• Trigger more ad units on page
• Experiment with new ad products

Low Propensity to Convert

04

• Push reg and other value-driven experiences

Regular/Unknown

03

• Paywall appears as usual – 3-5 articles

High Propensity to Convert

02

• Paywall to appear after 2 articles consumed

Ultra High Subscription Propensity (UHSP)

01

• Paywall is immediate, with customized

messaging based on their affinity/interest

Propensity to Subscribe

PROPENSITY & PAYWALLS

WHAT IF…

Drive acquisition + retention revenue!

• Increase conversion - Stricter paywall for Ultra High Subscription Propensity (UHSP) Lytics

segment only

• Increase free user LTV - lower paywall to encourage natural “influencer” ambassadorship
• Reduce churn - Personalise and surface relevant content per individual

Drive advertising revenue!

• Increase ad impressions - More ad units for Ultra Low Subscription Propensity (ULSP)

Lytics segment - “always free” visitors

• Ask ad blocking users to whitelist us

KPI IMPACTS

• Do you have the martech? Can you:
• Understand your content contextually? (AI or CDP)
• Understand and profile individual reading behaviour? (CDP)
• Push modeled audiences to your channels? (CDP)
• Organisational silos hide the bigger picture
• Unable to bring the best of martech, adtech, data insight leadership

together

CHALLENGES

Custom web/mobile Experiences $$$$$$$$$$$$$

THE RECIPE

(At The Economist):

Propensity models Customer Databases Content affinity Content Sources

• Number of visits
• Operating system
• Device
• Content being consumed
• Location
• Time
• Channel

Behavior & Propensity w/Lytics machine learning

CDPS PROVIDE CUSTOMER SIGNALS

Set up propensity audiences in Lytics Push Lytics audiences to Piano Test & Launch Setup new paywall rules in Piano

HOW QUICK IS QUICK?

Drive acquisition revenue:

• Solve the mobile conversion problem - different subs experience

Drive down churn:

• Offer unique site experiences based on subscription channel and method
• Dynamic “warm up” campaigns

Drive advertising revenue:

• Offer a new low-cost ad-free product

SUSTAIN PRODUCT INNOVATION

IT’S SIMPLE ENOUGH TO BAKE: DYNAMIC = SIGNALS + CAPABILITIES

THE CAKE IS NOT A LIE

THANKS FOR LISTENING!

Q&A: CDPs

SUBSCRIPTION ACCELERATOR: USE CASES, BENEFITS, COSTS

INMA SUMMIT WORKSHOP, NYC FEB 26 2020 STEVE LOK

CO-FOUNDER & CHIEF MARKETING TECHNOLOGIST, M MATTERS FORMER GLOBAL HEAD OF MARTECH, THE ECONOMIST & FRESHFIELDS steve@mmauthority.com https://www.linkedin.com/company/mmatters