Marketo Predictive content

Predictive ContentIntroduction Sept, 2016

Page 2Marketo Proprietary and Confidential | © Marketo, Inc. 05/02/2023

Overview

1. What is Predictive Content

2. How it works

3. Recommendation Algorithms

4. Case studies and real-life examples

5. How Marketo uses Predictive Content


Intro• Listening everywhere and communicating anywhere powered by

machine learning empowers marketers to maximize bottom line results leveraging marketing assets and audience data

• Automatically recommend the right content to the right prospect in a scalable way driven by data

• Self learning engine designed for rocket scientists marketers


Background• Marketers drive thousands of customers to their web site

• SMB’s: 10 - 50K visitors per month

• Enterprises: 100 - 500K visitors per month

• B2C’s: 100K - over 1 million visitors per month

• They also create hundreds of content assets to engage more effectively • Average SMB has +290 content assets

• Average Enterprise has thousands

• White papers, videos, case studies, blog posts, special offers


Engagement (R)evolution• Rule based targeting

• Great for small # of events, content products and verticals

• A/B testing• Further optimizing multiple engagement options for

segmented audience

• Predictive• A scalable way to auto-engage with the most relevant content

to each individual customer driven by data, powered by machine learning


Predictive Content

• Enables marketers to optimize engagements and conversions in

every online interaction

• Automatically presents the most relevant content for each

individual visitor

• Based on machine learning algorithms and predictive analytics


How it Works1. Auto-Discovers and maps all of your content assets

2. Learns which content works best and for who (training/testing)

3. Recommends relevant content to web visitors & leads

Increases content consumption and onsite engagement


• Auto-discover your content

• Analyze and track behavior to discover assets

DISCOVER LEARN RECOMMEND


Content Discovery


Content Discovery


• Record actual content consumption

• Analyze performance for different

audiences

• Run algorithms that learn what works

best for who


Machine Learning


Algorithm Types• Classification Algorithms (Logistic regression, Boosted decision Trees)

• Probabilistic algorithms Markov-chain

• Interest graphs


The models are tuned to..1. Detect successful and unsuccessful content engagements

2. Identify visitor/lead attributes in real-time

3. Discover positive correlations that could predict future engagements

4. Continuously optimize recommendations (every 8 – 24h)

5. Expose NEW content, detect trending and tune down old content


• Identify web visitor attributes

• Recommend most relevant content

• Optimize engagements



Example – Web Recommendation Bar


Example – Web Recommendation Bar


Example – Web Rich Media


Customer Example (Web Rich Media)


Customer Example• 75% direct lead conversion rate

from predictive content recommendation engine.

• Generated $100K in opportunities






Cross ChannelsWeb Mobile Email


Content Recommendation Results

315%Increase in site engagements (pages per visit, time on site) for targeted visitors

201%Increase in content specific lead conversion

3xMore views on recommended content compared to generic views

18%Increase in content consumption






1. Enable Content Discovery


2. Select Your Content







Advanced: use categories


3. Add Some Code

rtp('send','view');rtp('get','rcmd', 'richmedia');

<div class="RTP_RCMD2" data-rtp-template-id="template1"></div>

Add 2 lines to your standard RTP snippet:

Add a line of code where you want the recommendations to appear on your page:





A Measurable Lift in Downloads

New website launch

Recommendations added


Q4 – Predictive Content in Email

*not final



*not final



*not final



*not final



*not final


Predictive Content - Summary

• Enables marketers to optimize engagements and conversions

• Automatically recommends the most relevant content

for each individual visitor

• Based on machine learning algorithms and predictive

analytics

Thank you.

Marketo Predictive content

Marketing