Content Structure Workshop - Content Marketing World 2019

Content Structure: Your On-Ramp to AI
Laura Creekmore
Director of Taxonomy and Content Strategy, Syndigo
@lauracreekmore
@lauracreekmore • #CMWorld

SLIDES HERE:
https://slideshare.net/lauracreekmore

PHOTO FROM
DTS
@lauracreekmore #cmworld

THE REAL
SECRET–
CONTENT
STRUCTURE

AUTOMAGICALLY
My favorite internet word

• Intros
• Overview of content structure and advanced technology
• Exercise: Creating taxonomy with card sorting
• Designing content for advanced tech
• Exercise: Creating use cases for voice and chat
• Special characteristics of voice and chat
• Exercise: Creating structure for voice and chat
• Next steps

INTROS
Who are you, and what are you hoping for?

OVERVIEW
Content structure and today’s advanced technology

CHALLENGES FOR
MARKETERS
➤Extremely cluttered
landscape
➤Burdensome requirements
➤Endless content variations
➤Ineffective and expensive to
adapt for every channel
➤Channels not aligned

TRADITIONAL
SOLUTION

IT DOESN’T MEET YOUR
CUSTOMER’S NEEDS@lauracreekmore #cmworld

GET OUT OF THE
CAMPAIGN BUSINESS…
GET INTO THE
RELATIONSHIP BUSINESS

REBUILD YOUR MINDSET

WHAT
PROBLEM IS
MY CUSTOMER
TRYING TO
SOLVE?

HOW CAN I
SUPPORT
THEM?

STANDARD
CONTENT WORKFLOW
AUDIENCE STRATEGY OPERATIONS ANALYSIS GOVERNANCE
Identify
customers,
prospects,
people
Identify their
goals and
needs
Identify your
goals and find
synergy
Identify tools,
channels you
can use
Design
systems that
meet your
goals +
audience goals
Use systems to
create and
publish
content
Evaluate
success toward
goals
Iterate
Determine
ongoing
decisions to be
made
Define
framework for
decisions
Use
framework

ADVANCED
CONTENT WORKFLOW
AUDIENCE STRATEGY OPERATIONS ANALYSIS GOVERNANCE
Identify
customers,
prospects,
people
Identify their
goals and
needs
Identify your
goals and find
synergy
Identify tools,
channels you
can use
Design
systems that
meet your
goals +
audience goals
Use systems to
create and
publish
content
Evaluate
success toward
goals
Iterate
Determine
ongoing
decisions to be
made
Define
framework for
decisions
Use
framework

THE DIFFERENCE IS IN
THE FILLING

ADVANCED
CONTENT OPERATIONS
DIGITAL CHATBOTS VOICE SKILLS AI-DRIVEN ECOMMERCE
Topics
Formats
Write
Publish
Define
information
sets to support
conversations
Build
conversation
triggers
Test, rewrite!
Test, rewrite!
Determine
interactions
useful in voice
environment
Design
content for
common voice
interactions
Test! Test!
Test! Test!
Promote
Determine
where data
and needs
intersect
Jumpstart
with
taxonomy/
ontology/
semantic
content
Train! Train!
Train! Train!
Document
buyer’s
journey
Determine
content need
Structure to
create/store/
deliver content
Design
customer
experience

ADVANCED
CONTENT OPERATIONS
DIGITAL CHATBOTS VOICE SKILLS AI-DRIVEN ECOMMERCE
Topics
Formats
Write
Publish
Define
information
sets to support
conversations
Build
conversation
triggers
Test, rewrite!
Test, rewrite!
Determine
interactions
useful in voice
environment
Design
content for
common voice
interactions
Test! Test!
Test! Test!
Promote
Determine
where data
and needs
intersect
Jumpstart
with
taxonomy/
ontology/
semantic
content
Train! Train!
Train! Train!
Document
buyer’s
journey
Determine
content need
Structure to
create/store/
deliver content
Design
customer
experience
IT’S ABOUT THE
STRUCTURE

MISCONCEPTIONS
➤Throw it all in there; Google
will sort it out.
➤I don’t need to structure it.
➤I need to structure
everything.
➤My customer will see what I
want them to see.
➤I’m starting from scratch.

TRUTHS
➤Critical to start with
customer perspective.
➤You can’t script everything.
➤Creating the domain and
content model pays off.
➤Metadata for the win.
➤You should test some more.

WHY DOES IA MATTER
FOR AI?
You shortcut the pattern recognition.

TERMINOLOGY
INTERLUDE

DOMAIN MODEL
What it is: An organized way to describe the topics and
information related to your [business, industry, area of interest]
How you create it: Look for existing models, create your own
Format: No specific format. Could be an outline, a set of topics
and descriptions
What it’s good for: Overlay on your customer journey to find
where content is needed. A starting point for any content
strategy.

CONTENT MODEL
What it is: A structured representation of content types and
their relationships. The content version of a data model.
How you create it: Organize content types, determine
internal structure of each. Create metadata to make the system
work.
Format: Represented by fields in a content management
system. Designed in variety of software.
What it’s good for: Designing structure within a body of
content. Good starting point to write, even better for
determining how content can be reused.

HOW-TO: DOMAIN AND
CONTENT MODELS

KNOWLEDGE BASE
What it is: A group of electronic documents on a set of topics,
ideally in some organized form.
How you create it: Assemble all your stuff!
Format: Ideally, documents are text with metadata and
semantic structure. YMMV.
What it’s good for: Base of content for AI, chatbots.

WATCH OUT FOR:

TAXONOMY
What it is: A hierarchical structure for organizing information.
Could be topics, products, whatevs.
How you create it: Build categories that make sense. Create
metadata to extend the structure into other uses.
Format: Spreadsheet, folder structure, semantic tags, XML,
special software…possibilities are endless.
What it’s good for: Organizing items like content, products.
Structural base for AI, chatbot content, more.

HOW-TO: TAXONOMY

ONTOLOGY
What it is: A structure for demonstrating relationships. Could be
topics, products, whatevs.
How you create it: Group like entities, outline relationships
between them. Create metadata to extend the structure into other
uses.
Format: Semantic tags, XML, special software…many possibilities.
What it’s good for: Organizing items like content, products, and
especially complex information. Structural base for AI, chatbot
content, more.

KNOWLEDGE GRAPH
What it is: Google presents
structured data in a box on the
right side of the search results
page.
How you create it: Put good
metadata on your content, be
liked by Google’s algorithm.
What it’s mixed up with:
Non-branded information
representations.

WHATEVER YOU
CALL IT

COORDINATION
AND CAPABILITIES
➤Does any other system need to
understand your structure?
➤Do you need to map to any
other system?
➤Does your industry have a
standard structure?
➤What can your tech system do
now?

BUILDING A
TAXONOMY

BUILDING A TAXONOMY
GOALS INFORMATION ORGANIZATION GOVERNANCE
Manage
information well
Integrate with
other systems
Help people find
things
Support
advanced
content
What will be
covered?
How will info be
categorized?
What facets
matter?
How do you
identify the
characteristics?
Are relationships
important?
Document and
understand
integrations
Be flexible
Maintain
organizing
principles until
there’s reason
not to
Look for industry
standards
Group like things
together
Seek clarity
Use user language
for labels
Be cautious and
intentional about
redundancy
Test!@lauracreekmore #cmworld

HOW-TO: CARD SORTING

• You’re designing an ecommerce store.
• Decide how to organize your products.
• What are the product categories?
• Where do you stop having categories and start having
metadata?
• How will you group your metadata?
Tip: You may not have all the labels you need. Make more so that your information architecture
makes sense to you, and to the computer.

DESIGNING CONTENT
What makes content design different in advanced tech

ADVANCED CONTENT
OPERATIONS: CHAT & VOICE
CHATBOTS VOICE SKILLS
Define topics
covered
Define
unsuccessful
interactions
Q & A/
Triggers
Supporting
content
Similar to
chatbots, but add
challenges of
speech recognition
Evaluate voice vs.
typed conversation

ADVANCED CONTENT
OPERATIONS: AI
Define
knowledge base
Create
taxonomy/
ontology
Add semantic
structure
FOR WHAT?
➤PPC program management
➤Topic and editorial calendar recommendations
➤Competitor analysis
➤The Holy Grail: Content creation?

HOW MUCH DATA
DO I NEED?
No, even more.
Still more.
More than that, even.

Current experience shows that AI
initiatives often fail due to the lack of
appropriate data or low data quality.
--Introducing Semantic AI white paper
Pool Party

WHAT KIND OF
DATA?

ADVANCED CONTENT
OPERATIONS: ECOMMERCE
Determine
integrations
Define existing
structures
Structure at least
as granular as the
most granular
system
Structure to
create/store/
deliver content
FOR WHAT?
➤Integration between all suppliers/retailers
➤Make it easy to find and buy your products
➤Create faceted navigation, search filters
➤Create comparison tools, rich content
experiences

COMMON BOTS/SKILLS
➤Provide standard information [account, order, events]
➤Create a list
➤Place an order
➤Play media

USER INTENTS
➤Break down user “intent” into specific actions
➤Think of intents as the smallest possible step
➤For each intent, there may be a range of words and
phrasing your users will use
➤For each intent, you may need to provide a range of options

➤ Play music
➤ 80s music [need to have music grouped into genres or by decade of publication]
➤ Hound Dog by Elvis Presley [need to understand “You Ain’t Nothing But a Hound
Dog by Elvis Presley,” or perhaps “Ya Ain’t Nuthin’ But a Hound Dog by Elvis
Presley”]
➤ Pause music
➤ Stop music
➤ Turn it up
➤ Turn it down
INTENT EXAMPLES

INTENTS FROM AMAZON

• You’re designing a voice or chat bot.
• Decide what the bot will do.
• Who is your audience?
• What are the intents [mini-use-cases] they will have? How
might users express that intent? Use this format:
User need 1: Utterance 1 [Play music: Play “song name”]
User need 1: Utterance 2 [Play music: Play “song genre”]
Tip: Get specific. Really specific. Even more specific.

WHAT YOU CAN
ORGANIZE
➤Topics
➤Articles
➤Images
➤Videos
➤Products

BUILDING BLOCKS
OF CONTENT@lauracreekmore #cmworld

THE BATTLE FOR THE
BODY FIELD

TRAINING THE CMS

ENTITIES V. INTENTS
Play music: Play “song name”
Play music: Play “song genre”
Intent Entity

• Choose up to 3 intents.
• Walk through the intents and determine what information
is needed to support them.
• When will users already know the information?
• When will the bot need to support them with options?
• How should it best do that?
Tip: Think about how this information needs to be organized for the bot to understand it, and for
people to use the information when the bot shares options.

DESIGNING FOR FAILURE
➤Bots can’t do everything
➤Common scenarios they can’t accomplish
➤Plan for uncommon scenarios

TECHNOLOGY CHOICES
Platforms…
➤Amazon Alexa
➤Facebook Messenger
➤Google Home
➤WhatsApp
➤Your website
➤More
…v. builder tools
➤Dozens and dozens

MAKING THE RIGHT CHOICE
➤What platforms can the bot builder integrate with?
➤Can you create both voice and chat bots?
➤How easy is it to integrate with your existing content
and systems?

STRUCTURED
CONTENT
ENABLES THE
AUTOMAGIC
My favorite internet word

RESOURCES
➤ Taxonomy in the Age of Amazon Echo. Mike
Doane, Taxonomy Bootcamp 2018
➤ Taxonomies and AI Chatbots. Gary Carlson,
Taxonomy Bootcamp 2018
➤ AI UX: 7 Principles of Designing Good AI
Products
➤ Machine Learning for Everyone
➤ Designer’s Field Guide to Data
➤ How Voice-Activated Tech Will Change Content
Marketing
➤ Opportunities for AI in Content Marketing
➤ Google’s Machine Learning Crash Course
➤ Designing Connected Content by Mike Atherton
and Carrie Hane
➤ The Human Insights Missing From Big Data, Ted
Talk by Tricia Wang
➤ Why Big Data Needs Thick Data, Medium post by
Tricia Wang
➤ Designing Bots by Amir Shevat
➤ Designing Voice User Interfaces by Cathy Pearl
➤ The Accidental Taxonomist by Heather Hedden
➤ The Battle for the Body Field article by Jeff Eaton
➤ Training the CMS article by Eileen Webb
➤ Card Sorting by Donna Spencer
➤ Conversational Design by Erika Hall
➤ Content Design for the Conversational UI.
Melanie Seibert, STC Summit 2019
➤ Top 26 Chatbot Builders for 2019, CMS Wire
➤ IBM Watson Assistant
➤ Conversational Commerce: Using Product
Information to Drive ChatBot Dialogs, Seth
Earley and Dave Skrobela
➤ There’s No AI Without IA, Seth Earley
➤ Alexa Skills Kit
➤ Google Smart Home Device Traits

Laura Creekmore
@lauracreekmore
laura.creekmore@syndigo.com
https://slideshare.net/lauracreekmore

Content Structure Workshop - Content Marketing World 2019

Related slideshows

More Related Content

Content Structure Workshop - Content Marketing World 2019