Presented at CIDM RIDE conference, 2018
A customer-centric approach to managing content metadata, components, and deliverable assets using taxonomies, ontologies, and AI
1. Michael Priestley, IBM / @ditaguy
September 26, 2018
Customer-centric content and Info 4.0
2. Michael Priestley @ditaguy
Michael Priestley is a product owner and
content technology strategist, currently leading
the IBM Marketing Taxonomy Guild to revise
and align taxonomy initiatives across the
marketing ecosystem. He has experience
working with and across documentation,
support, training, and marketing content as an
enterprise content technology strategist. He
was one of the original architects and editors of
the DITA standard, was named an OASIS
Distinguished Contributor in 2017, and is
currently co-chairing the Lightweight DITA
subcommittee.
4. Building what customers want
Do we
have
it?
Did it
work?
What
do they
want?
Plan
Measure
Manage
Publish
5. Challenge – consistent tagging
Metrics
Tagging
Need
automated
service for
consistent
tagging
Model
depends on depends on
What do they want? Do we have it?
Did it work?
Who does the tagging?
Who benefits?
7. Requirements for customer-centric development
People
ProcessTools
Cultural shift from
organization
centric to customer
centric
Governance across
tools and authoring
communities
Cognitive tagging for quality
assurance and content coverage
Ontologies to assist planning
8. Giving customers what they want
Do we
have
it?
Did it
work?
What
do they
want?
Customer awareness
Improvement in selection
Content awareness
Measures of success
9. Requirements for customer-centric delivery
Common
customer data
Common
content
metadata
and
ontologies
Common
linking
components
Plus common KPIs, common process for improvement...
13. A chemistry of content
Test
hypotheses
Need data
Need
measurements
Need a
consistent
scale
13
What worked?
What didn’t?
Can we repeat the experiment?
Can we compare results?
From words to websites
14. Chemistry vs alchemy
Chemistry Alchemy
Tested, validated Based on intuition/gut
“You can tell the age of a tree by counting the rings” “You can tell the age of a horse by counting its teeth”
Has predictive power Has descriptive power
As many elements as we discover Four elements in various combinations
Make things that work Accidentally discover gunpowder
14
15. From words to websites
Words Components
•images, text, links
•cards, banners, phrases
•sections, bands, blocks
Assets
•data sheets, product
overviews
•explainers, case studies
•Reports, solution guides
Collections
•lesson, newsletter
•book, course
•knowledge base,
community
Websites
15
17. Customer-centric assets
What does the
user want from
the content?
Are there
patterns within
the content that
predict success?
What influences
success?
17
Identify types Structure types Identify and correct
for influences
Time of day, author, client...
18. This is a ___ delivered as a ____
What does
the customer
need to
understand?
Choose the
best content
type
How does the
customer
prefer to
learn/get info?
Choose the
best content
format
18
Example: this is a client story delivered as a video
19. A single asset may have multiple expressions
Interview
Video
Podcast
Transcript
19
20. A single event or engagement may drive multiple assets
Client
success
Client
story
Case
study
Solution
guide
20
21. Share components across assets
21
definitions
features
specs
video
scripts
Text
components
chat
scripts
diagrams
screen
caps
photos
Image
components
article
web
PDF
video
chat
presentation
social
Track
across
formats
Chat transcripts
% watched
downloaded
read/scrolled
engaged
22. Learning plan
Share components
and assets across
organizations
Training
Product docs
Marketing
Sales
Course
Tutorial
Feature
overview
Steps
Tutorial
RFP
Feature
Benefits
Case study
Feature
overview
Business
benefits
Knowledge center product collection
Tutorial Feature Steps
Feature
Steps
Benefits
23. Reuse catalog
§ Quickly assemble and filter
components for new
campaigns/journeys
§ Reuse of both assembled and
component content
§ Can reuse dynamically – always pick
best match for criteria, regardless of
origin
§ Measure performance at component
level
Runtime reuse
Benefits
Tutorial
Feature
Steps
Case
study
Dynamic
collection
For criteria:
• CIO
• Learning
• Interests in
cloud, security
24. Use a combination of dynamic and static pages
Dynamic
Always fresh and relevant
Poor for SEO
Great for site discovery
Use for personalization
Static
Can grow stale
Great for SEO
Great for repeat visitors
Use for evergreen content
24
Valuable assets should have a predictable, sharable, stable URLs of their own
25. Give customers what they want
Do we
have
it?
Did it
work?
What
do they
want?
Build it Deliver it