This deck covers both the abstract principles and tactical nitty-gritty of Schema.org semantic markup to help site content be better understood by search engines.
It was presented at an SEMpdx event on February 11, 2014.
1. Under the Hood
Advanced Semantic Markup for SEO
Will Hattman
Organic Search Specialist
will@anvilmediainc.com
2. The Information Superhighway
Remember this?
This was the dominant
metaphor for the wondrous
information-rich future that we
were promised by the TV
talking heads back in the 90s.
3. The Information Superhighway
Remember this?
This was the dominant
metaphor for the wondrous
information-rich future that we
were promised by the TV
talking heads back in the 90s.
[TV is this thing that we
had before we had the
internet.]
4. Try Information Heaven
The internet is not a physical
space, so the idea of being slave to
the limitations of the physical
world while exploring it is
unacceptable.
The internet is a space without
distances.
Therefore, it’s a space where you
get what you want the instant you
want it, even if you don’t know
where it’s going to come from.
That’s not any kind of highway.
That’s heaven.
5. The Hive Mind
The internet is where we decided to build our collective mind.
But it’s not enough
that it simply stores
information, like our
individual minds
do…
we also want it to
function the way
our minds function:
non-linearly.
Associatively.
6. Search Is What Makes This Possible
The power and availability of
search engines is the reason we
can navigate the hive mind in
much the way we navigate our
own minds.
Search makes the process of
retrieving content from the web
as easy as retrieving one of our
own memories.
In theory.
7. Why Do We Search?
All search depends on a fundamental premise…
8. Why Do We Search?
All search depends on a fundamental premise…
9. The Content Is Out There
…and not only is it a search engine’s job to find
it, but because of our love for the hive mind
model, we increasingly can’t get to the content
we seek without their help.
You are here, the content you seek is there,
and search paves the road between.
With each passing year, search is less and
less just one internet tool among many,
and more and more how we do the
internet.
10. The Content Is Out There
…and not only is it a search engine’s job to find
it, but because of our love for the hive mind
model, we increasingly can’t get to the content
we seek without their help.
You are here, the content you seek is there,
and search paves the road between.
With each passing year, search is less and
less just one internet tool among many,
and more and more how we do the
internet.
SO WHY DOES
SEARCH EVER
FAIL?
12. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
13. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
14. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
15. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
(indexation)
16. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
(indexation)
they need to know which sources to trust
17. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
(indexation)
they need to know which sources to trust
(ranking factors)
18. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
(indexation)
they need to know which sources to trust
(ranking factors)
they need to understand which of all the pages on the web
would best satisfy your query.
19. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
(indexation)
they need to know which sources to trust
(ranking factors)
they need to understand which of all the pages on the web
would best satisfy your query.
(the tricky part)
20. Information Retrieval
What do search engines need in order to be able to
give you what you’re looking for?
they need to know what’s out there to be served
(crawling)
they need to remember where it all lives
(indexation)
they need to know which sources to trust
(ranking factors)
they need to understand which of all the pages on the web
would best satisfy your query.
(the tricky part)
The evolution of the search engine is the evolution of machine understanding.
21. Search Engines Need Help
Search engines have always sought to
discover the meaning of webpages.
There is a difference between knowing the names of
foods and knowing what they are.
22. Search Engines Need Help
Search engines have always sought to
discover the meaning of webpages.
The problem is, for most of the web’s
life, they’ve had very little to go on.
There is a difference between knowing the names of
foods and knowing what they are.
23. Search Engines Need Help
Search engines have always sought to
discover the meaning of webpages.
The problem is, for most of the web’s
life, they’ve had very little to go on.
Page copy
Headings
Metadata
URLs
Anchor text of inbound links
No matter what your webpage is
about, these elements are
fundamentally the same.
Traditionally, everything rests on
how well you populate these fields,
and how well a search engine can
interpret their contents.
There is a difference between knowing the names of
foods and knowing what they are.
24. Making Machines Understand
If machines only know what you tell
them, how can they be made to
understand meaning?
We have to input not just data, but
connections between data.
We have to program not just values,
but relationships.
We have to show them how to
associate.
25. Triples
The concept of the knowledge “triple” is
the basis for conveying meaning to
machines.
It designates three elements:
subject
predicate
object
and arranges them to describe a
relationship.
The subject has a relationship with the
object. The nature of the relationship is
described by the predicate.
Thus, though a machine cannot actually
learn or experience meaning the way a
person can, it can simulate an
understanding of meaning by reference
to an ever-expanding web of relational
context.
courtesy of http://www.rdfabout.com/intro
26. Enter Semantic Markup
Semantic markup uses this relational
model, based in triples, to designate
various types of pages on the web
based on their content, as well as
various properties within that content,
in ways that search engines can
understand
concretely
and
unambiguously.
27. The Emergence of
The groundwork for the semantic
web was laid years ago, and several
different, competing standards arose
in a short period of time.
Schema.org, arriving in 2011, was the
first semantic markup vocabulary
with ambitions of universal adoption.
28. The Emergence of
The groundwork for the semantic
web was laid years ago, and several
different, competing standards arose
in a short period of time.
Schema.org, arriving in 2011, was the
first semantic markup vocabulary
with ambitions of universal adoption.
The fact that it was launched, funded,
and developed by these guys
probably is reason enough to believe
that it will become the standard.
29. The Emergence of
The groundwork for the semantic
web was laid years ago, and several
different, competing standards arose
in a short period of time.
Schema.org, arriving in 2011, was the
first semantic markup vocabulary
with ambitions of universal adoption.
The fact that it was launched, funded,
and developed by these guys
probably is reason enough to believe
that it will become the standard.
30. The Emergence of
The groundwork for the semantic
web was laid years ago, and several
different, competing standards arose
in a short period of time.
Schema.org, arriving in 2011, was the
first semantic markup vocabulary
with ambitions of universal adoption.
The fact that it was launched, funded,
and developed by these guys
probably is reason enough to believe
that it will become the standard.
Actually, it kind of already has.
31. Let’s Get Practical
How does it work?
As with most of the
world’s best things,
this question is best
answered with food.
Recipes constitute the
most visible semantic
search showcase on
the web. Search
Google for any dish by
name and behold the
wonder.
rich snippets galore
bipartite knowledge graph entry
32. Let’s Get Practical
How does it work?
As with most of the
world’s best things,
this question is best
answered with food.
Recipes constitute the
most visible semantic
search showcase on
the web. Search
Google for any dish by
name and behold the
wonder.
WHAT
SORCERY
IS THIS?
rich snippets galore
bipartite knowledge graph entry
33. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
34. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
specifies class of item
35. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
Then, you indicate which
properties of the item
you’re going to
designate.
specifies class of item
names
property
36. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
specifies class of item
names
property
Then, you indicate which
properties of the item
you’re going to
designate.
Then, you assign
a value to each
property.
specifies value of item
37. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
specifies class of item
names
property
Then, you indicate which
properties of the item
you’re going to
designate.
Then, you assign
a value to each
property.
specifies value of item
38. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
specifies class of item
names
property
Then, you indicate which
properties of the item
you’re going to
designate.
Then, you assign
a value to each
property.
specifies value of item
39. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
SUBJECT
PREDICATE
Then, you indicate which
properties of the item
you’re going to
designate.
Then, you assign
a value to each
property.
These are just glorified triples.
OBJECT
40. What Sorcery Is This?
In most cases, rich
snippets are
generated from
site-side markup.
First, you specify an
itemtype: a basic class
type for your item.
SUBJECT
PREDICATE
Then, you indicate which
properties of the item
you’re going to
designate.
Then, you assign
a value to each
property.
These are just glorified triples.
OBJECT
(as long as you’re comfortable with something like “has property” as a universal verb)
41. But I Don’t Have a Recipe Site
Schema.org semantic markup can also
generate rich snippets on searches for:
Events
supported properties include
—
—
—
—
location (including geographical coordinates)
event type & description
start date, end date, duration
ticket purchase CTA
Products
supported properties include
—
—
—
—
—
—
—
—
image
description
brand
category
reviews & ratings
price
SKU
offer details
42. Other Schema Types
People
supported properties include
—
—
—
—
—
—
name & nickname
photo
title/role
affiliations
social relationship to searcher, if any
address
Organizations and Local Businesses
supported properties include
—
—
—
—
name
address (and/or geographical coordinates)
telephone number
logo
43. Schema.org Is Ever-Growing
All of these are supported for at least some properties, and
more types are added to the list every year.
45. Specialized Search
Pages bearing semantic markup can
qualify for certain specialized searches
on Google as well.
This feature allows for search results to
be filtered
46. Specialized Search
Pages bearing semantic markup can
qualify for certain specialized searches
on Google as well.
This feature allows for search results to
be filtered
according to as many properties
47. Specialized Search
Pages bearing semantic markup can
qualify for certain specialized searches
on Google as well.
This feature allows for search results to
be filtered
according to as many properties
as the item class in question
48. Specialized Search
Pages bearing semantic markup can
qualify for certain specialized searches
on Google as well.
This feature allows for search results to
be filtered
according to as many properties
as the item class in question
allows.
49. The Knowledge Graph
What about the stuff
on the right?
rich snippets galore
bipartite knowledge graph entry
50. The Knowledge Graph
What about the stuff
on the right?
rich snippets galore
bipartite knowledge graph entry
51. The Knowledge Graph
This is Google’s Knowledge Graph.
It launched in 2012 and at the time constituted Google’s largest leap
yet toward the semantic web.
It is:
— gathered from trusted sources, defined by Google in familiarly
vague terms
— *the sources most commonly cited are Wikipedia,
Freebase, and the CIA World Factbook
— not dependent on site-side work
— aka impervious to your overtures
— designed to limit the world’s need for click-through altogether
— if Google feels equipped to answer the question directly on
the search results page, it’s not going to waste your time
making you perform an extra click
— based entirely on a relational model of understanding
— i.e. a thicket of triples in the billions
52. What Should I Do?
For Rich Snippets:
— determine which Schema.org itemtypes your content qualifies for
— determine whether your CMS will allow you to implement via a plugin, or whether you
need to hard-code
— test your code using Google’s Structured Data Testing Tool
www.google.com/webmasters/tools/richsnippets
— keep an eye out for evidence of new Schema.org support
— propose your own expansions to the vocabulary
For the Knowledge Graph:
— there is no direct path to earning a spot in the Knowledge Graph
— the long game involves getting yourself, your site, or your brand noticed and respected
— build trust over time via good acts (i.e. #RCS), good site content, and good SEO
— Google’s trust in you will increase with every forward-facing SEO measure your site
adopts
— Schema.org
— Google Authorship
— Full Google+ brand verification and community participation