4. 4
KEYNOTE WITH GARY ILLYES
• HTTPS – still his baby, says everyone should do it and cannot think of a reason not to
• Mobile
• Since 85% of all results are now mobile friendly, the mobile friendly, faulty redirect, and flash-enabled
labels were dropped
• AMP is mobile friendly by default, no direct ranking benefit at this time, just a lightening bolt label
• Rater guidelines are used to test and validate algo updates
• They either a) roll out the change to 1% of all users or b) have human raters decide which results are
better
• RankBrain was released (some rumored one year prior) and no one noticed, not even MozCast
• Only impacts super long tail and negative queries – think stop words like “without” or natural language
like “weak”
• Machine learning is hard, trying to learn by failure
• Panda is still continuously rolling out and Google will no longer mention anything about it because Jen
Slegg covered it so brilliantly in this article
• Penguin is in real-time and doesn’t demote, but discounts or ignores signals like links
• Don’t be afraid of legit link directories
5. 5
KEYNOTE WITH GARY ILLYES – NEXT YEAR
• Desktop index first will shift to mobile index first very soon (mobile-first indexing)
• Desktop will be ranked based on mobile index instead of other way around -- First time ever having two
separate indexes
• Not true, it was split into two but only during the initial switch-over
• Systems weren’t designed to handle two versions of the same page but they decided they would make
the switch because pretty much everything is mobile friendly and more than 50% of all queries are
mobile
• Users hate overly stripped down mobile sites that are missing content
• Mobile first indexing is not happening now but will be very soon as it’s an extended experiment that will
take a chunk of desktop index and create a separate mobile index which will eventually be the primary
index
• The desktop version will be considered the full deck index and will be less up to date since it won’t
be crawled nearly as often (way less crawl power)
• Current: 80% desktop / 20% mobile and Future: 80% mobile / 20% desktop
• Actually more like Current: 100% desktop and Future: 100% mobile
• Problems: Mobile pages typically have less real estate, less content (tokens = words), less links which
will make it harder for Google
• Migrate technical SEO tags, meta data, content, and structured data on mobile pages
• Decide what you care about and identify key differences in mobile to desktop pages
6. 6
KEYNOTE WITH GARY ILLYES – NEXT YEAR
• Assistant-type features like Google Home will propel voice search
• Think personal assistant, navigation, question answering, although not sure how Publishers will be
integrated
• Don’t rush to jump on this unless you are a massive company like Amazon, it’s purely experimental and
will leverage an API platform that is going to change a lot
• Progressive Web Apps (PWA)
• Makes it so that you engage with the users via app without them having to download it using push
notifications -- it essentially takes away the step but offers the same features of having the app
• Don’t do this unless you already have a native app
• GSA timeline data will be longer eventually – feedback from Google Dance
8. 8
KEYNOTE WITH BILL HUNT
• C-suite still doesn’t care about DM enough
• Bill creates badass data models for his clients
• Model to show why 1 is 1 and 2 is 2 to cut out the BS
of SEO
• Graph to map out organic CTR by search position
• Model to show searcher interest
• Model to show rank to click by persona
• Model to show keyword opportunity analysis
• Model to show ranking changes by business unit
(change by embarrassment)
• Model to show paid and organic keyword co-
optimization
• Model to show keyword clusters based on intent and
step in decision making process
• Searcher content requirements have shifted now that there
are no 10 blue links anymore
• Don’t act on BS Twitter journalism
11. 11
SEO 2016
BILL HUNT, DUANE FORRESTER, & GREG BOSER
http://www.pubcon.com/vegas2016/pubcon65/229-bill_hunt-1-final.pdf
http://www.pubcon.com/vegas2016/pubcon65/813-duane_forrester-1-final.pdf
http://www.pubcon.com/vegas2016/pubcon65/439-greg_boser-1-final.pdf
12. 12
SEO 2016: NEW REALITIES
• Mobile now, not mobile first = default web experience
• PWA will be very important for the future
• Don’t use mobile emulators, use actual smartphones to test
• You’re biggest mistake is poor UX and not moving fast enough in terms of execution and speed
• The next gen consumer isn’t as brand loyal, a good mobile UX helps them become loyal
• SEO is now about maximizing findability in a multi-channel world
• FS is an example of Google absorbing more media
• FS is very fluid, but worth the effort
• Voice search sucks right now, but it will get better
• AI is like a 2 year old right now – it’s learning and evolving
• Believe that RankBrain is learning through failure and even FS
• SEO is less technical only because everyone’s already doing it
• SERP is still shrinking
• Can’t ignore other marketing channels, especially email and loyalty programs = most effective
• RankBrain is overhyped
• AI is the future, forget about it for now
• Write long tail content or leverage an authoritative video hosting service
13. 13
SEO 2016: GREG BOSER RANT
• Twitter style “journalism” is making us SEO stupid
• There’s been an explosion of bad, context-less content
14. 14
SEO 2016: GREG BOSER RANT
• HTTPS
• No reward was seen (similar to Mobilegeddon)
• 2017: starting in January, anything that should be HTTPS (pages with forms, secure data) that is HTTP
will show a Chrome browser error and freak out your users
• Eventually we plan to label all HTTP pages as non-secure
• Doubt that they will flag all types of content but do it by end of 2017
• Move to HTTPS – Greg’s official standpoint because of browser alerts, thinks it’s overkill but it’s where
they are going, make it a priority now
• It’s not a quick and easy transition, unless on WP
• Ton of dependencies, complete commitment from dev team, tons of research needed, especially
third-party content (CDN)
• Allocate proper time for planning
https://www.wired.com/2016/09/wired-completely-encrypted/
https://docs.google.com/spreadsheets/d/1ixlsxheqvazivao9Joc3p57XZbRwgB2XI4_R1NLhqwQ/edit#gid
=1975121463
15. 15
SEO 2016: GREG BOSER RANT
• Penguin 4.0
• Don't know a lot so far
• Real-time, as they crawl they make decision kind of on the fly to automate a formerly manual process
• Less punitive because
• Page-centric vs. site-centric = penguin filter used to be page-centric, panda is more site-centric
• Demote vs. devalue = doesn’t demote, it devalues it and ignores it in theory
• Going to be multiple iterations, not as sophisticated as Google wants you to believe
• Fake it till they make it
• Cuts out human review element
• It’s time to disavow the disavow file!
• Disavow file is about data collection
• AI is learning from disavow files
• Hugely valuable for building a better algo
16. 16
TECH ISSUES
BARRY ADAMS & DAWN ANDERSON
http://www.pubcon.com/vegas2016/pubcon65/208-barry_adams-1-final.pdf
http://www.pubcon.com/vegas2016/pubcon65/208-dawn_anderson-1-final.pdf
17. 17
TECH ISSUES: AMP
• Mobile first but really mobile only
• Gives Google full control over pages
• Serves up alt versions of web pages
• Uses caching and pre-rendering which is why it’s so fast,
served from Google server
• Probably AMP support for all sorts of content by EOY
• Severely restrictive and it’s an all or nothing sort of thing for
validation
• HTML AMP declaration
• Structured data is necessary and must be integrated (in
head)
• AMP must use inline CSS instead of a single external CSS
file
• You have to use JS and it has to be specific to AMP
• There is AMP specific analytics even, recommend separate
GA profile to clean up the data, most likely not accurate
though (under-reporting)
https://ampbyexample.com/
https://validator.ampproject.org/
18. 18
TECH ISSUES: AMP
• AMP has “expanded exposure” which basically means you have no choice.
• Recommend looking into it regardless of what kind of site it is
• WordPress is easy, just use a plugin (AMP & PageFrog)
• For other platforms, plugins are probably coming but for now, it’s a huge manual effort to undertake
• Full end-to-end ecommerce and most JS-based interactivity doesn’t work with AMP yet
• Everything breaks AMP like social sharing, video embeds, and comments
19. 19
TECH ISSUES: CRAWL BUDGET MYTHS
• Crawl budget = host load + URL scheduling
• Host load = what can you handle?
• Not just about your site
• Relates to an IP level and is shared amongst the sites there
• Host load is mostly about server capacity than SEO
• URL scheduling = what is important to crawl and how often?
20. 20
TECH ISSUES: CRAWL BUDGET MYTHS
• Crawl budget CANNOT be defined using GSC crawl stats report
• Includes all types of files and not just single web pages
• Includes all 10 bots of the Googlebot family including paid search bots
• Includes all 200 and 30X responses, but does not include 40X or 50X
• Doesn’t show you what URLs have been crawled and when
• Frequency is at URL basis
• You cannot just divide the number of pages crawled to find the number of
pages crawled per day
• The only way to find out which URLs are being crawled and when is to
constantly review server logs and log analysis – especially important for
large sites
• Can be done through automatic CRON jobs then analyze the data with tools
https://support.google.com/webmasters/answer/35253
https://goo.gl/1pToL8
21. 21
TECH ISSUES: CRAWL BUDGET MYTHS
• Googlebot does NOT crawl through your site from one end to the other then repeats
• Crawling is very complicated and the web is littered so bots look for certain elements for crawl
efficiency with the purpose of minimizing visibility of stale content
• There is no duplicate content penalty but duplicate content is filtered and not indexed
• The content checksum Google patent and other crawl efficiency patents de-dupe and sorts out the
noise
• Keep internal signals consistent
• Depth of crawl is greater in higher quality sections of site
• Important parents make important children
• Quality can only be shown in server logs
• Low quality, stale sites get crawled less frequently
• Not based on PageRank -- used to, but not anymore
22. 22
TECH ISSUES: CRAWL BUDGET MYTHS
• URL scheduling is not based on just PageRank, it’s mainly driven by URL importance which is
independent from the query
• URL importance may include:
• Location in site
• PR and internal PR
• Page / file type
• Internal links
• Internal anchor text consistency
• Relevance to a topic (content, anchors, elements)
• Meta robots and robots.txt directives
• Parent quality
• Inclusion in XML sitemap and the index
• Emphasis on IA, internal links, technical SEO tags, merging/updating thin content, XML sitemaps,
etc.
• Crawl rank is correlated to SERP ranking, but it’s not exactly causation
• URL importance dissipates when signals are inconsistent, especially with URI/URL name changes
23. 23
TECH ISSUES: CRAWL BUDGET MYTHS
• IBM’s Joel Wolf coined the theory of “Search Engine Embarrassment”
• Basically says you need to update/change pages at a certain threshold to keep it
fresh and therefore crawled more often
• Cannot be randomized or manipulated
• Importance > change
• XML sitemap tags
• If accurate, <lastmod> is taken into consideration
• <priority> and <changefreq> are ignored, don’t make it up
26. 26
FEATURED SNIPPETS
• Google loves FS because it’s quicker, less hassle, and better for mobile
• Google says to state the question implicitly or explicitly on the page
• How does Google decide which queries get FS according to the GSQRG?
• Know Queries = Usually don’t get FS
• Informational queries that are broad, complex and/or in-depth
• No short, or “right” answer
• Controversial, opinion, broad
• Know Simple Queries = Many have FS
• Queries that seek a very specific answer
• Fact, diagram, mostly everyone agrees with it, one size fits all
• Answer must be correct, complete, and displayed in the size of a mobile screen
27. 27
FEATURED SNIPPETS
• How to get images in FS
• “og:image”
• <meta property=“og:image” content=http://example.com/image-description.jpg” />
• Alt keywords
• Keywords in image name
• Doesn’t have to be a brand new page
• Landscape works better than portrait
• Changing the image can entice CTs
• How to get your YouTube FS
• YouTube will show a screenshot – nothing on YouTube is too dumb
• Required: voice
• The source is the transcript not the description
• How to get recipe FS
• Most FS are recipes
• Use schema markup
• Expands the regular results
• Use AMP because AMP carousels are coming
• Be careful with recipe plugins which could spark a manual action
• Bing recipes produce tabs
28. 28
FEATURED SNIPPETS
• How to find your FS opportunities
• Identify implied Know Simple queries by mining
GSC SA queries for those that include “how”, “why”,
“when”, “who”, “where”, and “what”
• SEMRush Featured Snippet and long tail queries in
the Phrase Match report
• Moz Keyword Explorer SERP Analysis Features
• Buzzsumo
• Mine your competitors
• How to change the snippet
• Feedback button
29. 29
FEATURED SNIPPETS
• Voice search is the next generation of FS
• Voice searches return more FS than text searches
• Google reads out the FS answer by starting with “according to
[domain name]…”
• Yay branding!
• AMP may get preference for FS results along with PWAs
• People Also Ask popping up more than ever
• Google gets it wrong still
• May be the same results getting selected as a FS
• No way to track it and most likely not including in the rich result
report of GSC SA
30. 30
STRUCTURED DATA
BILL SLAWSKI & CARRIE HILL
http://www.pubcon.com/vegas2016/pubcon65/218-bill_slawski-1-final.pdf
http://www.pubcon.com/vegas2016/pubcon65/218-carrie_hill-1-final.pdf
31. 31
STRUCTURED DATA
• Google loves it because it basically does their work for them by organizing and defining data that they
can use for their KG API and building entities
• Vocabulary = Schema, GoodRelations
• Markup = JSON, metadata
• Can be penalized, not a bandaid for bad SEO
• Schema.org updates 2-3 times / year http://schema.org/docs/releases.html
• Make Make – latest and added markup for accommodations
• Deimos and Phobos – not major, but added geocircles and special open hours
• Review markup is strict and cannot exist anywhere else
• Using third party reviews should use different markup (critic)
• Structured data helps with voice search
• Table suggestions for FS placement:
• Limit boilerplate, use table headings <th> to add columns, use meaningful attribute names in table
headings, use meaning titles, captions, and semantically related text surrounding the table
33. 33
OMNI-CHANNEL DM STRATEGY
BILL HUNT
http://www.pubcon.com/vegas2016/pubcon65/149-dave_roth-1-final.pdf
http://www.pubcon.com/vegas2016/pubcon65/149-rhea_drysdale-1-final.pdf
34. 34
OMNI-CHANNEL DM STRATEGY
• Requirements
• You have multiple channels to market on
• You're confident in your strategy and know it will resonate with stakeholders
• You have the resources needed to coordinate and implement campaigns
• Omni-channel = great brands = coordinated efforts = highly effective = great reputations
• Reputation marketing is the 2nd phase of reputation management which only happens once everything
is going well
• Marketers have bad reps and are siloed, wasteful, and often prioritize short-term growth over long-term
strategy
• Especially true for agency, we forget about the non-campaigns
35. 35
OMNI-CHANNEL DM STRATEGY
• Great brands need:
• Internal brand resource center that
includes all marketing assets,
editorial calendar, and brand
guidelines
• Cohesive design and singular voice
• Develops a verbal identify (style,
tone, promotional copy, organic
intent documentation, SM
promotions / replies)
• Identify how stakeholders are
engaging with you and invest in
technology that makes you visible
• Personas don’t need to be
fancy. You can literally just
email your customers and ask
them who they are, what they
do, which devices they use, and
which channels they are on
37. 37
THE NEW SERP: PAID & ORGANIC
ERIC ENGE & NAVAH HOPKINS
http://www.pubcon.com/vegas2016/pubcon65/331-eric_enge-1-final.pdf
http://www.pubcon.com/vegas2016/pubcon65/331-navah_hopkins-1-final.pdf
38. 38
THE NEW SERP: PAID
• Organic gets shafted once again
• Queries with a research intent will always favor organic
where transactional oriented SERP prefers paid
• Shopping CPCs are so much cheaper and excellent when
someone is looking for a specific product
• Green ad boxes make it harder to decipher paid vs. organic
• Extensions are cool, but you still have to prioritize them for
conversion and audit them thoughtfully
• For mobile, it pays off big time to appear #1 because of
screen size
• A/B test new ad copy now because your old ads are
going to suck 2017 with the new ETA
39. 39
THE NEW SERP: ORGANIC
• KG = database on facts and relationships
• Enhanced snippets increased 195%
• Charts increased 116%
• FS sources are changing more than they are
staying the same
• 55% shows different URL when comparing
January 2016 to July 2015)
• Comparison chart KB have now disappeared
• FAQ pages usually don’t get selected
• Schema is not required
• Always submit URL to GSC after making changes
41. 41
TAKEAWAYS
1. Google’s index will be based entirely on mobile Googlebot
1. Maintain content, tags, essential structured data, links, etc.
2. HTTP browser errors are coming soon and by the end of 2017, you should migrate to HTTPS
3. Mobile is the default web experience and if you aren’t investing, you’re screwed
1. Consider AMP but hold off on PWA for now
2. Responsive still preferred
4. Real-time Penguin doesn’t demote, it ignores or discounts links
5. FS is extremely fluid but worth it, optimize for Know Simple Queries
1. For images: landscape only, alt text, og:image markup
6. Crawl budget = host load (server capacity) + URL scheduling (URL importance and
frequency)
1. Crawl budget is not defined with the GSC crawl stats report – you have to look at server
logs and log analysis
2. URL Importance consists of factors including external/internal PR, external/internal links,
internal anchor text consistency, topic relevance, robots, page/file type, XML sitemap
and index inclusion, and parent quality
7. Reviews that you mark up with schema cannot exist anywhere else
Google gives exclusives to bloggers they like which is why they get the info before everyone else
Jen is one, so is
If content is missing on mobile but exists on desktop, Google will probably ignore desktop still
Need to fetch as google smartphone
Disavow, don’t disavow, disavow but only the bad stuff?
If content is missing on mobile but exists on desktop, Google will probably ignore desktop still
If you have an m. or separate mobile url, you don’t need to update the canonicals
Overall, mobile title tags are about 8 characters longer compared to desktop ones. But when optimizing those title tags for mobile, if you go over about 70 characters, the title tag on desktop will either be rewritten by Google, or will be shown in a truncated form, cutting off the end of the tag.
There will only be one Google search index and it will be entirely based on what the mobile Googlebot sees when it crawls sites. There will NOT be a different desktop index for searches done on desktop. There will not be a different desktop index for those sites that only have a desktop version.
Away from “seo-friendly”
Findability -- not just optimized but findable
Voice system of Cortana was partnered with US military. Had to be accurate. High level of tech.
RankBrain is so overhyped
AI is cool, but more of future state, so forget about it for now
Use common sense SEO with exact match meta data if you want to show up for uncommon queries
Example: 301s don't lose PR caused explosion of shitty context-less content
Espn.go.com to espn.com, no discussion of redirects
301s to homepage is soft 404 = no PageRank passing
Value of redirect is completely dependent on the context of which it is used
If you get straight a’s, I'll take you to Disneyland = HTTPS
Example: 301s don't lose PR caused explosion of shitty context-less content
Espn.go.com to espn.com, no discussion of redirects
301s to homepage is soft 404 = no PageRank passing
Value of redirect is completely dependent on the context of which it is used
If you get straight a’s, I'll take you to Disneyland = HTTPS
Most amp pages load in under 1 second
When does host load matter?
Your site is massive, like a Amazon (100K pages)
Your site is massive and youre on a shared hosting
Your site is massive and your using a CDN
Crawl figures can change a lot especially from CDN
You have lots of arge subdomains sharing the space
You have test and staging sites that are crawlable
You keep throwing server errors dueing crwaling
Crawl stats in GSC is just a temperature reading
It will likely just a few URLs being crawled very often, some very rarely and most others somewhere in between – YOU NEED TO KNOW
Need to review your site from the spider’s eyes
Fresh results are vital to searchers
– this was the case in the past with PR
Same could be said about ranking
Internal links huge
User_View_Rate – Likelihood of the document being seen + Document_Update_rate – How often it has material changes + Web_Crawl_Interval – How often is it currently crawled COMBINED TO CALCULATE Probability(Seen_Stale_Data) = Risk of Search Engine Embarrassment? ‘JUST IN TIME SMART CRAWLING’
– this was the case in the past with PR
Same could be said about ranking
Internal links huge
User_View_Rate – Likelihood of the document being seen + Document_Update_rate – How often it has material changes + Web_Crawl_Interval – How often is it currently crawled COMBINED TO CALCULATE Probability(Seen_Stale_Data) = Risk of Search Engine Embarrassment? ‘JUST IN TIME SMART CRAWLING’
Why isn’t ASQ selected as “six sigma” FS?
Too wordy, overly technical, confusing
Revise to have natural language, lower reading level, FK score
There’s no reason not to mark everything up as long as it exists on the page – not exactly true for JSON
Used for Google Trends, locations and images of different entities
Mobile friendly if they are on mobile devices – use GSA SA by device type and segment nobranded queries comparing desktop vs. mobile
Apps if it makes sense
Email
Social
Push notifications
TV
Home assistants
Example: cross device sign in
With organic intent documentation: use query intent to identify content gaps and stay on brand in meta descriptions for CTR boost
Mobile friendly if they are on mobile devices – use GSA SA by device type and segment nobranded queries comparing desktop vs. mobile
Apps if it makes sense
Email
Social
Push notifications
TV
Home assistants
Example: cross device sign in
With organic intent documentation: use query intent to identify content gaps and stay on brand in meta descriptions for CTR boost
Mobile friendly if they are on mobile devices – use GSA SA by device type and segment nobranded queries comparing desktop vs. mobile
Apps if it makes sense
Email
Social
Push notifications
TV
Home assistants
Example: cross device sign in
With organic intent documentation: use query intent to identify content gaps and stay on brand in meta descriptions for CTR boost
Must be #1 for mobile
Dr. Pete on Penguin
Barry Shwartz
Glenn Gabe
Jenn Slegg
Aaron Walls called this out
URL scheduling is not based on just PageRank, it’s mainly driven by URL importance which is independent from the query
URL importance may include:
Location in site
PR and internal PR
Page / file type
Internal links
Internal anchor text consistency
Relevance to a topic (content, anchors, elements)
Meta robots and robots.txt directives
Parent quality
Inclusion in XML sitemap and the index