What are “fraggles” in SEO and how do they relate to mobile-first indexing, entities, the Knowledge Graph, and your day-to-day work? In this glimpse into her 2019 MozCon talk, Cindy Krum explains everything you need to understand about fraggles in this edition of Whiteboard Friday.
Hi, Moz fans. My name is Cindy Krum, and I’m the CEO of MobileMoxie, based in Denver, Colorado. We do mobile SEO and ASO consulting. I’m here in Seattle, speaking at MozCon, but also recording this Whiteboard Friday for you today, and we are talking about fraggles.
So fraggles are obviously a name that I’m borrowing from Jim Henson, who created “Fraggle Rock.” But it’s a combination of words. It’s a combination of fragment and handle. I talk about fraggles as a new way or a new element or thing that Google is indexing.
Fraggles and mobile-first indexing
Let’s start with the idea of mobile-first indexing, because you have to kind of understand that before you can go on to understand fraggles. So I believe mobile-first indexing is about a little bit more than what Google says. Google says that mobile-first indexing was just a change of the crawler.
They had a desktop crawler that was primarily crawling and indexing, and now they have a mobile crawler that’s doing the heavy lifting for crawling and indexing. While I think that’s true, I think there’s more going on behind the scenes that they’re not talking about, and we’ve seen a lot of evidence of this. So what I believe is that mobile-first indexing was also about indexing, hence the name.
Knowledge Graph and entities
So I think that Google has reorganized their index around entities or around specifically entities in the Knowledge Graph. So this is kind of my rough diagram of a very simplified Knowledge Graph. But Knowledge Graph is all about person, place, thing, or idea.
Nouns are entities. Knowledge Graph has nodes for all of the major person, place, thing, or idea entities out there. But it also indexes or it also organizes the relationships of this idea to this idea or this thing to this thing. What’s useful for that to Google is that these things, these concepts, these relationships stay true in all languages, and that’s how entities work, because entities happen before keywords.
This can be a hard concept for SEOs to wrap their brain around because we’re so used to dealing with keywords. But if you think about an entity as something that’s described by a keyword and can be language agnostic, that’s how Google thinks about entities, because entities in the Knowledge Graph are not written up per se or their the unique identifier isn’t a word, it’s a number and numbers are language agnostic.
But if we think about an entity like mother, mother is a concept that exists in all languages, but we have different words to describe it. But regardless of what language you’re speaking, mother is related to father, is related to daughter, is related to grandfather, all in the same ways, even if we’re speaking different languages. So if Google can use what they call the “topic layer”and entities as a way to filter in information and understand the world, then they can do it in languages where they’re strong and say, “We know that this is true absolutely 100% all of the time.”
Then they can apply that understanding to languages that they have a harder time indexing or understanding, they’re just not as strong or the algorithm isn’t built to understand things like complexities of language, like German where they make really long words or other languages where they have lots of short words to mean different things or to modify different words.
Languages all work differently. But if they can use their translation API and their natural language APIs to build out the Knowledge Graph in places where they’re strong, then they can use it with machine learning to also build it and do a better job of answering questions in places or languages where they’re weak. So when you understand that, then it’s easy to think about mobile-first indexing as a massive Knowledge Graph build-out.
We’ve seen this happening statistically. There are more Knowledge Graph results and more other things that seem to be related to Knowledge Graph results, like people also ask, people also search for, related searches. Those are all describing different elements or different nodes on the Knowledge Graph. So when you see those things in the search, I want you to think, hey, this is the Knowledge Graph showing me how this topic is related to other topics.
When you put this in that context, it makes more sense. He wants the entity understanding, or he knows that the entity understanding is really important, so the href lang is also really important. So that’s enough of that. Now let’s talk about fraggles.
Fraggles = fragment + handle
So fraggles, as I said, are a fragment plus a handle. It’s important to know that fraggles — let me go over here —fraggles and fragments, there are lots of things out there that have fragments. So you can think of native apps, databases, websites, podcasts, and videos. Those can all be fragmented.
Even though they don’t have a URL, they might be useful content, because Google says its goal is to organize the world’s information, not to organize the world’s websites. I think that, historically, Google has kind of been locked into this crawling and indexing of websites and that that’s bothered it, that it wants to be able to show other stuff, but it couldn’t do that because they all needed URLs.
But with fragments, potentially they don’t have to have a URL. So keep these things in mind — apps, databases and stuff like that — and then look at this.
So this is a traditional page. If you think about a page, Google has kind of been forced, historically by their infrastructure, to surface pages and to rank pages. But pages sometimes struggle to rank if they have too many topics on them.
So for instance, what I’ve shown you here is a page about vegetables. This page may be the best page about vegetables, and it may have the best information about lettuce, celery, and radishes. But because it’s got those topics and maybe more topics on it, they all kind of dilute each other, and this great page may struggle to rank because it’s not focused on the one topic, on one thing at a time.
Google wants to rank the best things. But historically they’ve kind of pushed us to put the best things on one page at a time and to break them out. So what that’s created is this “content is king, I need more content, build more pages” mentality in SEO. The problem is everyone can be building more and more pages for every keyword that they want to rank for or every keyword group that they want to rank for, but only one is going to rank number one.
Google still has to crawl all of those pages that it told us to build, and that creates this character over here, I think, Marjory the Trash Heap, which if you remember the Fraggles, Marjory the Trash Heap was the all-knowing oracle. But when we’re all creating kind of low- to mid-quality content just to have a separate page for every topic, then that makes Google’s life harder, and that of course makes our life harder.
So why are we doing all of this work? The answer is because Google can only index pages, and if the page is too long or too many topics, Google gets confused. So we’ve been enabling Google to do this. But let’s pretend, go with me on this, because this is a theory, I can’t prove it. But if Google didn’t have to index a full page or wasn’t locked into that and could just index a piece of a page, then that makes it easier for Google to understand the relationships of different topics to one page, but also to organize the bits of the page to different pieces of the Knowledge Graph.
So this page about vegetables could be indexed and organized under the vegetable node of the Knowledge Graph. But that doesn’t mean that the lettuce part of the page couldn’t be indexed separately under the lettuce portion of the Knowledge Graph and so on, celery to celery and radish to radish. Now I know this is novel, and it’s hard to think about if you’ve been doing SEO for a long time.
But let’s think about why Google would want to do this. Google has been moving towards all of these new kinds of search experiences where we have voice search, we have the Google Home Hub kind of situation with a screen, or we have mobile searches. If you think about what Google has been doing, we’ve seen the increase in people also ask, and we’ve seen the increase in featured snippets.
They’ve actually been kind of, sort of making fragments for a long time or indexing fragments and showing them in featured snippets. The difference between that and fraggles is that when you click through on a fraggle, when it ranks in a search result, Google scrolls to that portion of the page automatically. That’s the handle portion.
So handles you may have heard of before. They’re kind of old-school web building. We call them bookmarks, anchor links, anchor jump links, stuff like that. It’s when it automatically scrolls to the right portion of the page. But what we’ve seen with fraggles is Google is lifting bits of text, and when you click on it, they’re scrolling directly to that piece of text on a page.
So we see this already happening in some results. What’s interesting is Google is overlaying the link. You don’t have to program the jump link in there. Google actually finds it and puts it there for you. So Google is already doing this, especially with AMP featured snippets. If you have a AMP featured snippet, so a featured snippet that’s lifted from an AMP page, when you click through, Google is actually scrolling and highlighting the featured snippet so that you could read it in context on the page.
But it’s also happening in other kind of more nuanced situations, especially with forums and conversations where they can pick a best answer. The difference between a fraggle and something like a jump link is that Google is overlaying the scrolling portion. The difference between a fraggle and a site link is site links link to other pages, and fraggles, they’re linking to multiple pieces of the same long page.
So we want to avoid continuing to build up low-quality or mid-quality pages that might go to Marjory the Trash Heap. We want to start thinking in terms of can Google find and identify the right portion of the page about a specific topic, and are these topics related enough that they’ll be understood when indexing them towards the Knowledge Graph.
Knowledge Graph build-out into different areas
So I personally think that we’re seeing the build-out of the Knowledge Graph in a lot of different things. I think featured snippets are kind of facts or ideas that are looking for a home or validation in the Knowledge Graph. People also ask seem to be the related nodes. People also search for, same thing. Related searches, same thing. Featured snippets, oh, they’re on there twice, two featured snippets. Found on the web, which is another way where Google is putting expanders by topic and then giving you a carousel of featured snippets to click through on.
So we’re seeing all of those things, and some SEOs are getting kind of upset that Google is lifting so much content and putting it in the search results and that you’re not getting the click. We know that 61% of mobile searches don’t get a click anymore, and it’s because people are finding the information that they want directly in a SERP.
That’s tough for SEOs, but great for Google because it means Google is providing exactly what the user wants. So they’re probably going to continue to do this. I think that SEOs are going to change their minds and they’re going to want to be in those windowed content, in the lifted content, because when Google starts doing this kind of thing for the native apps, databases, and other content, websites, podcasts, stuff like that, then those are new competitors that you didn’t have to deal with when it was only websites ranking, but those are going to be more engaging kinds of content that Google will be showing or lifting and showing in a SERP even if they don’t have to have URLs, because Google can just window them and show them.
So you’d rather be lifted than not shown at all. So that’s it for me and featured snippets. I’d love to answer your questions in the comments, and thanks very much. I hope you like the theory about fraggles.
While SEOs have been doubling-down on content and quality signals for their websites, Google was building the foundation of a new reality for crawling — indexing and ranking. Though many believe deep in their hearts that “Content is King,” the reality is that Mobile-First Indexing enables a new kind of search result. This search result focuses on surfacing and re-publishing content in ways that feed Google’s cross-device monetization opportunities better than simple websites ever could.
For two years, Google honed and changed their messaging about Mobile-First Indexing, mostly de-emphasizing the risk that good, well-optimized, Responsive-Design sites would face. Instead, the search engine giant focused more on the use of the Smartphone bot for indexing, which led to an emphasis on the importance of matching SEO-relevant site assets between desktop and mobile versions (or renderings) of a page. Things got a bit tricky when Google had to explain that the Mobile-First Indexing process would not necessarily be bad for desktop-oriented content, but all of Google’s shifting and positioning eventually validated my long-stated belief: That Mobile-First Indexing is not really about mobile phones, per se, but mobile content.
I would like to propose an alternative to the predominant view, a speculative theory, about what has been going on with Google in the past two years, and it is the thesis of my 2019 MozCon talk — something we are calling Fraggles and Fraggle-based Indexing.
I’ll go through Fraggles and Fraggle-based indexing, and how this new method of indexing has made web content more ‘liftable’ for Google. I’ll also outline how Fraggles impact the Search Results Pages (SERPs), and why it fits with Google’s promotion of Progressive Web Apps. Next, I will provide information about how astute SEO’s can adapt their understanding of SEO and leverage Fraggles and Fraggle-Based Indexing to meet the needs of their clients and companies. Finally, I’ll go over the implications that this new method of indexing will have on Google’s monetization and technology strategy as a whole.
Ready? Let’s dive in.
Fraggles & Fraggle-based indexing
The SERP has changed in many ways. These changes can be thought of and discussed separately, but I believe that they are all part of a larger shift at Google. This shift includes “Entity-First Indexing” of crawled information around the existing structure of Google’s Knowledge Graph, and the concept of “Portable-prioritized Organization of Information,” which favors information that is easy to lift and re-present in Google’s properties — Google describes these two things together as “Mobile-First Indexing.”
Fraggles represent individual parts (fragments) of a page for which Google overlayed a “handle” or “jump-link” (aka named-anchor, bookmark, etc.) so that a click on the result takes the users directly to the part of the page where the relevant fragment of text is located. These Fraggles are then organized around the relevant nodes on the Knowledge Graph, so that the mapping of the relationships between different topics can be vetted, built-out, and maintained over time, but also so that the structure can be used and reused, internationally — even if different content is ranking.
More than one Fraggle can rank for a page, and the format can vary from a text-link with a “Jump to” label, an unlabeled text link, a site-link carousel, a site-link carousel with pictures, or occasionally horizontal or vertical expansion boxes for the different items on a page.
The easiest way for an SEO to think about a Fragment is within the example of an AJAX expansion box: The piece of text or information that is fetched from the server to populate the AJAX expander when clicked could be described as a Fragment. Alternatively, if it is indexed for Mobile-First Indexing, it is a Fraggle.
We have also recently discovered that Google has begun to index URLs with a # jump-link, after years of not doing so, and is reporting on them separately from the primary URL in Search Console. As you can see below from our data, they aren’t getting a lot of clicks, but they are getting impressions. This is likely because of the low average position.
Why index fragments & Fraggles?
If you’re used to thinking of rankings with the smallest increment being a URL, this idea can be hard to wrap your brain around. To help, consider this thought experiment: How useful would it be for Google to rank a page that gave detailed information about all different kinds of fruits and vegetables? It would be easy for a query like “fruits and vegetables,” that’s for sure. But if the query is changed to “lettuce” or “types of lettuce,” then the page would struggle to rank, even if it had the best, most authoritative information.
This is because the “lettuce” keywords would be diluted by all the other fruit and vegetable content. It would be more useful for Google to rank the part of the page that is about lettuce for queries related to lettuce, and the part of the page about radishes well for queries about radishes. But since users don’t want to scroll through the entire page of fruits and vegetables to find the information about the particular vegetable they searched for, Google prioritizes pages with keyword focus and density, as they relate to the query. Google will rarely rank long pages that covered multiple topics, even if they were more authoritative.
With featured snippets, AMP featured snippets, and Fraggles, it’s clear that Google can already find the important parts of a page that answers a specific question — they’ve actually been able to do this for a while. So, if Google can organize and index content like that, what would the benefit be in maintaining an index that was based only on per-pages statistics and ranking? Why would Google want to rank entire pages when they could rank just the best parts of pages that are most related to the query?
To address these concerns, historically, SEO’s have worked to break individual topics out into separate pages, with one page focused on each topic or keyword cluster. So, with our vegetable example, this would ensure that the lettuce page could rank for lettuce queries and the radish page could rank for radish queries. With each website creating a new page for every possible topic that they would like to rank for, there’s lot of redundant and repetitive work for webmasters. It also likely adds a lot of low-quality, unnecessary pages to the index. Realistically, how many individual pages on lettuce does the internet really need, and how would Google determine which one is the best? The fact is, Google wanted to shift to an algorithm that focused less on links and more on topical authority to surface only the best content — and Google circumvents this with the scrolling feature in Fraggles.
Even though the effort to switch to Fraggle-based indexing, and organize the information around the Knowledge Graph, was massive, the long-term benefits of the switch far out-pace the costs to Google because they make Google’s system for flexible, monetizable and sustainable, especially as the amount of information and the number of connected devices expands exponentially. It also helps Google identify, serve and monetize new cross-device search opportunities, as they continue to expand. This includes search results on TV’s, connected screens, and spoken results from connected speakers. A few relevant costs and benefits are outlined below for you to contemplate, keeping Google’s long-term perspective in mind:
Why Fraggles and Fraggle-based indexing are important for PWAs
What also makes the shift to Fraggle-based Indexing relevant to SEOs is how it fits in with Google’s championing of Progressive Web Apps or AMP Progressive Web Apps, (aka PWAs and PWA-AMP websites/web apps). These types of sites have become the core focus of Google’s Chrome Developer summits and other smaller Google conferences.
The answer is because PWA’s require ServiceWorkers, which uses Fraggles and Fraggle-based indexing to take the burden off crawling and indexing of complex web content.
ServiceWorkers and SEO
For a PWA to be indexed, Google requires webmasters to ‘register their app in Firebase,’ but they used to require webmasters to “register their ServiceWorker.” Firebase is the Google platform that allows webmasters to set up and manage indexing and deep linking for their native apps, chat-bots and, now, PWA’s.
Direct communication with a PWA specialist at Google a few years ago revealed that Google didn’t crawl the ServiceWorker itself, but crawled the API to the ServiceWorker. It’s likely that when webmasters register their ServiceWorker with Google, Google is actually creating an API to the ServiceWorker, so that the content can be quickly and easily indexed and cached on Google’s servers. Since Google has already launched an Indexing API and appears to now favor API’s over traditional crawling, we believe Google will begin pushing the use of ServiceWorkers to improve page speed, since they can be used on non-PWA sites, but this will actually be to help ease the burden on Google to crawl and index the content manually.
It’s important to remember that this is how AMP, Schema, and many other types of powerful SEO functionalities have started with a limited launch; beyond that, some great SEO’s have already tested submitting other types of content in the API and seen success. Submitting to APIs skips Google’s process of blindly crawling the web for new content and allows webmasters to feed the information to them directly.
It is possible that the new Indexing API follows a similar structure or process to PWA indexing. Submitted URLs can already get some kinds of content indexed or removed from Google’s index, usually in about an hour, and while it is only currently officially available for the two kinds of content, we expect it to be expanded broadly.
How will this impact SEO strategy?
Of course, every SEO wants to know how to leverage this speculative theory — how can we make the changes in Google to our benefit?
The first thing to do is take a good, long, honest look at a mobile search result. Position #1 in the organic rankings is just not what it used to be. There’s a ton of engaging content that is often pushing it down, but not counting as an organic ranking position in Search Console. This means that you may be maintaining all your organic rankings while also losing a massive amount of traffic to SERP features like Knowledge Graph results, Featured Snippets, Google My Business, maps, apps, Found on the Web, and other similar items that rank outside of the normal organic results.
These results, as well as Pay-per-Click results (PPC), are more impactful on mobile because they are stacked above organic rankings. Rather than being off to the side, as they might be in a desktop view of the search, they push organic rankings further down the results page. There has been some great reporting recently about the statistical and large-scale impact of changes to the SERP and how these changes have resulted in changes to user-behavior in search, especially from Dr. Pete Meyers, Rand Fishkin, and JumpTap.
Dr. Pete has focused on the increasing number of changes to the Google Algorithm recorded in his MozCast, which heated up at the end of 2016 when Google started working on Mobile-First Indexing, and again after it launched the Medic update in 2018.
Rand, on the other hand, focused on how the new types of rankings are pushing traditional organic results down, resulting in less traffic to websites, especially on mobile. All this great data from these two really set the stage for a fundamental shift in SEO strategy as it relates to Mobile-First Indexing.
The research shows that Google re-organized its index to suit a different presentation of information — especially if they are able to index that information around an entity-concept in the Knowledge Graph. Fraggle-based Indexing makes all of the information that Google crawls even more portable because it is intelligently nested among related Knowledge Graph nodes, which can be surfaced in a variety of different ways. Since Fraggle-based Indexing focuses more on the meaningful organization of data than it does on pages and URLs, the results are a more “windowed” presentation of the information in the SERP. SEOs need to understand that search results are now based on entities and use-cases (think micro-moments), instead of pages and domains.
Google’s Knowledge Graph
To really grasp how this new method of indexing will impact your SEO strategy, you first have to understand how Google’s Knowledge Graph works.
Since it is an actual “graph,” all Knowledge Graph entries (nodes) include both vertical and lateral relationships. For instance, an entry for “bread” can include lateral relationships to related topics like cheese, butter, and cake, but may also include vertical relationships like “standard ingredients in bread” or “types of bread.”
Lateral relationships can be thought of as related nodes on the Knowledge Graph, and hint at “Related Topics” whereas vertical relationships point to a broadening or narrowing of the topic; which hints at the most likely filters within a topic. In the case of bread, a vertical relationship-up would be topics like “baking,” and down would include topics like “flour” and other ingredients used to make bread, or “sourdough” and other specific types of bread.
SEOs should note that Knowledge Graph entries can now include an increasingly wide variety of filters and tabs that narrow the topic information to benefit different types of searcher intent. This includes things like helping searchers find videos, books, images, quotes, locations, but in the case of filters, it can be topic-specific and unpredictable (informed by active machine learning). This is the crux of Google’s goal with Fraggle-based Indexing: To be able to organize the information of the web-based on Knowledge Graph entries or nodes, otherwise discussed in SEO circles as “entities.”
Since the relationships of one entity to another remain the same, regardless of the language a person is speaking or searching in, the Knowledge Graph information is language-agnostic, and thus easily used for aggregation and machine learning in all languages at the same time. Using the Knowledge Graph as a cornerstone for indexing is, therefore, a much more useful and efficient means for Google to access and serve information in multiple languages for consumption and ranking around the world. In the long-term, it’s far superior to the previous method of indexing.
Examples of Fraggle-based indexing in the SERPs
Google has dramatically increased the number of Knowledge Graph entries and the categories and relationships within them. The build-out is especially prominent for topics for which Google has a high amount of structured data and information already. This includes topics like:
TV and Movies — from Google Play
Food and Recipe — from Recipe Schema, recipe AMP pages, and external food and nutrition databases
Science and medicine — from trusted sources (like WebMD)
Businesses — from Google My Business.
Google is adding more and more nodes and relationships to their graph and existing entries are also being built-out with more tabs and carousels to break a single topic into smaller, more granular topics or type of information.
As you can see below, the build-out of the Knowledge Graph has also added to the number of filters and drill-down options within many queries, even outside of the Knowledge Graph. This increase can be seen throughout all of the Google properties, including Google My Business and Shopping, both of which we believe are now sections of the Knowledge Graph:
Other similar examples include the additional filters and “Related Topics” results in Google Images, which we also believe to represent nodes on the Knowledge Graph:
The Knowedge Graph is also being presented in a variety of different ways. Sometimes there’s a sticky navigation that persists at the top of the SERP, as seen in many media-oriented queries, and sometimes it’s broken up to show different information throughout the SERP, as you may have noticed in many of the local business-oriented search results, both shown below.
Since the launch of Fraggle-based indexing is essentially a major Knowledge Graph build-out, Knowledge Graph results have also begun including more engaging content which makes it even less likely that users will click through to a website. Assets like playable video and audio, live sports scores, and location-specific information such as transportation information and TV time-tables can all be accessed directly in the search results. There’s more to the story, though.
Companies who want to leverage the Knowledge Graph should take every opportunity to create your own assets, like AR models and AMP Stories, so that Google will have no reason to do it. Beyond that, companies should submit accurate information directly to Google whenever they can. The easiest way to do this is through Google My Business (GMB). Whatever types of information are requested in GMB should be added or uploaded. If Google Posts are available in your business category, you should be doing Posts regularly, and making sure that they link back to your site with a call to action. If you have videos or photos that are relevant for your company, upload them to GMB. Start to think of GMB as a social network or newsletter — any assets that are shared on Facebook or Twitter can also be shared on Google Posts, or at least uploaded to the GMB account.
You should also investigate the current Knowledge Graph entries that are related to your industry, and work to become associated with recognized companies or entities in that industry. This could be from links or citations on the entity websites, but it can also include being linked by third-party lists that give industry-specific advice and recommendations, such as being listed among the top competitors in your industry (“Best Plumbers in Denver,” “Best Shoe Deals on the Web,” or “Top 15 Best Reality TV Shows”). Links from these posts also help but are not required — especially if you can get your company name on enough lists with the other top players. Verify that any links or citations from authoritative third-party sites like Wikipedia, Better Business Bureau, industry directories, and lists are all pointing to live, active, relevant pages on the site, and not going through a 301 redirect.
While this is just speculation and not a proven SEO strategy, you might also want to make sure that your domain is correctly classified in Google’s records by checking the industries that it is associated with. You can do so in Google’s MarketFinder tool. Make updates or recommend new categories as necessary. Then, look into the filters and relationships that are given as part of Knowledge Graph entries and make sure you are using the topic and filter words as keywords on your site.
Featured Snippets or “Answers” first surfaced in 2014 and have also expanded quite a bit, as shown in the graph below. It is useful to think of Featured Snippets as rogue facts, ideas or concepts that don’t have a full Knowledge Graph result, though they might actually be associated with certain existing nodes on the Knowledge Graph (or they could be in the vetting process for eventual Knowledge Graph build-out).
Featured Snippets seem to surface when the information comes from a source that Google does not have an incredibly high level of trust for, like it does for Wikipedia, and often they come from third party sites that may or may not have a monetary interest in the topic — something that makes Google want to vet the information more thoroughly and may prevent Google from using it, if a less bias option is available.
Like the Knowledge Graph, Featured Snippets results have grown very rapidly in the past year or so, and have also begun to include carousels — something that Rob Bucci writes about extensively here. We believe that these carousels represent potentially related topics that Google knows about from the Knowledge Graph. Featured Snippets now look even more like mini-Knowledge Graph entries: Carousels appear to include both lateral and vertically related topics, and their appearance and maintenance seem to be driven by click volume and subsequent searches. However, this may also be influenced by aggregated engagement data for People Also Ask and Related Search data.
The build-out of Featured Snippets has been so aggressive that sometimes the answers that Google lifts are obviously wrong, as you can see in the example image below. It is also important to understand that Featured Snippet results can change from location to location and are not language-agnostic, and thus, are not translated to match the Search Language or the Phone Language settings. Google also does not hold themselves to any standard of consistency, so one Featured Snippet for one query might present an answer one way, and a similar query for the same fact could present a Featured Snippet with slightly different information. For instance, a query for “how long to boil an egg” could result in an answer that says “5 minutes” and a different query for “how to make a hard-boiled egg” could result in an answer that says “boil for 1 minute, and leave the egg in the water until it is back to room temperature.”
The data below was collected by Moz and represents an average of roughly 10,000 that skews slightly towards ‘head’ terms.
SEO strategy for featured snippets
All of the standard recommendations for driving Featured Snippets apply here. This includes making sure that you keep the information that you are trying to get ranked in a Featured Snippet clear, direct, and within the recommended character count. It also includes using simple tables, ordered lists, and bullets to make the data easier to consume, as well as modeling your content after existing Featured Snippet results in your industry.
This is still speculative, but it seems likely that the inclusion of Speakable Schema markup for things like “How To,” “FAQ,” and “Q&A” may also drive Featured Snippets. These kinds of results are specially designated as content that works well in a voice-search. Since Google has been adamant that there is not more than one index, and Google is heavily focused on improving voice-results from Google Assistant devices, anything that could be a good result in the Google Assistant, and ranks well, might also have a stronger chance at ranking in a Featured Snippet.
People Also Ask & Related Searches
Finally, the increased occurrence of “Related Searches” as well as the inclusion of People Also Ask (PAA) questions, just below most Knowledge Graph and Featured Snippet results, is undeniable. The Earl Tea screenshot shows that PAA’s along with Interesting Finds are both part of the Knowledge Graph too.
The graph below shows the steady increase in PAA’s. PAA results appear to be an expansion of Featured Snippets because once expanded, the answer to the question is displayed, with the citation below it. Similarly, some Related Search results also now include a result that looks like a Featured Snippet, instead of simply linking over to a different search result. You can now find ‘Related Searches’ throughout the SERP, often as part of a Knowledge Graph results, but sometimes also in a carousel in the middle of the SERP, and always at the bottom of the SERP — sometimes with images and expansion buttons to surface Featured Snippets within the Related Search results directly in the existing SERP.
Boxes with Related Searches are now also included with Image Search results. It’s interesting to note that Related Search results in Google Images started surfacing at the same time that Google began translating image Title Tags and Alt Tags. It coincides well with the concept that Entity-First Indexing, that Entities and Knowledge Graph are language-agnostic, and that Related Searches are somehow related to the Knowledge Graph.
This data was collected by Moz and represents an average of roughly 10,000 that skews slightly towards ‘head’ terms.
SEO STRATEGY for PAA and related searches
Since PAAs and some Related Searches now appear to simply include Featured Snippets, driving Featured Snippet results for your site is also a strong strategy here. It often appears that PAA results include at least two versions of the same question, re-stated with a different language, before including questions that are more related to lateral and vertical nodes on the Knowledge Graph. If you include information on your site that Google thinks is related to the topic, based on Related Searches and PAA questions, it could help make your site appear relevant and authoritative.
Finally, it is crucial to remember that you don’t have a website to rank in Google now and SEO’s should consider non-website rankings as part of their job too.
If a business doesn’t have a website, or if you just want to cover all the bases, you can let Google host your content directly — in as many places as possible. We have seen that Google-hosted content generally seems to get preferential treatment in Google search results and Google Discover, especially when compared to the decreasing traffic from traditional organic results. Google is now heavily focused on surfacing multimedia content, so anything that you might have previously created a new page on your website for should now be considered for a video.
Google My Business (GMB) is great for companies that don’t have websites, or that want to host their websites directly with Google. YouTube is great for videos, TV, video-podcasts, clips, animations, and tutorials. If you have an app, a book, an audio-book, a podcast, a movie, TV show, class or music, or PWA, you can submit that directly to GooglePlay (much of the video content in GooglePlay is now cross-populated in YouTube and YouTube TV, but this is not necessarily true of the other assets). This strategy could also include books in Google Books, flights in Google Flights, Hotels in Google Hotel listings, and attractions in Google Explore. It also includes having valid AMP code, since Google hosts AMP content, and includes Google News if your site is an approved provider of news.
Changes to SEO tracking for Fraggle-based indexing
The biggest problem for SEOs is the missing organic traffic, but it is also the fact that current methods of tracking organic results generally don’t show whether things like Knowledge Graph, Featured Snippets, PAA, Found on the Web, or other types of results are appearing at the top of the query or somewhere above your organic result. Position one in organic results is not what it used to be, nor is anything below it, so you can’t expect those rankings to drive the same traffic. If Google is going to be lifting and representing everyone’s content, the traffic will never arrive at the site and SEOs won’t know if their efforts are still returning the same monetary value. This problem is especially poignant for publishers, who have only been able to sell advertising on their websites based on the expected traffic that the website could drive.
The other thing to remember is that results differ — especially on mobile, which varies from device to device (generally based on screen size) but also can vary based on the phone IOS. They can also change significantly based on the location or the language settings of the phone, and they definitely do not always match with desktop results for the same query. Most SEO’s don’t know much about the reality of their mobile search results because most SEO reporting tools still focus heavily on desktop results, even though Google has switched to Mobile-First.
As well, SEO tools generally only report on rankings from one location — the location of their servers — rather than being able to test from different locations.
The only thing that good SEO’s can do to address this problem is to use tools like the MobileMoxie SERP Test to check what rankings look like on top keywords from all the locations where their users may be searching. While the free tool only provides results with one location at a time, subscribers can test search results in multiple locations, based on a service-area radius or based on an uploaded CSV of addresses. The tool has integrations with Google Sheets, and a connector with Data Studio, to help with SEO reporting, but APIs are also available, for deeper integrations in content editing tools, dashboards and for use within other SEO tools.
At MozCon 2017, I expressed my belief that the impact of Mobile-First Indexing requires a re-interpretation of the words “Mobile,” “First,” and “Indexing.” Re-defined in the context of Mobile-First Indexing, the words should be understood to mean “portable,” “preferred,” and “organization of information.” The potential of a shift to Fraggle-based indexing and the recent changes to the SERPs, especially in the past year, certainly seems to prove the accuracy of this theory. And though they have been in the works for more than two years, the changes to the SERP now seem to be rolling-out faster and are making the SERP unrecognizable from what it was only three or four years ago.
SEOs need to consider the opportunities and change the way we view our overall indexing strategy, and our jobs as a whole. If Google is organizing the index around the Knowledge Graph, that makes it much easier for Google to constantly mention near-by nodes of the Knowledge Graph in “Related Searches” carousels, links from the Knowledge Graph, and topics in PAAs. It might also make it easier to believe that featured snippets are simply pieces of information being vetted (via Google’s click-crowdsourcing) for inclusion or reference in the Knowledge Graph.
Fraggles and Fraggled indexing re-frames the switch to Mobile-First Indexing, which means that SEOs and SEO tool companies need to start thinking mobile-first — i.e. the portability of their information. While it is likely that pages and domains still carry strong ranking signals, the changes in the SERP all seem to focus less on entire pages, and more on pieces of pages, similar to the ones surfaced in Featured Snippets, PAAs, and some Related Searches. If Google focuses more on windowing content and being an “answer engine” instead of a “search engine,” then this fits well with their stated identity, and their desire to build a more efficient, sustainable, international engine.
SEOs also need to find ways to serve their users better, by focusing more on the reality of the mobile SERP, and how much it can vary for real users. While Google may not call the smallest rankable units Fraggles, it is what we call them, and we think they are critical to the future of SEO.