Jason Barnard does the SEOisAEO webinar (now a podcast.) He has done 60+ interviews with very smart people in the industry, including myself of course ;), and has covered various SEO topics. The AEO bit stands for Answer Engine Optimisation so it’s that idea of how do we empower Google to answer those questions that users have.
In this episode, I speak to Jason about his latest article that he wrote on Search Engine Journal around How Google Search Rankings Works. The idea was that instant answers had a different ranking algorithm. In a conference in Australia, Jason asked Gary Illyes whether there is a different algorithm for featured snippets. Gary said ‘no’, to which Jason’s face dropped as he had a theory, along with Eric Enge, Cindy Krum and Hannah Thorpe, that it did. Don’t worry though Gary goes on to then explain how it might actually work so keep reading/listening…
Firstly, we discuss the complexities of a customer journey. Kevin Gibbons for example says (in episode 8: Future of Search,) that we might not even need websites but it’ll still be about optimising the channels to grab your users attention in one way or another.
Final steps of the journey, using other platforms first, such as social media, and approaching digital marketing is a way to communicate to customers, where people you’re trying to get will be hanging out, they’ll be ready to convert once they get to the website. Move away from the website as default place to communicate.
So, back to featured snippets. We, as the unreasonable consumer want the answer right there in front of us and fast. Google provides this but seemingly different to the classic links we see. What is it then if the algorithm is the same for both?
Jason states that it’s a parallel algorithm that has different ranking factors and weightings.
One of them, is almost certainly the knowledge graph. Someone in Google wanted the Knowledge Graph to be a ranking factor in Google and walked out as this wasn’t implemented. However, the 10 blue links are the basic foundation blocks of Google, on which everything else is built. It’s the stability of the platform and therefore they’re unlikely to change this too dramatically. They can throw their whole business model out of the window due to the advertising model collapsing so can’t mess the fundamental functions and insert major changes to the way it functions. So, this change, although seemingly may improve search results, wasn’t taken on board.
Gary Illyes mentions the top ranking factors to be:
Each one has a score so it’s a bidding system of value. For example ‘My bid value is this’. The important thing is that it multiplies it together, so if any drops under 1, the bid moves from thousands to hundreds very quickly. It’s better to be a straight C student than mostly A’s then an F as the latter is likely to kill your bid.
No point as to concentrating on one particular ranking factor but all need to be above that ‘1’ in order to improve that multiplication. Weightings of rankings change constantly Jason confirms.
Tip: The rankings factors within a particular situation are going to be different. A factor can be important in one instance and not another.
After the bidding, explained above, there is a second algorithm which refines and sorts the top results. It tends to give a negative effect, meaning it’s to check what has got to the top which shouldn’t have and so results can score a final rank of ‘0’ which means they’ll no longer be on that page.
Then we have the same system on ‘candidate result sets’ such as rich results like a featured snippet. If they can outbid the top one of the ten blue links then they get a place because they have more value and deserve a place at the top as they serve the query better, thus it takes that top stop and one of the blue links drops off. Same with video etc.
[International tip] So there are less French featured snippets and why? Jason explains that it’s because there isn’t enough content that’s of good enough value to deserve that space. They’re not sufficiently strong to outbid the well-established blue links.
If we push this to an extreme, for certain queries, Google has become a media page like Yahoo but per query, contextual, time-based basis. Algorithmically generated; something that Jason calls Google ‘Darwinism’.
I’ve read a patent around an algorithm that reconfigures the algorithm (so an extra layer) on top of what has already been chosen. Annoyingly I can’t find the patent but I’ll add it in if I ever do. Jason said that he also theorised an algorithm based on a topic layer. Although it’s not true, it may be in the future because the idea is based on results that are dependant on the ‘level’ of where you are in something. For example if you’re searching for how to learn an instrument, you get one set of results but as you progress you’ll also get a different answer to your query.
It’s another form of personalisation as Google is already doing this in a sophisticated way, showing results depending on the intent, device and location etc.
Everything you need to do needs to focus on:
Deliverability. Make sure you make your content as easy as possible to understand. Add schema markup, tables, lists etc. Check site speed, use correct format and so on.
Accessibility. Have you used alt tags on imagery for example.
Credibility. With EAT (Expertise, Authority, Trust) being so important you need to ensure you’re trustworthy.
You need all three and in that order too. It’s a very simple approach but a great strategy.
Music credit: I Dunno (Grapes of Wrath Mix) by spinningmerkaba (c) copyright 2017 Licensed under a Creative Commons Attribution (3.0) license. http://dig.ccmixter.org/files/jlbrock44/56346 Ft: Jlang, 4nsic, grapes.