SEO Predictions 2020 (Australia)
- Search Engine Marketing in 2019: A Quick Review
- Zero-Click SERPs
- Google Algorithm Updates During 2019
- Creating Clicks in a Zero Click Search Landscape
- Will SEO Become Automated?
- The Influence of AI on Google and on Marketers
- SEO for Voice Search Queries
- Final SEO Predictions for 2020
Peter: We're live across Australia, across the UK, and around the world. Wow, it's that time of year again: SEO predictions for 2020. This is the Australian version. Thanks very much to SEMrush for allowing us this platform.
Here's the episode where we talk about did 2019 predictions come true? What kind of Google roller coaster ride will we see in 2020? To help me answer these questions, I have three amazing search marketing professionals. Ric, Dan and Nik, and I will go through and introduce these guests.
Ric joining us from the UK. It's early in the morning. He's an SEO consultant at Yext. Ric has worked agency side for the past six years with brands including L'Oreal, Regus, Santander Bank, and others. He's a frequent conference speaker and a writer.
I'll go on and introduce Dan, managing director of DEJAN Marketing and a futurist at heart who loves to predict future events. Dan is also known for his SEO experiments and many years of speaking at various industry events around the globe. He's focusing on search optimization, CTR optimization, web content, link building, and creative SEO. Welcome, Dan. Thank you so much for joining us.
Nik is an SEO specialist at StudioHawk, overseeing SEO strategy for large enterprises and local businesses. Nik's interested in site structure, SERP watching, structured data…
Search Engine Marketing in 2019: A Quick Review
I'm going to start the first question and I'm going to ask you first of all, Ric, if I can? At the beginning of this year, you may have had some certain ideas in mind of what a year ahead people were talking, rumors, predictions, that kind of thing. Do you think much of this came true?
Ric: I believe that search has headed in the direction I thought it would. I wrote a piece for The Drum in sort of tail end of 2018 talking about why Google needs to have the answers in it. It focused on the changing search landscape, Google moving from keyword focus search to longer strings to questions and asking direct asks in search.
We've seen really in 2019 is that Google has moved further and further and further forward with that agenda and pushing more direct information search results, pushing people away from websites in some instances.
I'm sure it will be something we'll talk a lot about in the rest of this webinar, but moving to this knowledge-based future.
Peter: Moving away from keywords, right? Dan, is this what you thought would happen in 2019 moving away from the old ways of SEO optimizing for keywords?
Dan: Well, I wouldn't say it's not about the keywords. That's just a little more new weird keywords that we haven't considered in full. A lot of the changes that took place didn't affect SEO specifically or how SEO is done, it affected how Google presents new information to people who are searching for random things in different ways.
They found new ways to connect potential information with the person searching for it that we haven't considered before, to the point where I find Google useless in certain instances and I go to Bing to find the results that I want when I want a very specific answer to my query.
Exact match queries are almost useless at the moment because Google tries to hone you in on a certain direction that it thinks you want to go into. And if that direction is wrong, you need to completely rephrase your query. So, a lot of weird things are going with Google at the moment, particularly because they'll let the algorithm decide a lot more.
I don't think Google's results are as robust as accurate as before, but they're just a lot more diverse and a lot more random. Something tells me that they're using a lot less resources than before. I think they've optimized their systems to make a lot more guesswork than actual results. I think that they're saving a lot of money that way.
Peter: They're getting a lot of repeat searches, right? I mean, the more searches you do, the more you're using Google, right?
Dan: Well, one thing that really wasn't a surprise for me, and I predicted it, is that Google wants you to stay at Google. They don't want you to click anywhere. The zero-click results, the zero-click searches and something that I wrote about a couple of months ago, Google refers to as the Good Abandonment. Basically we're facing the zero-click and Good Abandonment era. That's going to just, I think, become more prominent in the next year.
Peter: Yeah, this is really interesting. Nik, can I bring you in now because this is seriously becoming a trend is quite obvious is the zero-click SERPs, right?
The results that we're seeing, we don't need to click on it, it's just there for us. Also this combination of zero-click results, but also we do a search, the kind of result we get, it just forces us to search again, is this a good thing? I mean, is it good for Google? Is it good for us as SEO people? But what about is it good for the general public doing their searching?
Nik: I kind of feel like ultimately Google really wants to be able to serve users a lot better and making a lot of improvements that they're hoping that with the long game that this is going to all kind of even itself out. I do agree with Dan, I find that a lot of broad match keywords, a lot of people aren't finding their intent as well as that they might've had.
I know that a lot of people towards the start of this year was like, "Yeah, this is definitely going to be something that we're going to have to look out for." Rand Fishkin did some really interesting research, specifically about that where for zero-click searches, there was a massive increase in impressions and things like that through search console because Google was just kind of like eating up this SERP.
I think a myth that I'd like to also maybe even bring up right now is that position zero is good for brands. It's to be able to just be seen, but a lot of the time you might see a link, but a lot of the time you get a lot of impression increases, but not a lot of CTR. I know in some instances people definitely see positions zero working for them. But I think the general chatter is overall that hasn't really helped a lot of businesses and a lot of brands.
Google Algorithm Updates During 2019
Peter: But let's just maybe change tact a bit here. I want to start talking about some of the penalties. I mean, this year in 2019 we've seen a lot of penalties... When I say penalties, I mean we've seen algorithm updates and some people suffered and some people benefited from that. In general, the people who got hit considered themselves to be penalized, right?
Some of them couldn't really work out what they did wrong. Some of the communication from the official channels was quite a bit blurry too. What do you think, Ric? Were they hard-hitting? Was it unfair? Should we see more of this next year?
Ric: Ultimately, an algorithm is just a set of instructions that facilitate actions, right? It's the instructions for the machine learning models to continuously learn and to learn by. In that regard, sure, there have been lots of core updates as Google tune them.
Core updates, as you quite rightly say, there's very little information around them. Google doesn't say, "Hey, we focused on this one thing and this is what you need to do to start killing it now."
It is broad, but I think that's purely because there are so many factors and so many things going on with ranking that actually all they're doing is changing a criteria or releasing a new model into the wild, they've been testing on the side that is getting better quality results than the one before.
Does this surprise me? No, I think this has been how it's worked for quite some time. Do I expect to see more of it next year? Yes. I think we'll see a lot more of this next year. But I think just like Penguin became completely real-time, just like Panda became real-time. I think these kind of core updates will be more ongoing rather than as a sort of jittery as they have been.
Peter: Dan, what do you think about the algorithms? Obviously, a lot of SEOs, they optimize for the algorithm, right? Which I'm not a fan of that myself but again, even if we're doing all of our things, sometimes you can lose...when an algorithm update is done.
Dan: Well, I don't think we should completely ignore when something changes at Google. It's our duty towards our clients to be on top of it. That's why I created Algoroo, because it started off as an internal set of metrics that we use and we thought, "Well, let's make it public to give the Australian side of what's going on rather than just having MozCast and the others."
When something happens with one of my clients, the first question I ask myself, "Is it something that we did or the client did? Or is it something that's global, a global event?" That's an example when it's useful to chase algorithms. We just want to make sure that we can say, "Okay, it's a global event. Something happened on an algorithm level." There’s not much you can do about that.
If there isn't something that's showing up on the trackers, then you go and have a look at your onsite stuff to see if you've messed up something.
What you often see is that there are updates that fly under the radar. It's not just when the announcements happen, you should be looking at the trackers and going, "Ah, something happened there, I should check my client's rankings." In terms of the jargon of penalties, not even Google uses that term. They say manual actions and they say algorithmic changes.
Peter: I do want to hear Nik's take on some of these algorithm updates, which seem to be a lot to do with the same kind of things, they always have been, but maybe tougher on the links and maybe tougher on the EAT sort of the stuff that I've seen the trends of.
Nik: We saw a lot of volatility in the SERPs and with rankings and with organic traffic with a lot of our clients with the June 2019 core updates. A lot of sites with reciprocal link schemes, especially around about September they had a lot of issues. There's a lot of digging and scratching into the background profile of those just a sort of weed it out and try and figure out what was going on.
I think quite honestly machine learning is having a massive impact in the SERP, and if you really want to make sure that you're going to be fine for 2020, I really think that all the traditional factors like EAT, the Expertise, Authority and Trustworthiness, but also quality, let's not forget quality, is definitely something that we should really can definitely make sure that we're seriously adding to the logic of why we do things when we're looking to optimize our sites for 2020.
Really wanting to make sure that queries and things like that will be much better understood. I think there's going to be a lot more scrutiny on black hat, gray hat kind of techniques.
Peter: That's a pretty in-depth answer. Thanks for that, Nik. I've got a question for you, Ric, but do you have any more comments about the algorithms before we move onto the next topic?
Ric: I'm going to touch on what Nik was saying around links. I've gone backwards and forwards as to whether I think links are important and not important, and are they important?
Regardless of how I've tried to disprove the value of links multiple times, purely because I want to understand a bit more about what the influence they have is. Every time I've done this, I've hit a bit of a dead-end purely because links are ultimately the lifeblood of the Internet.
I was going to say the change with the different tagging around links, to me it's very much seems like they need better quality data. If we start tagging things as sponsored or as UGC and start giving them more context to the links that they exist, they can feed that into their systems, their algorithms and the machine learning processes to actually better classify links they discover in the wild that maybe haven't been tagged as X, Y or Z.
Creating Clicks in a Zero Click Search Landscape
Peter: Maybe, Dan, I'll bring you into this next question because we've had one, "What is the best way to set a website up for the result for zero search?" I mean, people want position zero, right? I'm going to mix this up a little bit with the other topic here, which is we've been thinking about, which is the zero-click result. Is this right? I mean, does position zero also kind of double as a zero-click?
Dan: Couple of things that are very interesting at the moment is that position number one and position number zero, the featured, snippets, it's not necessarily a major traffic driver. There's increasing evidence and some of my own research that supports this, as well as Nielsen's research, that tells us that people's eyes dot around the SERP page very randomly these days. Nielsen refers to as a pinball effect, and Google in their own research talks about it as a nonlinear search pattern behavior.
What does nonlinear mean? Well in the traditional sense, you have the top 10 results, and then you go number one, number two, number three and then 10 blue links and the good old days. What was nice about that is that you can easily predict the likelihood of each result being clicked on based on its position bias. It's just statistical stuff. That helped Google normalize their results when they evaluate the popularity of each of the 10 results on the page.
For example, if a significant amount of people clicks on fifth result, once you iron out all the bias and when you compare to the results one through to four, that result suddenly becomes very interesting to Google. That's how they analyze what's interesting.
I think webmasters now have an opportunity regardless of whether the results zero or if you're facing a query that provides zero clicks, you face an opportunity now to create the clicks, if they weren't going to happen. It's like writing an ad copy.
People put so much time and effort in Google Ads, find crafting the titles and descriptions and the ad copy and use AB testing. Nobody, almost nobody, does it on SERPs on the organic side. Why? I don't know. Because it's free I guess, but it's not free because we invest so much money into it.
A/B testing or sequential testing of SERPs, regardless of where they are, and it's a whole new topic. I'm not going to take too much time, but it's a topic of using power words and little hooks and little catchphrases to lure people in.
A lot of my own research shows that if you give an answer in your meta description and provide a reason for in-depth, so you give the answer and you say, "Why?" click through to find out more detail. You provide value... You're actually more likely to get the click.
Dan: It's our responsibility as SEOs to provide that hook to attract the click through. That's what's happening now. That's 2020 challenge. How do you make your stuff clickable? Because what's happening with the zero-click searches, it's all like free marketing for us. Even if people don't click through, they're seeing our brand. But we have a huge opportunity to cause the clicks to happen. If the click doesn't happen, we didn't do our job.
Ric: I completely agree that we want clicks to come to our website. We need people to stay and to convert. But actually, if people are asking you a question, the search results and you as a brand are not providing the answer, the customers can't find that information. Actually, they're not going to blame Google they can't find information, they could blame you as a brand for not putting it there.
People are being retrained to look for information in different ways, in more complicated ways. Voice has done a lot on that. We're going to see this more and more in 2020. I completely agree about getting the clicks. I think trying to drive clicks, however you can, to your site is important.
But I also think that if you're not focusing on how you're getting your information into Google's graph, into the knowledge base that it sits upon, and you're not answering those questions... Two things. One, your users aren't going get any information. Two, someone else probably will instead.
Nik: I'd like to just add one more thing there as well. I think whenever you see positions zero... they're getting a lot shorter and a lot more qualified. I also just think that they're doing a lot better at understanding the user's query and actually like delivering a little bit featured-like answer there and that's definitely something that I've seen.
But also too, if you get a featured snippet on mobile, for instance, that might not also translate when you go and sit back at your desk and on your desktop computer.
Will SEO Become Automated?
Peter: Look, I've got a very quick question while we've got you there, Nik. It's a short question, but it actually is maybe a little bit of a scary one for SEO people who make their money out of selling SEO. Martin Victor says, "Is SEO going to be automated?"
Nik: Like I was saying before, machine learning is definitely something that is impacting the SERPS a lot more. There's a lot of people that are going into that and trying to understand how Google has organized their... I'll try and remember the full name of it...Bidirectional Encoder Representations from Transformers (BERT).
There's a lot of people who are also saying, "Should you know Python?" There's mixed opinions about that. I myself personally don't know Python. Though it's an enormously interesting field of script to get into and to see the potential of what it can do. I think when it boils down to it, grit, logic and perseverance and trying to do that quite well is going to be your best tools for the future.
Ric: II think BERT is one of the largest changes in search for many, many, many years. I agree that in part that you cannot optimize for it. It's not designed to change ranking, it is designed to understand the query better so that the system that do change the ranking can know what to change the ranking towards.
BERT is a part of a system whereby Google is trying to really identify what people want, and that transcends the query, that goes throughout everything. What I think is really interesting is there's a stat that I read recently that 15% of searches are totally brand new. Globally, every year, every day, 15% of searches, I'll tell you, brand new, which if you think is something like 5.6 billion searches a day and puts around 800 million searches, but Google has never ever seen before.
To the point around automating SEO and can you automate it and what does this mean and what does this mean in the context of BERT and everything? I think it's not so much automation, but it's about data. It's about pushing information out into sources where search engines can find it.
This is great because in local SEO, we've been doing this for ages. We've had a direct interface with Google's knowledge graph of how to Google My Business. You put your name, address, phone number, opening hours, categories, et cetera, et cetera, into Google My Business.
What we're finding is with things like structured data and schema, I'm thinking sites that are deploying schema markup and for lots and lots of different things. They're able to push key information they want to be known for into the SERP, and that is what's now being surfaced when you look for these 15% of new queries because some of them are incredibly long.
BERT is almost part of the larger trends was this idea of entity-based search engines. As far as automation's concerned, I don't think we will automate because there's going to be an element of, "Well, we still need to write content. We still need to know what we want to be found for and we still want to know what we want to put out there." But we can start pushing to data feeds and we can stop presenting information in a way that makes sense to these systems that want to learn from us.
The Influence of AI on Google and on Marketers
Peter: Yes, so it's a big deal. AI becoming more and more prevalent. But Dan, just following on this topic about BERT, because obviously it's not just BERT and NLP, what about Hummingbird, any other, RankBrain, all the AI type of things, right?
Now, what about this theory or the concept that because of all of this that it has now made it more important for perhaps more people to sit around and work for the manual review team?
Dan: I don't think Google will ramp up the manual interventions anytime soon. That's definitely not going to happen. If anything, they're going to scale it down. First of all, it costs money to employ people and algorithms can just do everything faster and better than humans. They're not new into machine learning and they're just going to keep getting better and better.
The amount of research that they pump out is staggering. I'm personally not a big fan of what they've done with the recent... just letting the unsupervised algorithms loose and guessing what the user wants. I really don't like that. I feel like I'm being forced into a certain search factor, and if that doesn't suit me, it's hard to get out of it.
As much as I hate it personally, I think it benefits more users. It's just that for advanced users, if you know what you're after and you're using advanced search queries, it's becoming less and less useful. I prefer Bing to Google now when I need to do my very complicated search queries.
You mentioned automation. Google's automating, but are we? I think what we're going to see next year and moving forward every year more and more is we're going to see a lot more automation with content. Natural language generation is right, it's ready and there are platforms out there that go beyond Quill and all the others that are very expensive and in an enterprise level.
What we, as SEOs, can do and webmasters can do, we can start auto-generated content that's actually of good quality and driven by data. It's not going to apply to every industry, but it's... NLG is something to look out for, and if you're looking to automate, that's one thing.
Another thing that's easily automated is outreach through social media ads. Basically buying the presence of your content. If content is good enough, people earn its links. Some things that I'm doing on an automated level already, I've been doing it for a few years now, is automated link detection through analytics referrals because I'm earning my links, I'm not making them. I like to know when they come.
The next level we're planning to do after that is trying to find patterns. When something is consistently improving the CTR on your snippets, like presence of a dollar sign or absence of it, that will be alerted to you as a webmaster. You'd be like, "Oh, okay, well keep putting that symbol or that word more because it's improving your CTR."
One other thing that I'm doing already is automated event detection, whether it be onsite or global and annotating it within your traffic timeline. There's a lot that we can automate as SEOs, and you want to leave time for creative thought and you can't do that if you bogged down in manual work. I'd like to encourage everyone to automate a lot more so they can think creatively while machines are doing all the mundane work.
Peter: It's very creative, so this is creative SEO, right?
Dan: When you get the machines to do all the leg work for you, you have time to breathe and think creatively. Creative ideas don't just come to you straight away.
SEO for Voice Search Queries
Peter: Muncie is also from the UK, but she is asking, "Should we specifically optimize for voice search?" Do you have any thoughts about that? I'm throwing you right in the deep end there on the spot.
Nik: I think even at the start of 2019 voice sessions are definitely something that a lot of people were focusing on and still what of SEOs are still talking about voice search. I think that that's definitely something that will be coming a lot more interesting of a space as a natural language understanding, it's going to be working, and working through the search and being tested and retested and queries, hopefully will be getting a little bit more qualified.
As far as voice search for now, I mean, I like to use voice search all the time, but I still think that it's in its infancy and for questions like, "Oh, can you please play this song? Or can you please..." Kind of like VoiceGrid. They're still not well understood to be able to give you something that is deep and meaningful around your voice query.
I think that is very much like going to be filling out a lot of the structured data that you add to your sites. If you're marking up your pages and you're going to be doing that, that makes it statistically more likely that maybe an answer will be able to be read out like with voice search.
If it makes sense for your business that your target audience may be a little bit more adept at using voice search to make their queries, and I think that's something that you might consider, but for the future.
Ric: Think about a voice search, right? You ask your Alexa device, you ask your Google home device, a question. The first thing or the device does after answering the question is turn around to you and say, "Hey, was that the right answer? Did I answer it correctly?" Maybe it doesn't happen every time, but it happens a lot.
When Google or Alexa or Cortana asks you that question and you say yes or no, you've fundamentally changed the results of that query for yourself and for every other user moving forward because that's another data point they can use in that system.
How do you optimize for voice search? Well, I say you can't for that specific query in that specific moment of time because you are trying to optimize moving target, I don't think you can.
But what you can do, and this is where I think it comes on to the point around technical SEO and content, is create a knowledge base that sits underneath it. Build a knowledge of information around it so that search engines can query that knowledge and pull back the right information that is relevant to you or that you have that would be relevant to that user.
Technical SEO, I think, is always going to have a place. Making sure your site is usable, making sure it's user-friendly, making sure that it can be found by search engines. It's always going to be critically important because if you can't be found your information can't be distributed. If users have a terrible experience, they bounce, they don't want to convert.
Will content be king? I think content will change from being about content in and of itself. Just having blocks of texts to tell you things...to being split into two things. One, brand building, so content that drives people to be inspired, to be engaged, who wants to enjoy your brand in the same way that a TV ad does or an out of home advert does. I think the other side of that is the information, the knowledge side of the content, which is then linked to structured data and sort of links back into technical. That is a great recipe for optimizing for voice search. In fact, that's a great recipe for optimizing for search, full-stop.
Final SEO Predictions for 2020
Peter: I mean, we're out of time but I'll just go around to each guest one more time and I just want to ask what are your final thoughts about... What's your SEO prediction for 2020?
Nik: I think I'm just going to keep it really, really simple and just kind of boil it down to three things. Those three things being content: make sure that you're writing really, really good content, thinking about EAT, Expertise, Authority and Trust, and quality. Use heading tags and use structured data because of NLU, it's going to be understanding a lot more in conjunction with your structured data to be able to fill out that knowledge graph. I think links are still important. Based on what I was saying with the link graph, still having some reliance, and I also think that... Oh gosh, content, links, and...user experience.
Peter: Well, thank you Nik, and thanks for joining us. How do people stay in contact with you?
Nik: You can tweet at me @nikrangerseo or you can find me at studiohawk.com.au.
Ric: I think users will get used to searching for things in even more complex ways and the way they ask voice devices and other things. I think Google will get better at understanding what people are looking for and create new formats and new interesting ways to present information in search results.
I think 2020 is the year that everyone needs to get behind building knowledge, building data, building information that sits in Google's knowledge graph and making sure that they're answering the customers’ questions as opposed to someone else.
Actually 2020 is going to be the year that we start focusing on things like site search as well. Most websites right now, they don't offer that type of experience. Actually, if websites work to start offering that Google-like Bing-like search engine-like experience on their own site, would users bounce? Would users look in search results for them or would they start to go to the brand directs?
Dan: I'm going to make one simple clean prediction. Google is going to start deep indexing video content and audio content on the web. You'll be able to search for the very specific point within the video using a normal query. For example, you're going to type in a sentence that appears within this webinar, and you're just going to find this video in the results and go straight to that point where we say that thing.
Peter: Thanks for everyone who submitted questions, and thanks to SEMrush for making this platform available for us to share our ideas and we'll catch up with all of you very soon and most likely see you in 2020. Bye for now.
Nik: See you.
Ric: Thanks, everyone.
Nik: Thank you.
Peter: Thanks guys.