The latest Google update is here, and I wanted to present a few ideas to help you take advantage of it. BERT was trained on Wikipedia among others, using 2,500M words and now it’s here to help Google present better ‘question answering’ in the results.

Back after a long hiatus from writing anything SEO related, I’m back as this topic has really got me interested again in what Google is up to.

I’ll cover what can we, SEOs, do about this new BERT update and what we should be doing for our clients to future proof their rankings.

Just as a note, I’ve based these thoughts largely on Google’s Paper and github page, sprinkled by what I’ve seen ‘in the wild’ over the last few years. References at the end of the post.

What is BERT according to Google’s own page?

BERT is a method of pre-training language representations, meaning that we train a general-purpose “language understanding” model on a large text corpus (like Wikipedia), and then use that model for downstream NLP tasks that we care about (like question answering). BERT outperforms previous methods because it is the first unsupervised, deeply bidirectional system for pre-training NLP.

What is the BERT Google update for us, SEOs?

BERT Google update is a strong move towards more Natural Language Processing from Google, showing us, SEOs, that we can’t ignore the intent, context and semantics of any written content on our clients’ websites.

A few terms to understand first

Pre-training is a stage where BERT was essentially trained, Google taught BERT based on millions of sentences.

NLP = Natural Language Processing

OK, enough with the theory, what are the 5 actionables?

1 Google BERT is all about ‘question answering’ so give them questions and answers.

Add FAQs to your key pages, list questions and answers – mark them up using FAQ schema.

Here’s why Google cares: voice search and teaching BERT. Google cares most about ‘question answering’.

https://github.com/google-research/bert

Add in example of money supermarket page with FAQs.

https://www.moneysupermarket.com/breakdown-cover/

2 Optimising for entities and tokens, is your content ‘deep’ enough or just shallow?

Google is really going after sentence prediction and named entity recognition. The more entities you mention, the stronger ‘deep’ support will there be for your content.

Devlin, J., Chang, M.W., Lee, K. and Toutanova, K., 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.

Here’s an example of depth of content your competitors (best rankings ones) may have in their copy:

Let’s take ‘blue widgets’ as an example: they mention buying them, installing, maintaining and cleaning a ‘blue widget’ 3 times per week, but with a special blue widget oil diluted in water 5% for that extra blue widget shine, ensuring that the key part of the ‘blue widget’ = the blue widgetator is tuned to 10 degrees south when operating blue widget.

If in your text, you just say ‘we sell the best ‘blue widgets’ and the oil to clean them. And then you talk for 300 words about how your blue widgets are cleaner and shinier than theirs, it’s not gonna work. Google will look at their copy, analyse it based on entities, concepts and context they contain, understand that their copy is way DEEPER into the topic, provides answers to most common questions that happen during ‘blue widget’ buying journey, and rank theirs instead of yours.

So the whole notion of ‘you can’t improve your SEO to work together with Google’s algorithms’ is not true, yes we can improve it, it just takes a whole different approach to presenting data and helping users through your content.

3 Ensure that all your targeting is in the main copy

Not just in headings, tables or lists, it MAY BE that Google treats the main text data closer (sentences and paragraphs) than list items or tables or headings.

Devlin, J., Chang, M.W., Lee, K. and Toutanova, K., 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.

BERT was ‘pre-trained’ largely using ‘body copy’ – avoiding bullet lists and tables. My inclination is that Google bot may do similar things on your website. Although Direct Answers use tables and lists very often, I think they may not be included as a ranking factor itself.

4 Write some of your content Wikipedia-like

As per above screenshot from Google’s paper, BERT was taught largely using Wikipedia. Wikipedia is full of:

These are the kinds of aspects that I can already see playing a large part in helping pages rank, that we can take advantage of.

I’m not saying that every single page has to look like a Wikipedia page, but looking at some of best ranking sites in multiple industries – I can see how ‘heavy’ and meaningful their copy is, and how many facts and terms they present with good interlinking. This is not ‘brand new information’ its’ just another Google’s move in rewarding that approach.

5 Sentences, paragraphs and next steps (next sentences)

In addition to optimising for keywords and intents, we should optimise the copy for sentences, context and ‘next step’ that a user may want to take.

Google has multiple times mentioned that BERT is all about understanding sentences better.

If your copy only talks in short bullets, and doesn’t include ‘a flow of thought’ – it may perform weaker.

The analysis also talks a lot about ‘next sentence prediction’.

Devlin, J., Chang, M.W., Lee, K. and Toutanova, K., 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.

I hope you enjoyed this read and that you can and will apply some of these ideas.

If you’ve any need for a SEO consultant (in London or anywhere else) drop me a message and let’s arrange a time to speak.

References:

Google’s paper: https://arxiv.org/pdf/1810.04805.pdf 

Devlin, J., Chang, M.W., Lee, K. and Toutanova, K., 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.

Github page for the project: https://github.com/google-research/bert

A great Medium post on BERT: https://medium.com/the-artificial-impostor/news-topic-similarity-measure-using-pretrained-bert-model-1dbfe6a66f1d

And part 2 from 2018 dissecting BERT: https://medium.com/dissecting-bert/dissecting-bert-part2-335ff2ed9c73 

It has now been enough time since the September 2019 update to gauge what it targeted more or less and what changes should be made to regain your rankings before or during the next big core update.
What I present here are not individual fixes as they depend on each website, but more of a concise approach that I have used in the past to recover multiple websites from the rankings peril.

Why ‘Leave No Stone Unturned’?

I will also talk about an approach of ‘leave no stone unturned’ SEO auditing to ensure recovering from a penalty. Inspired heavily by Glenn Gabe with his American version ‘kitchen sink’ approach (read more about it here). For some it’s a nobrainer that with multiple Google algorithms at play, you need to cover everything on a site and not just cherrypick your fixes. However – I now use Glenn’s articles (linked above) and research to explain to fellow SEOs or/and clients that this is a valid approach. That often a Google query penalises a site for a number of reasons and if we just carry out one test – the chances of that one element being the only thing at play are miniscule.

I’ve been a SEO consultant in London for many years and I have never seen so much turmoil in Google’s updates as during 2019.

In a nutshell – noone but Google know definitely what exactly triggered the penalty. They did state they’re looking for a big improvement in the website quality overall. So by covering all main aspects of SEO and not only implementing one or two tests that seem most likely – we have the biggest chance of seeing a reversal of any penalty/filter they have applied.

Why it’s necessary to look at these updates ‘holistically’?

These big core updates should be looked at not as individual factors. We don’t get ‘penguin’ or ‘panda’ anymore – they’re not single big changes of one factor.
These new core updates, as much as multiple smaller updates throughout the year, are a collection of sets of ranking adjustments. Alltogether designed to change the landscape.

Sure you may say – the ‘Medic’ update targeted one particular niche more than others, for me – it was targeting trust on the SERPs landscape and, the accuracy of data presented, depth of topic coverage and things like (perhaps) authority or trustworthiness of each individual author.

What does Google want from us?

Here are a few directions Google moves towards with their individual events/focus areas I’ll be discussing:
* Security: HTTPs as ranking factor, recent announcement on blocking partial content
* Mobile experience: speed on mobile and smartphone crawler being used, prioritising mobile experience over desktop for rankings
* Trust: schema and multiple coverage of author as part of the algorithm, knowledge graph taking over more and more, Q&A schema in the SERPs gaining prominence
The new Google core algo update in September 2019 should be looked at through the lens of these changes, as it’s part of a continued process to bring us mobile first, secure, trustworthy search results.

Do you have any case studies of a site making a recovery during the September 2019 Google Core update?

Yes I do, as a freelance SEO consultant it is my job to record and after having asked permission from clients, present the results of the SEO capmaigns. The below is one of the clients I’ve worked with prior to September 2019:

Google Medic recovery in September 2019 Core Update.

As you can see they’ve experienced a drop back in September 2018, then had a small very short recovery in March 2019 and May 2019 in time with other Google updates, but then a slow decline followed.

What were the adjustments carried out to address the September 2019 Google Core update?

Here are the main themes of the updates carried out to improve the situation:
1 Improving the navigation across the site, interlinking and parts of the architecture (there were some potential duplicates, or pages belonging to a section but being in a different folder due to CMS).
2 Addressing the intent vs search query – on multiple pages the visitors were looking for one thing, say ‘apples’ and because our page was about ‘fruit’ in general, we believed we meet the intent, but we never really talked about ‘apples’ specifically. Updates were made to address keywords and queries more appropriately and in a more indepth approach.
Here’s another example of this:
3 Using the ‘left no stone unturned’ and improving the overall website quality, tackling anything significant from the common ‘Google Quality Guidelines’ issues – speed of the site, content duplication, bad experience on the pages, lowering bounce rate etc.

What are the most important (in my opinion) elements to address on any site that is struggling post Medic from 2018 OR September 2019 Core update?

Here is a step by step process I use to review query-to-page relevancy and if the intent is met:

What are some of the most common technical onsite issues that combined may cause problems with Google core updates?

End notes

I hope this post and ideas help you, if you’ve been struck by one of past Google updates. Feel free to get in touch if you’re a bit stuck on what to tackle next on your site, always happy to help.

Hi, my name is Krystian. I’m an SEO consultant and I’ve been doing it for a long, long time, about 13 years now.

What I wanted to talk to you today about is optimizing for Google’s search intent and how to match this intent with our web page.

This is a big, ongoing move by Google. They have been executing on this for years and years. Now, more than ever, they are trying to make sure that whatever your intention is when searching for a given keyword, a given page will be able to deliver you that intention.

First, we’ll take a look inside the Google Search Console.
You should have set up for your website and this is going to be helpful whether you are a web designer, marketing manager or business owner.

I’m looking here on a specific page because I’m doing the recording. Um, I was preparing some notes here, so I use my own data, using an old dataset, keeping it all the new data secret.

These are some of the keywords I used to rank for.

When we talk about intent, we’re talking about specifically making sure that when someone searches they find the right information to give you an idea in the old times, when I would be searching for green card or visa for a given country, Google would very often display news results or editorial coverage about the problem of someone not being able to find the correct answer.

Now that wasn’t matching my intent. My ‘intent’ was to find out their legal requirements. So I had to change the theory a lot to explain to Google what I’m trying to do.

So they got a lot smarter in this, but that meant that some businesses lost a lot of visits while some other businesses gained.

Why was this page getting a lot of impressions, but a few clicks?

It’s about matching that intent of the visitor a lot better. So in this example, let’s have a look at this page. These days I had a big problem, I looked at my technical SEO audits, spammy URL, huh?
You can see the issue here: it had 11,000 impressions and only eight clicks, so a very bad ratio. Now I’m seeing queries to the single page and we will see what was happening here. I was getting a tiny amount of visits, but a lot of impressions. Why?

And so the clicks, I was getting match the intent of the page, right? It was technical SEO, audits, technical SEO etc.

What was the searchers intent? What did they want to find?

Many of them, over 400, wanted to find the cost and other six hundred wanted to find the pricing and other a is 372 wanted to find the price etc.

So what can I do with this information?

If this information, what I need to consider to do, is to take my SEO audit page and update it with the cost information. Or maybe I create a new page.

How do I find out what to do with search intent? New page or updating an old one?

I go to Google results and I will check and Google for that keyword and look at the results. There was one company that had a good page, I’m seeing a provider here and provider here.

I’m getting a provider here and another provider here and another provider, two or three other providers. So it’s looking like a SERP that I can enter.
It’s a results page. I go, Google should allow me to enter it because it’s from the kind of providers, especially if this American information here, and this is a search being done from United Kingdom, so they should prefer UK information. I’m seeing an opportunity here.

Now do I want to make a new page or do I want to update my current page to match the search intent?

Well, in this case I must say I’m probably gonna go with a new page. Now, why is that?

So I opened this pages and what I’m trying to find is, if they’re optimizing one page specifically for this keyword or did they just include this information within the bigger page.

Now it seems to me that’s a specific page, um, that a specific date that’s actually a huge page with a lot of text on.
Interesting. So that’s a bigger ‘what the cost of SEO’ page, but Google still decided to rank it. ‘How much does SEO cost’? Again, general page. On this search we have some example costs.
So specific information that’s mostly news coverage or editorial coverage.

How big is this page?

That’s my next question. If I want to rank for this keyword, how much do I need to put in, how much information?

That’s a huge page – interesting to see how much you can write about the cost of SEO audit. And surely it’s a one of two paragraphs to say how much does it cost? Huh? Ah, okay. Interesting. So they do an audit, market-leading code, it one is to print six months, 2.8. Very interesting pricing strategy here. So it seems to me, eh, again, intent optimising, there is no point of me including this on my page. Technical SEO page, I need to make a new page.

How about another keyword?

It’s a very large page that optimises for ‘SEO audit costs’. And another example. Let’s see.

Should I add information about ‘considerations’ to my page?

If I Google for technical SEO considerations, uh, could those results here? And if you look at the comparison here, there are some providers bidding here again, but when you look at the results, um, what do we have?

Uh, we have editorial covered mostly actually fully. And so there’s no point of me making a page about this because Google is not likely going to allow me to rank for this keyword.

How is this useful for me?

So that’s how I would pick the keywords that I found in this my page.

And then on any page on your website, you can do this with even today:

I did view it in search console, view the keywords against X, how many impressions is even worth pursuing?

Well that’s just because I included it in my technical page. That’s not because I want to present this information.

That’s maybe I should. But then again, one impression, it’s maybe too low.

So I do some bit more research. I can mine all those keywords and compare the intent, demand, clicks vs impressions. And then maybe, maybe I don’t want to make a little guide on how much does an SEO cost, um, for SEO did cost. But that would have to be a separate guide, bigger article.

And that’s how you do search intent optimisation, in a very basic way.

You may find that your page on, you may be selling cars other than service or selling others products. People might be asking questions about this product or asking if you do certain things or out to this product and you may not be providing enough information about it. And that’s when you can find out all the queries that the given page offers.

What would I do in this situation to optimise for search intent?

I would probably do a two fold action here.

One: I’d improve some of this information on my page, and 2: I set up some new pages to optimize for those keywords. Maybe if I make a checklist throughout, there were some impressions for the checklist.

Maybe I should just rework the article as a checklist.

So that’s, that’s just a few ideas here. Guys. I hope this was helpful to you. That’s a, in short, doing this kind of stuff, it’s very, very useful for other websites you can, we gain a lot of visits. You can capture a lot more traffic. And if you get stuck with any of it, just leave a comment there and I will drop you a message.

Session 2 transcript:

Hi, it’s Krystian and again, and I’m here with you for a little supplementary video on the Google intent matching and optimizing pages for intent.

This is from Google’s own blog post, about understanding searches better than before.

This is when they show a very good example on a query when someone wants to know if in 2019 if they’re brazilian and they’re traveling to your site, do they need a visa? This is the old result which had a news article. As you can see from 2019 which was helpful, but it was talking about some problems and it was talking about, it was a news article. So can you really fully trust it? It could be talking about anything to do about US travellers. Now, what actually they wanted to find is the official information from the tourism board, from the embassy, right? Uh, the government, uh, website.

To give you another example, visa to Vietnam.

You can see it’s very specific. Google understands that when I searched for Visa to Vietnam, I don’t want to see a picture of it or example, I actually, what I want to find out is the insurance requirement. That’s my intent here, right?

My intention before visiting Vietnam is ‘what are the entry requirements’. That’s the question I really ask when I type this.

Another example here is um, peanut butter, Quito friendly. Um, Google has to figure it out, right? If I’m searching this though, I wanted to buy um, peanut butter that’s a keto friendly or am I asking ‘is peanut butter keto friendly’? Do I want to see a recipe? And as you can see that Google is divided so they let companies bet here on those results on paid product carousel, they do show a search result here and too, that actually answers the question. They provide similar questions but then they go into recipes and they show quite a few recipes here.

So is Google not understanding the intent here?

It’s an interesting case where Google is really trying to figure out this query. What are they trying to do? We’ll have to have it match the intent. Another example, keto friendly peanut butter, almost the same query, same or just in a different order. And we can see I’m disabled slightly higher. Maybe the results are that asked the question I would say right?

And another example was, uh, asked. So these are the results. So we didn’t see, um, and this is the expert article. Here’s how they, they didn’t write too much. They run this in the article here. The women’s magazine one one, a very similar layout, very similar length, maybe a bit longer is peanut butter kiddo. Frankly. So this is a, this different, uh, theory.

Now they list specifically the ‘answer article’ as number one above everything else as that’s the main intent.
You can see the more commercial on the other SERP. Clearly the ads were first, the recipes done after questions and answers. And then only the article here we can see the article is number one, the recipes are lower down. So we’ll go figure it out. Okay. If they, if they’re asking, they’re definitely asking. And yet people are showing quite fine with the recipes here as well.

Again, a lot of, again, a few more questions here ranking. So whether you’re, um, whether you’re one of those, one of those recipes, a few recipe websites, whether you’re running a recipe website, you should consider probably having some kind of an answer in question. So you can see that even though this, this article here is about, um, it shows a recipe here. Uh, it actually, uh, where do we have the recipe, how to store now, how to make, how to make that with a Lothrop peanut butter ball.

What’s the situation with recipe results and search intent?

So we will have some kind of a recipe here and they have recipes on keto peanut butter. So you’d have that, they have the recipe, you can see that what actually they had to do in order to rank for those skills is have to, uh, is, is it, is it okay like keto diet? Can I eat peanuts on the keto diet? It’s right. And how to eat peanut butter on Keith to dietary goals. So there’s a lot of questions on answering going here and there is a lot of benefits of it. Um, and yeah, so that’s, that’s it in a short, to give you another example. Um, Trinzic tires. You can see here that Google understands that when I’m writing tinting tires don’t I need, uh, a local fig suffocate to 20 of my tires in the car. Um, are, do I want the, how tall do I want to do with myself?

So they’re going to have to understand the intent of that plate. So Bridgestone tires is ranking kill very nicely with the article because RSA is trying to capture some of that action. And then you have a few other websites there against step-by-steps of Google decided that actually if I search this, I most likely want to know how to do this right, how to change the tires if I Google it for uh, how to change a tire, the same guide. So Google understands that this theory is act specifically the same as this clear from my intent. I, the only difference is the circular changing tires actually shows me places you to teach my tire. So they are trying, again, they are helping health meting the Patriots, the results for the intent and here they don’t show that map. So here they understood that if I, if I wanted to know this, I just want to know this, I’m not interested in prices to, they don’t even know I imagine, but they may not even be willing to show them for bidding or maybe they put this in a negative.

Maybe they have, how is a negative cured. So they don’t want to show for this. Maybe treat feet aside. We don’t do, they don’t want those mistake clicks to their website from, from this curate because those people are not interested in them. Don’t make the job for them. So just wanted to give you guys this little overview on how those search results pages are composed by Google and obviously how to rank right. And I think what these guys are doing killer on this website is very, very clever. Even though they wanted to run with their recipe and they showing the guts, they’re showing the poll results because they are, they have a good recipe. Actually, they had to include all the background stories here with questions and answers on the topic, right. In order to keep those results. So there you go. I hope you enjoyed it and stay tuned for more episodes.

NewTide Digital Ltd, 86-90 Paul Street, London, EC2A 4NE. Phone: +447455183193 Company Reg: 9685847 VAT: 265 4314 08
linkedin-squaretwitterbehancecross