In 2023, we can expect further updates from Google which will shape the SEO landscape. Our Head of SEO, Dave Cousin, outlines five key trends to look out for next year.
Google’s use of machine learning technologies to weed out content created by machine learning is likely to be one of the main SEO talking points of 2023.
This could become a game of cat and mouse, where Google gets better at spotting AI-generated content only for the AI to get better at masquerading human-created content. This is likely to be a battle Google will eventually win – because they have an index of the entire (discoverable) internet and the resources to allow their machine learning to improve at an exponential rate, perhaps matched only by Baidu, which is trying to achieve a similar objective.
Already those using AI to create content are evolving from purely AI-created content to AI content created with some level of human input. How well this can work remains unclear, mainly because Google’s Helpful Content update – which aims to penalise AI content – is still rolling out. What proportion of page content needs to be new, original, and helpful to avoid falling foul of the update? SEOs will spend 2023 trying to work out the right formula.
There has been a measurable impact from the Helpful Content update (HCU). But for most sectors, the impact has not been as dramatic as some previous Google updates. Where impact has been significant, it has affected sites which contain a high proportion of generic content similar to others (such as lyric sites, for example). These sites are similar to those which have been impacted by some previous core updates which implies that some form of ‘helpful and original content’ ranking factor almost certainly already existed.
There should be more to the Helpful Content update though, including penalties for sites with ‘unhelpful’ pages and benefits for those showing expertise with a cluster of related helpful content. It’s likely that we haven’t yet seen the full impact of the HCU and that Google will continue to tweak it and slowly increase its weighting over time. This is standard for a new element of the algorithm so we may see the HCU ramping up after core updates in 2023.
While Google try to prevent AI-generated content from ranking well, it appears as though the features on SERPs pages will be based on Google’s own machine learning.
MUM, the Multitask Unified Model, is a machine learning model that is already improving semantic matching between users’ searches and the contents of Google’s index and is starting to be rolled out to allow searches with multimodal inputs, such as text and images (think ‘I want to buy a shirt like this’).
One aspect of MUM that we have been expecting, but are yet to see, will analyse content from various sources and languages to provide a bespoke, machine-generated answer to a user’s query directly in the SERPs. Think of this as a sort of super snippet. This will even be able to predict and answer users’ follow up questions before they ask them. We saw the announcement back in 2021 and analysed the research papers released by Google employees. Surely in 2023 this will be ready to roll out fully?
Look out too for similar changes on Baidu as they make use of their rival technology to MUM, which is known as Ernie 2.0.
One publicised aspect of MUM was its ability to source information to answer questions from content in other languages – and to input that answer within the SERPs. Google already translate in real time content from English into certain languages where there are limited results in those languages (such as Indonesian, Hindi, Kannada, Malayalam, Tamil, Telugu and others).
The Catch-22 is that to provide high quality translations, there has to be enough content in a language for the machine learning-based translation to learn from, and the languages that will benefit most from these translated results are those with little content. Google is therefore adding more and more previously offline content, such as books in those languages, to improve translations.
Google is likely to add more languages in the near-future once it has sufficient confidence in the translations of pages it can provide. Whether Google will, in future, also add languages with more results already available remains to be seen, but we may well find out in 2023.
A non-Core Web Vital is a web page performance KPI that is not yet part of Google’s algorithm but for which scores are available. In 2022, we found out about a new non-Core Web Vital called INP, which stands for Interaction to Next Paint, and many are predicting it will become a fully-fledged Core Web Vital in 2023. And this may not be the only change coming to the Core Web Vitals’ part of the Page Experience aspect of Google’s algorithm in 2023.
INP is likely either to replace or complement FID (First Input Delay) as a Core Web Vital. FID measures only where there is one long process that delays a browser from reacting to a user action. This omits multiple types of perceived delay that might frustrate users. INP will be based more on what a user actually sees, as it measures the total wait between a user interacting and seeing a visual change or confirmation on their screen.
FID isn’t the only current CWV that some think is flawed though, and there are other non-Core Web Vitals waiting in the wings. LCP (Largest Contentful Paint) is meant to be a way of measuring how fast a page loads as perceived by a user. The problem is that it allows sites to optimise for one specific element on the page to obtain a good score but potentially still have a bad load time for everything else on the screen. Perhaps a Web Vital like SI (Speed Index) could be added to reflect the loading of multiple elements, or a completely new Web Vital could be created.
Google announced CWVs with much fanfare and many SEOs and site owners expected Armageddon for poor websites. The reality was that so many sites would have been penalised that Google could only make it a fairly minor ranking factor or else many top sites would have dropped out of rankings. However, this doesn’t mean that Google have given up on CWVs so their importance could still be tweaked, especially if Google have new CWVs in which they are more confident. Expect CWVs to raise their head again in 2023. Oban’s advice is to optimise for INP and a truly fast web page load time now.
For more on Core Web Vitals in 2023, please see Oban’s useful webinar here.
. . .
Planning an international SEO campaign in 2023? Oban can help. To find out how, please get in touch.
Oban International is the digital marketing agency specialising in international expansion. Our LIME (Local In-Market Expert) Network provides up to date cultural input and insights from over 80 markets around the world, helping clients realise the best marketing opportunities and avoid the costliest mistakes.
