22nd September 2021

Where is Google heading in the next generation of search?

User Photo
James Gardner
Head of SEO
Read time: 5min
Left Angle Mask

In June 2021, Sundar Pichai the boss of Google stated, “AI will have a bigger impact than fire”.  Even if we dial down the hyperbole by a few orders of magnitude, this is a huge statement from someone respected for their expertise.

AI impact on Search

Leaving the element of fire to Pitchai, let’s look at how his company is planning to use AI to power the next generation of search.

Google recently introduced the concept behind its next-generation search algorithm called Multitask Unified Model (MUM). To be clear, this is not ready yet and may not be widely available for 12 – 24 months. But it will be 1000 times more powerful than Google’s current algorithm BERT that was launched in 2019. It will integrate information from 75 world languages to provide the searcher with the most relevant answer.

Okay Google

Let’s ask Google a couple of questions that MUM is aiming to answer:

“Okay Google, I climbed Ben Nevis last summer, but I want to climb it this November, what should I do differently to prepare?” 

“Okay, Google, here is a photo of the boots I wore last time, will they be suitable this time?”

Images source: Google
Image source: https://blog.google/products/search/introducing-mum/

I am asking Google to understand a highly complex question requiring a significant grasp of the context and very specific expertise. The exact expertise of a mountain guide living in Fort William, perhaps?

Let’s ask Google the question today.

Google does a decent job of returning sites and blogs written by experts on the topic. I just need to read several of them and then join the dots to find the answer to my questions about winter mountaineering in Scotland.

Ben Nevis SERP

Joining the dots

MUM promises to do the reading for me and provide a tailored answer to my nuanced question. It will probably look like today’s featured snippet but rather than pulling a paragraph from a single source, it may curate an answer from multiple sources.

Here is a simpler question for Google which it does well to answer by pulling the appropriate paragraph from the Abacas Mountain Guides blog.  The author of this blog seems to be someone with the right expertise and authority to allow us to trust (EAT) their advice: he is an IFMGA Mountain Guide who provides ‘world-class mountain adventures on the west coast of Scotland’.)

Ben Nevis SERP with Images


But what if we ask the same question in French?  Google provides the rather less useful: “It is this dreadful weather, influenced by the nearby Atlantic Ocean, which creates climbing conditions that cannot be found anywhere else. And above all, this experience is possible at all difficulty levels, without being altered. We can almost speak of winter mountaineering for everyone!”

Translated Search Results

This is probably because there is far less and more generalised information about Ben Nevis published in French.

A key element of MUM is that it will be able to scour the web for the best answer across 75 languages. Perhaps it will simply translate the answer from Abacus Mountain Guides into French.   Most likely, it will also draw on additional sources to tailor the content of the answer specifically for a French climber.  

All very helpful, but we can see why Google needs extensive testing of this algorithm to ensure that the information that it provides is accurate especially for YMYL (your money or your life) sectors. What if Google says my boots are just fine for climbing Ben Nevis in November, but then I injure myself falling in the snow?  It seems likely that Google will continue to link to the source of its answers.

Impact on SEO

If Google is personally curating my answers, why should I ever need to leave the search results page (SERPs) and visit any websites?  Certainly, MUM will keep users on SERPs for longer for informational type searches but more than ever we need to be highly present for when the searcher is ready to click through to engage with our brand.

The challenge will be twofold:

a. Rather than just competing against websites in our language, we are now competing against sites in all 75 world languages.

b. If searchers are less likely to click through to our website, how can we build our brand, start them on a transactional journey and ultimately monetise our expertise?

What can you do now?

We need to be redoubling on our efforts to demonstrate our Expertise, Authority and Trust (EAT) in all the content that we produce so that Google surfaces it. Back to the “10x” mantra: how is our content 10 x better than other content already on the web?

International sites that create content centrally in English and translate to local languages need to reconsider the value of this. Content needs to be strongly localised rather than simply translated.   

SEO evolving

SEO will continue to change but remain central to a brand's digital strategy. It will remain laser-focused on providing the most relevant answers to match the search intents of the target audiences.

If you’re interested in learning more about how equimedia can help you prepare your SEO strategy for changes such as the Google MUM algorithm change, or if you’d like a more general chat about our SEO services then please get in touch.