Google is known for continually updating and adapting its search algorithms to ensure it’s always providing the most relevant search results for users as search behaviours change. Today, we simply have to type ‘weather in Lanzarote’ and Google will tell us in an instant what the temperature, precipitation, and wind is on the island, right now.

Searches like this are known as simple requests. It means that there is a very specific question being asked, which results in a very specific answer being given. Until recently, Google’s algorithm hasn’t been quite as skilled in handling complex requests.

Complex requests are those that ask a broader question, and require more data sources – and an understanding of how to connect these sources – to answer.

For example:  ‘What do I need to pack for a holiday to Lanzarote?’

To answer this question, Google can’t just consider weather data alone; it has to consider this data alongside a whole host of other possible variables, such as:

  • Interests – are you going to be relaxing on a beach, or hiking volcanoes?
  • Travellers – do you need to pack for a family holiday, or a couples getaway?
  • Location – will you be taking a short haul flight or a long haul flight?
  • Duration – is this going to be a long weekend, or a two week stay?

All of these elements will have a big impact on what needs to be packed.

Most of us would prefer to ask this sort of question to a friend who’s familiar with the island, and who understands our circumstances and needs. They’ll be able to provide a single answer, tailored to our situation, encompassing everything we need to know.

Google can’t do that. In fact, research shows that users typically need to make eight searches on average to gather all the data they need to answer a complex question. That’s because Google can’t fully connect different concepts, topics, and ideas.

But it might be possible soon. One of the most exciting updates to be made to Google Search this year is an advanced artificial intelligence (AI) algorithm called the Multitask Unified Model – or MUM for short – that’s designed to answer complex requests better.

How Does MUM Work?

The Multitask Unified Model works by using Google’s T5 text-to-text framework. T5 is considered to be a ‘big brother’ to the previous 2019 BERT (Bidirectional Encoder Representations from Transformers) update, and uses natural language processing benchmarks to understand and connect information 1000 times more accurately.

Ultimately, MUM achieves two things:

1. Greater understanding of search terms: MUM is better equipped to analyse natural language patterns and derive more meaning from search terms, enabling Google to source the most relevant data and, most importantly, bring all that data together.

2. Greater access to relevant data: One of the most notable advantages of MUM is that it doesn’t just use accessible textual resources to provide answers to questions. MUM can pull data from images, and from texts written in 75 different languages.

Right now, we’re still very much in the very early days of MUM. However, the model is looking to be very promising, and Google has already hinted that future updates may include the ability to draw data from video and audio files, and the incorporation of Google Lens to enable users to make complex requests through their own photographs.

What Does MUM Mean for You?

A big question that many are asking right now is how MUM affects SEO. Right now, the answer to that is that it doesn’t. At least not yet. MUM is still a very young model, and it hasn’t yet been developed and rolled out to a point where SEO is being affected.

That being said, it’s always important to plan for the future. While there’s no official word from Google about how MUM will impact SEO, we can take a guess that less importance will be placed on exact keywords, and more importance on experience.

That’s because, when answering complex queries, Google needs to be able to consider a broad range of data, rather than a single data source connected to a single keyword. With MUM, it’s more important for Google to deliver results that provide a satisfactory, comprehensive experience, than specific keyword based information.

So if you haven’t already started to ensure you’re writing for people, not just for machines, then it never hurts to start optimising your approach for experiences.

If you’re ready to discover more opportunities to elevate your marketing in 2022, contact the Adrac team today.

Author Rebecca

More posts by Rebecca
shares