On May 18 (May 19 in Australia), Pandu Nayak, Vice President of Search at Google announced a huge algorithm update.
The update uses something called a Multitask Unified Model, or MUM, to provide solutions to complex queries that would usually take a user multiple searches to answer.
What exactly will MUM do?
In the example provided a hiker has previously climbed Mt. Adams (in the state of Washington) and is now preparing to climb Mt. Fuji in Japan and they want to know what they need to do to prepare for the trip.
This type of informational search would usually require multiple steps and would require the user to consider equipment, physical preparation, and more. According to Google, this type of informational search usually takes ~8 queries to be answered.
Despite not being live yet, Nayak did give us an example of how MUM would work. MUM does not just understand language, but also generates it. The example shows the user’s query and a number of responses by Google on an interface that resembles live chat.
Nayak also stated that MUM is multimodal. This means that it can digest information from text, images, and video. The second example in the article shows an image of some hiking boots tagged with the question ‘Can I use these to climb mount Fuji?’
Nayak explained that this image could then be used to point users towards a blog post with a list of recommended gear.
What it could mean for SEOs
Reduced website traffic from search
I’m not as manic as some of the doomdayers in the comments section over at Search Engine Roundtable, but I do see the potential for the MUM update to reduce traffic from search.
If Google is answering complex queries in a chat-like interface and reducing the number of searches (around 8) that a user has to perform this means that the user will not visit those 8 websites and will not exposed to their brands.
At the end of the day we have to remember that Google has two major goals:
- Create a better search engine (their main product) that more effectively answers users queries
- Keep users inside the Google ecosystem so it can serve them revenue-generating ads
The MUM update could do both of these things, which would have a big impact on blog/website traffic.
Increased reach across countries/different languages
MUM is trained in 75 languages. The objective will be for MUM to to collect and digest data from sources that the end user would not be able to translate themselves.
This means that the user will be exposed to information that originates from a much broader variety of websites than they would usually have access to.
For us SEOs, that means our content may be retrieved by MUM and served to users that are unable to speak English – thereby extending the overall reach of the content (however, it is unclear at this stage whether the source of that information will be attributed).
Multi modal – blog & product opportunities
In the example above, Google is able to understand and assess the image of a hiking boot and decide whether the boot would be appropriate for Mt. Fuji.
If the boot provided in the example was not appropriate the search engine would suggest products or content that would help the user find appropriate footwear.
This is a potential source of sales for SEOs/marketers.
The MUM update is going to be big for users and for SEOs. If successful, users will access a much more powerful search engine with the ability to answer complex questions using a variety of modalities as input.
I believe this update could significantly impact SEOs and digital marketers – reducing traffic to blog posts, but also providing opportunities to sell products that are recommended through the MUM interface.
If you’d like to connect with experts on the MUM expert – check out my new agency: mumseo.com