fbpx

Google Ads Makes Keyword Matching More Predictable Search Engine Rolls Out Improved Understanding Of Search Intent And More Predictability In How Keywords Match

author image Written by: Nicole McCormick           Categories - In The News, Paid Ads

Some big news came from Google ads yesterday. To summarize, the search engine announced changes to how it uses BERT to better understand keywords and search intent. However, there’s a bit more to unpack.

The biggest changes relate to how Google handles phrase and broad match keywords when they match, what happens when they are not identical, and how BERT is used to match keywords to queries.

See Google’s full statement below:

“What consumers want and how they search for it is constantly changing. In fact, we see billions of searches every day, and 15% of those queries are ones we haven’t seen before1. To help you reach these new, relevant searches while still meeting your performance goals, we recommend using broad match paired with Smart Bidding. On average, advertisers that switch their exact match keywords to broad match can see more conversions and a higher conversion value.”

Over the past year, we’ve made it easier to find success with this strategy by improving broad match relevance, recommending new opportunities to switch to broad match, and making keyword matching more predictable. However, we’ve also heard that you want higher quality and more control over which keywords match a search. That’s why we’re rolling out new improvements based on this feedback, including improved understanding of search intent and more predictability in how keywords match.”

Now, let’s dive a bit deeper.

Improved Understanding Of Language And Intent

Because keywords can match to multiple different queries, Google is using technology like BERT to interpret language, queries, and search intent, which helps with broad match.

Google provided the following example:

“A highly specific query like “1995 5 speed transmission seal input shaft” is now able to match with the broad match keyword auto parts because we can tell they’re related, even though none of the words in the query and in the keyword actually match. This means that broad match can now help you find more relevant, high-performing traffic using fewer keywords.”

For those who are unfamiliar, BERT is a neural network-based technique for natural language processing pre-training. Google uses this tool to better understand the context of words in search queries.

READ MORE: An In Depth Look At Google’s Most Recent Colossal Update BERT

Better Predictability in Matching Keywords

It’s a common occurrence for multiple keywords to match with a particular search, search, which often leads to searches being matched with less relevant keywords.

Google has made the following improvements to help ensure the most relevant keywords are being selected:

When A Search Is Identical To One Of Your Keywords

Google previously announced that exact match keywords that are the same as a search query will now always be preferred. Now, Google says that both phrase and broad match keywords will follow the same rule.

See Google’s example below: 

“Let’s say someone searches for “sushi delivery near me”, and you have the broad match keywords sushi delivery and sushi delivery near me. Before this update, both of these keywords would be eligible to serve. Now, the keyword sushi delivery near me is preferred because it is identical to the search term. Note that if you have an eligible exact match keyword that is identical to the query, it will still be preferred over the phrase and broad match keyword.”

When A Search Is Not Identical To Any Of Your Keywords

In the good old days, you would have multiple keywords eligible to match, none of which were identical to a search. In these instances, your Ad Rank would decide which keyword matched best. Now, Google is looking at relevance signals along with Ad Rank to help determine which keyword will be chosen.

Learn more about the keywords that are eligible along with how they are selected by checking out the official Google Ads announcement.

Nicole McCormick

Nicole is a wordsmith wizard, passionate about the written word and an avid storyteller who uses creatively crafted prose to help bring your brand’s story to the next level. A former journalist with writing credits in both local and national news publications and a few newspaper awards under her belt, Nicole now enjoys telling your stories and finding new and creative ways to create valuable content that resonates with audiences in the digital landscape.

Google Uses MUM to Identify Vaccine Variations in Over 50 Languages In Seconds

07/08/2021

Nearly two months after announcing the Multitask Unified Model (MUM), Google began testing out its first application of the technology and is reporting some pretty impressive results. According to a blog post published on July 6, Google was able to identify more than 800 variations of vaccine names in over 50 languages in just seconds during its first-ever application of MUM. The blog post stated that without MUM, this would have taken the search engine weeks to complete.

After validating MUM’s findings, Google applied them to Search so that people could “find timely, high-quality information about COVID-19 vaccines worldwide.”

“Our ability to correctly identify all these names is critical to bringing people the latest trustworthy information about the vaccine,”said Google. “But identifying the different ways people refer to the vaccines all over the world is hugely time-intensive, taking hundreds of human hours.”

READ MORE: Google Announces New MUM Technology 1,000 Times More Powerful Than BERT  

What is MUM and How Does It Work?

Announced back in May at Google I/O, MUM is a technology developed by Google that is built on a transformer architecture. It has been compared to the BERT algorithm that rolled out in 2019,  however, Google is reporting that MUM is 1,000 times more powerful and has incredible multitasking capabilities.

READ MORE: BERT Explained – Everything You Need to Know About Google’s Biggest Update in 5 Years

Google explained that MUM can learn from and transfer knowledge across 75+ languages.

Here is a great example provided by Google that demonstrates what this looks like:

“Imagine reading a book; if you’re multilingual, you’d be able to share the major takeaways of the book in the other languages you speak — depending on your fluency — because you have an understanding of the book that isn’t language- or translation-dependent. MUM transfers knowledge across languages much like this. Similarly, with its knowledge transfer abilities, MUM doesn’t have to learn a new capability or skill in every new language — it can transfer learnings across them, helping us quickly scale improvements even when there isn’t much training data to work with.”

This means that MUM requires far fewer data inputs than previous models to accomplish the same task.

While its first application focused on identifying vaccines, going forward, Google will be looking at ways to use MUM to dramatically improve Search as a whole.

“Our early testing indicates that not only will MUM be able to improve many aspects of our existing systems, but will also help us create completely new ways to search and explore information,” said Google.

read more

Google Announces New Technology 1,000 Times More Powerful Than BERT

05/20/2021

Exciting news from Google this week – the search engine revealed some cool new technology that will dramatically affect the way Google understands and responds to complex search queries.

The technology is called Multitask Unified Model (MUM), and was announced this week at Google’s I/O conference.

Similar to BERT, which was launched in 2019, MUM is built on a transformer architecture but is 1,000x more powerful. MUM is also capable of multitasking in order to connect pieces of information for users in ways that no other search engine can do.

What makes MUM different from BERT is that it is trained to understand 75 languages and can perform numerous tasks simultaneously. It’s also able to understand information in the form of images, text, and video.

During the I/O conference, Google SVP Prabhakar Raghavan used the following query example to demonstrate how MUM works:

“I’ve hiked Mt. Adams and now want to hike Mt. Fuji next fall, what should I do differently to prepare?” Leveraging MUM, Google will provide the user with highly relevant results, including articles for the equipment to hike Mt. Fuji, along with the differences and similarities between the two mountains.

This would enable users to conduct searches that were previously thought to be far too complex for search engines. It’s not yet known what this will mean in terms of SEO, as MUM is still in the pilot phase, and no launch date has been announced.

read more