'Reviving a Classic': The Traditional Marketing Tool Reshaped for Modern Success

Dorothea Wiesmann Rothuizen and Wojtek Skut, Google engineers, continuously work towards the enhancement of AI and large-scale language models. They are integral members of the team responsible for infusing Google Ads with new functionalities, such as broad match.

In recent times, we've seen a surge in the interest surrounding artificial intelligence (AI), largely propelled by considerable enhancements in its fundamental technology. Large-scale language models are now capable of deeper language understanding, leading to more natural and intuitive interactions with AI.

Broad match, a setting within Google Ads that aids businesses in expanding their ad reach and minimising manual keyword efforts, benefits significantly from these advancements.

While broad match has been a staple of Google Ads since its inception, it wasn't flawless. It began to lose its lustre approximately 15 years ago. It had the tendency to match unrelated keywords to search queries, for instance, a query such as "home pet care" might be matched with an irrelevant keyword like "pet treats".

Moreover, it couldn't establish a connection between related phrases if they lacked identical keywords. Thanks to the strides in AI and continuous upgrades, broad match can now better decipher nuances and context.

Now, broad match can probe deeper into individuals' search queries, understand intent, and has become one of the most potent tools for search advertising.

The Broadening Scope of Marketing through AI Advances

When broad match was initially launched, the engineering team at Google would manually formulate synonyms for a keyword and cast a broad net. For instance, an ad keyword like "cheap" could be matched with searches for "low-cost" or "shoestring". While "shoestring" is another term for "budget", it could also be used by users to search for shoelaces.

With new machine learning capabilities, large language models have become more adept at comprehending a user's intent. They are trained with billions of pieces of text to grasp the various usages and connotations of a word or phrase and to understand which sequences make sense. This fundamental understanding is the foundation before assigning it to specific tasks, such as matching ads to search queries.

While the language technology can comprehend the potential demand, its effectiveness amplifies when paired with Smart Bidding. This automated bidding strategy considers a variety of signals and surrounding data points in a privacy-safe manner — like search history, interests, and previous purchases — to construct a predictive model that aids advertisers in pinpointing audiences with a high likelihood of conversion.

Now, broad match can interpret individuals' search queries more profoundly and is capable of discerning their intent.

Broad Match in 2023: What's in Store for Marketers

Google’s engineering team has made substantial improvements to broad match. Here are a few:

The order of words matters, and broad match acknowledges this too

Initially, broad match would establish connections between keywords in a search query and an ad but wouldn't necessarily consider the order of the words in the query.

Suppose someone was searching for a flight from Johannesburg to Lagos; they might have been presented with an ad for a flight from Lagos to Johannesburg. To avert such occurrences, search specialists at Google would have suggested a different keyword strategy to advertisers. Nowadays, broad match understands that going from A to B isn't the same as the reverse. This enhancement has been applied to other areas where a search query is uni-directional.

A user might want to "convert nairas to shillings" or be looking for a "USB to USB-C cable". Ads that swap these words around would not be pertinent to them — but now, broad match comprehends their intent.


Channelling Traffic to Appropriate Keywords in Google Ads

A previously raised query was: "What occurs when two ads from the same advertiser match the same search query"? Earlier, broad match would consider the Ad Rank of an advertiser to decide which ad to display, which might not always have been optimal.

To bolster relevance — for both the advertiser and the user — it now takes supporting details more into consideration, like other keywords in the ad group and the ad's landing page. Moreover, with the upgraded large language model powering broad match, it can understand the meaning of the search query and its variations better, further enhancing relevance.


Multilingual Matching

Considerable progress has been made in making connections on multilingual searches. Broad match now acknowledges that some people might interchange languages when searching online and utilises that information to serve appropriate ads.

This is often the case with expatriates, such as Nigerian natives living in South Africa or Europeans in Kenya. They may be searching in their native language but are looking for results in the local language of the country they reside in.

We're not suggesting that running your campaign in a single language will garner traffic for all, as your campaign's success depends on your language settings and localised creatives. However, in countries with numerous bilingual expatriates, it's beneficial to know that broad match can match relevant, local traffic to an ad, even if the search is in another language.

AI-powered language improvements have enhanced broad match as a tool for search advertising, but it still requires a savvy marketer at the helm.

The Current Role of Broad Match in Search Advertising

While broad match may not be as old as "Back to the Future", it's been a mainstay for quite a while. Much like in the iconic 1985 film, it demonstrates that sometimes, we must revisit the past to recognise its worth in the present.

Even though AI-powered language improvements have made broad match a potent tool for search advertising, it will continue to need an astute marketer in control.

These sharp marketers will monitor performance, choose the appropriate bidding strategy, and supervise the new queries that broad match identifies to draw in new customers. You might uncover new insights about your audience and their behaviour during the campaign.

Broad match’s success relies on the team utilising it, similar to how the time-travelling car from "Back to the Future" required Doc Brown's ingenuity to journey into the past — and the future.

For a more in-depth discussion on how large language models are enhancing broad match, tune into the Google Marketing Platform podcast episode "Talking AI: How a deeper language understanding is transforming broad match". It can be found on the GMP Academy website and all major podcast platforms.